THE RESEARCH AND DEVELOPMENT PROGRAM IS FUNDED BY THE AUSTRALIAN GOVERNMENT DEPARTMENT OF HEALTH AND AGEING

AS PART OF THE FOURTH COMMUNITY PHARMACY AGREEMENT

FULL FINAL REPORT

A national funding model for pharmacotherapy treatment for opioid dependence in community pharmacy

Researchers: Dr Anne-Marie Feyer, PricewaterhouseCoopers

Professor Richard P Mattick, National Drug & Alcohol Research Centre, University of New South Wales

Carrie Schulman, PricewaterhouseCoopers

Rebecca Jessop, PricewaterhouseCoopers

Jeremy Solomon, PricewaterhouseCoopers

Deanna Pyper, PricewaterhouseCoopers

2

FULL FINAL REPORT

Acknowledgement

PricewaterhouseCoopers would like to thank the Department of Health and Ageing for funding this project as part of the Fourth Community Pharmacy Agreement, which has been administered by the Pharmacy Guild of Australia.

The researchers would like to make special thanks to the pharmacies and pharmacists in addition to the clients who participated in this trial, without whom this study would not have been possible.

Thank you to both the Advisory Panel and the Reference Group who provided guidance and input into the study and trial design.

This report was produced with the financial assistance of the Australian Government Department of Health and Ageing. The financial assistance provided must not be taken as endorsement of the contents of this report.

The Pharmacy Guild of Australia manages the Fourth Community Pharmacy Agreement Research & Development which supports research and development in the area of pharmacy practice. The funded projects are undertaken by independent researchers and therefore, the views, hypotheses and subsequent findings of the research are not necessarily those of the Pharmacy Guild.

3

FULL FINAL REPORT

Project team

Lead researchers

Dr Anne-Marie Feyer, PricewaterhouseCoopers

Prof Richard P Mattick, National Drug and Alcohol Research Centre, University of New South Wales

Quality and assurance partner

John Cannings, PricewaterhouseCoopers

Project director

Carrie Schulman, PricewaterhouseCoopers

Project manager

Rebecca Jessop, PricewaterhouseCoopers

Research team

Jeremy Solomon, PricewaterhouseCoopers

Deanna Pyper, PricewaterhouseCoopers

Brian Sabet, PricewaterhouseCoopers

Data management team

Monica Brabant, PricewaterhouseCoopers

Michael Allwright, PricewaterhouseCoopers

4

FULL FINAL REPORT

Advisory Panel Members

Toni Riley (Chair, from March 2010) - The Pharmacy Guild of Australia / Community Pharmacist

Chris Boag (Chair, project commencement to February 2010) - Intergovernmental Committee on Drugs

Paul Feldman – Department of Health and Ageing

Kim Bessell – Department of Health and Ageing

Dr Robyn Davies - Department of Health and Ageing

Denis Leahy - The Pharmacy Guild of Australia / Community Pharmacist

Khin Win May – The Pharmacy Guild of Australia

Irvine Newton - Pharmaceutical Society of Australia / Community Pharmacist

Sarah Lord – Australian Injecting and Illicit Drug Users League

5

FULL FINAL REPORT

Acronyms

Acronym Explanation

ACT Australian Capital Territory

BTOM-C Brief Treatment Outcome Measure – Concise

CI Confidence Interval

EOI Expression of interest

GEE Generalised Estimating Equation model

Guild Pharmacy Guild of Australia

NDARC National Drug and Alcohol Research Centre

NSW New South Wales

NT Northern Territory

PhARIA Pharmacy Access/Remoteness Index of Australia

PwC PricewaterhouseCoopers

QLD Queensland

RCT Randomised Controlled Trial

SA South Australia

SAS Statistical Analysis Software

SQL Structured Query Language

TAS Tasmania

VIC Victoria

WA Western Australia

WHOQoL-8 World Health Organisation Quality of Life-8 scale

6

FULL FINAL REPORT

Table of contents

1. Background and purpose of the research .......................................................................... 9

1.1. Contextual background ............................................................................................ 9

1.2. Rationale for the research ...................................................................................... 13

1.3. Objectives and hypotheses .................................................................................... 14

1.4. Research project outline ........................................................................................ 14

1.5. Purpose of the report ............................................................................................. 15

1.5.1. Roadmap to the report .......................................................................................................... 15

2. Approach to the design of the trial ................................................................................... 17

2.1. Advisory Panel and Reference Committee for the research project ........................ 17

2.1.1. Advisory Panel ...................................................................................................................... 17

2.1.2. Reference Group ................................................................................................................... 17

2.2. Literature review .................................................................................................... 18

3. Trial design ..................................................................................................................... 19

3.1. Development of the randomised controlled trial approach ...................................... 19

3.2. Design of the funding model for the trial ................................................................. 20

3.3. Collection of data from pharmacies ........................................................................ 21

3.3.1. Activity based costing study .................................................................................................. 22

3.4. Collection of data from clients ................................................................................ 22

3.4.1. Brief Treatment Outcome Measure – Concise ..................................................................... 22

3.4.2. Quality of life ......................................................................................................................... 22

3.4.3. Financial strain and satisfaction ............................................................................................ 22

3.4.4. Administration of the client survey ......................................................................................... 22

3.5. Collection of data from health authorities ............................................................... 23

3.6. Timeframes for the research project ....................................................................... 23

3.7. Ethics approval ...................................................................................................... 24

4. Trial methodology............................................................................................................ 25

4.1. Overview of trial activities and timelines ................................................................. 25

4.2. Sampling framework .............................................................................................. 25

4.3. Recruitment process .............................................................................................. 28

4.3.1. Pharmacy recruitment ........................................................................................................... 28

4.3.2. Client recruitment .................................................................................................................. 30

4.4. Allocation to intervention and control groups .......................................................... 31

4.5. Pharmacy data collection ....................................................................................... 32

7

FULL FINAL REPORT

4.5.1. Pharmacy information form ................................................................................................... 32

4.5.2. Pharmacy satisfaction survey ................................................................................................ 33

4.5.3. Activity based costing study .................................................................................................. 33

4.6. Client data collection .............................................................................................. 35

4.6.1. Client survey ......................................................................................................................... 35

4.6.2. Client dosing data .................................................................................................................. 36

4.6.3. Client drop-out data ............................................................................................................... 36

4.7. Data management processes ................................................................................ 36

4.7.1. Access database ................................................................................................................... 36

4.7.2. Data quality ........................................................................................................................... 37

4.8. Data analysis design .............................................................................................. 37

5. Trial results – Client ........................................................................................................ 38

5.1. Client recruitment and retention ............................................................................. 38

5.2. Client demographics .............................................................................................. 39

5.3. Retention in treatment ............................................................................................ 42

5.4. Compliance with treatment ..................................................................................... 45

5.5. Drug and alcohol use and related behaviours ........................................................ 46

5.5.1. Baseline drug and alcohol use and related behaviours ......................................................... 46

5.5.2. Six-month drug and alcohol use and related behaviours ....................................................... 49

5.6. Health and social outcomes ................................................................................... 51

5.6.1. Baseline health and wellbeing ............................................................................................... 51

5.6.2. Six-month health and wellbeing ............................................................................................ 53

5.6.3. Baseline quality of life ............................................................................................................ 54

5.6.4. Six-month quality of life ......................................................................................................... 56

5.7. Client financial strain .............................................................................................. 56

5.8. Client satisfaction ................................................................................................... 58

5.8.1. Baseline client satisfaction .................................................................................................... 58

5.8.2. Six-month client satisfaction .................................................................................................. 59

5.9. Costs of pharmacotherapy treatment to clients ...................................................... 59

5.9.1. Background and context ........................................................................................................ 59

5.9.2. The costs of participating on the program for clients ............................................................. 60

5.10. Summary of client results ....................................................................................... 62

6. Trial results – Pharmacy .................................................................................................. 63

6.1. Pharmacy recruitment and retention ...................................................................... 63

6.2. Pharmacy demographics ....................................................................................... 64

6.3. Pharmacy satisfaction ............................................................................................ 64

6.3.1. Baseline pharmacy satisfaction ............................................................................................. 65

6.3.2. Six-month pharmacy satisfaction ........................................................................................... 65

8

FULL FINAL REPORT

6.4. The service delivered by community pharmacists and the associated costs .......... 67

6.4.1. The service delivered by the community pharmacist ............................................................. 67

6.4.2. Average time spent by pharmacists on regularly occurring activities .................................... 69

6.4.3. The time and cost drivers of the provision of pharmacotherapy treatment in the

community pharmacy ............................................................................................................ 70

6.5. Summary of pharmacist results .............................................................................. 75

7. Key elements of a national funding model for pharmacotherapy treatment in community pharmacy: Implications of the national trial ............................................................................... 76

7.1. Implementation of a nationally consistent funding model ........................................ 76

7.2. Subsidisation of client payments: the impact on affordability for treatment for clients ............................................................................................................... 76

7.3. Reimbursement of pharmacies: how the service is delivered and the associated costs .................................................................................................... 77

7.4. Strengths and limitations of the study ..................................................................... 78

7.5. Recommendations for key elements of a national funding model ........................... 78

9

FULL FINAL REPORT

1. Background and purpose of the research

1.1. Contextual background

Benefits of pharmacotherapy treatment

Since Methadone was introduced as a substitution treatment for opioid dependence just over four decades ago, studies have consistently shown that opioid users reduce their illicit opioid use and improve their health and social wellbeing while maintained on Methadone (Hall & Wodak, 1999; Mattick, Breen, Kimber, & Davoli, 2004, 2008; Mattick, Kimber, Breen, & Davoli, 2004; Ward, Hall, & Mattick, 1999; Ward, Mattick, & Hall, 1992; Ward, Mattick, & Hall, 1998). Despite often quite ferocious opposition to these treatments, orally administered Methadone and Buprenorphine maintenance treatments have become accepted as effective and safe treatments for opioid dependence, supported by policy-makers, clinicians and researchers.

While there are criticisms of these treatments, the evidence which has accumulated from randomised controlled trials shows that these interventions retain clients in treatment, and while in such treatment, there is a marked and sustained reduction in injecting opioid use (Mattick, Breen et al., 2004, 2008). Higher daily doses of Methadone are associated with better outcomes and maintenance treatment approaches (rather than abstinence-oriented methadone-to-withdrawal) are likely to enhance client wellbeing (Ball & Ross, 1991). When they leave treatment, however, most resume illicit opioid use (Ball & Ross, 1991; Dole & Joseph, 1978).

While in maintenance treatment, the physical health and wellbeing of clients improves and there is evidence of markedly reduced mortality (Caplehorn, Dalton, Cluff, & Petrenas, 1994; Caplehorn, Dalton, Halder, Petrenas, & Nisbet, 1996; Darke, Degenhardt, & Mattick, 2007). Additionally, there is very substantially reduced crime (Lind, Chen, Weatherburn, & Mattick, 2005), so that there is a cost-benefit of being in treatment from reduced crime and other cost offsets (P. G. Barnett & Hui, 2000; P.G. Barnett, Zaric, & Brandeau, 2001; Harwood et al., 2002; Shanahan, Hetherington, Mattick, & Weatherburn, 2007).

Retention in treatment is the major variable used by all commentators to assess the effects of pharmacotherapy treatment, as it is the best single predictor of treatment success. Without retention in treatment, all of the other treatment gains quickly evaporate. All major reviews rely on retention in treatment to assess treatment effectiveness (Amato et al., 2005).

Maximising the health benefits from the treatment of opioid dependence is reliant on a number of factors to do with the effectiveness of the pharmacotherapy used, its accessibility and its attractiveness to the target group – dependent opioid users - however the ability to retain these clients in treatment and provide long-term care depends on a number of factors, including the affordability of the treatment.

Pharmacotherapy treatment in Australia

In January 2007, the Australian Government Department of Health and Ageing released the report National Pharmacotherapy Policy for People Dependent on Opioids. This provides a broad policy context and a framework for state and territory policies and guidelines that are concerned with the treatment of opioid dependence using:

• Methadone hydrochloride - Methadone Syrup, Biodone Forte

• Buprenorphine - Subutex

• Buprenorphine/Naltrexone – Suboxone.

In Australia, pharmacotherapy treatment is focused on either harm minimisation using opioid substitution treatments, such as Methadone or Buprenorphine maintenance therapy or abstinence-based therapy using Naltrexone. Methadone is currently the most widely used pharmacotherapy in Australia. Buprenorphine is an alternative treatment to Methadone and has been approved for pharmacotherapy treatment since 2000.

10

FULL FINAL REPORT

The goals of Australian pharmacotherapy treatment are:

• to reduce or cease harmful heroin/opioid use

• to improve physical health

• to reduce risk for infectious diseases

• to improve psychological, social and occupational functioning, including a reduction in crime.

The Australian Government funds the cost of Methadone and Buprenorphine for the treatment of opioid dependence. In most states, treatment is provided free through publicly funded clinics and public hospitals. In community pharmacies and private clinics, clients are charged a fee. The Australian Government does not directly contribute to the cost of pharmacotherapy services at the community pharmacy level and a patchwork of funding and support programs is provided by states and territories. These factors contribute to a wide variation in client fees across and within jurisdictions.

In the period from 1987 to 2007, the largest change in the supply of pharmacotherapy treatment has come from the expansion of interventions provided through the private sector, rather than an expansion of publicly-funded treatment programs. That is, there has been a larger increase in persons receiving pharmacotherapy treatment from private medical practitioners than from publicly-funded treatment programs.

Nationally, the number of clients enrolled in public Methadone treatment increased from 2,701 in June 1987 to 6,541 in June 1994 and to 10,695 by 2007. Over the same period the numbers enrolled in private Methadone treatment increase from 1,745 to 8,449 in June 1994 and then to 25,153 in 2007 (Australian Institute of Health and Welfare, 2008; Commonwealth Department of Human Services and Health, 1995). The participation rates from 1987 to 2007 have increased from 28 to 116 per 100,000 for public settings, compared with 19 to 212 per 100,000 for private settings, most of the expansion being due to general practitioner and community pharmacy involvement.

The cost-benefit of pharmacotherapy treatment

Research has demonstrated that pharmacotherapy treatment is beneficial to society, cost-effective, and pays for itself in basic economic terms. Pharmacotherapy has been identified as the most cost effective treatment available in Australia for the management of opioid dependence (Department of Health and Ageing, 2004). Some key findings from the literature in relation to the cost-benefit of pharmacotherapy treatment include:

• A study undertaken by the National Drug and Alcohol Research Centre (NDARC) found a reduction in the cost of crime associated with enrolment in pharmacotherapy treatment consistent with the broader literature. In addition it found that every day an individual was enrolled in pharmacotherapy treatment, it paid for itself in terms of a decrease in jail and crime costs (Shanahan et al., 2007).

• A comprehensive examination of economic benefits and costs was performed on data from the Treatment Outcome Perspective Study. After examining the average cost of pharmacotherapy treatment per day, detailed measurements of criminal activities rates and the cost to society of various crimes, the study yielded a final benefit-to-cost ratio of 4 to 1 (Harwood et al., 1988).

• A cost effectiveness analysis undertaken by Barnett and Hui (2000) found that expanded access to pharmacotherapy treatment had an incremental cost-effectiveness ratio of less than $11,000 USD per quality-adjusted life year.

• Rufener et al. (1977), studied the cost-effectiveness of pharmacotherapy treatment (and other treatment modalities) and determined a benefit-to-cost ratio of 4.4 to 1.

• A study of the cost benefits of pharmacotherapy treatment showed that the costs to society of the criminal activities related to active heroin use can run as high as four times more than the costs for pharmacotherapy (Harwood et al., 1988).

11

FULL FINAL REPORT

• Researchers estimated the yearly cost to maintain an opioid addict in New York: untreated and on the street ($43,000 USD), in prison ($34,000 USD), in a residential drug-free program ($11,000 USD), and in pharmacotherapy treatment ($2,400 USD) (New York State Committee of Methadone Program Administrators, 1991).

Overall, the scientific evidence clearly supports the cost-effectiveness of pharmacotherapy treatment and the benefits to society.

The role of the community pharmacist in pharmacotherapy treatment

The community pharmacy sector has taken on a large and central role in the administration of pharmacotherapy treatment, and they do this for several reasons, including client care and community service and also for a fee-for-service. The inability of public clinic systems to provide the accessibility and hours of operation achieved by community pharmacies makes the involvement of this sector critical to the effective reach of treatment in Australia. In addition, community pharmacies are seen as a normal environment by pharmacotherapy clients for receiving their doses. There is no evidence of any alteration in the safety or the effectiveness of community pharmacy-based pharmacotherapy dosing and administration. In addition, the uptake by the community pharmacy sector has been impressive in Australia and internationally (Fleming et al., 2001; Matheson, Bond, & Tinelli, 2007; Sheridan et al., 2007).

A recent estimation of the number of clients who are on pharmacotherapies for opioid dependence (as of June 2008) is 41,347 people across all jurisdictions (Australian Institute of Health and Welfare). This is represented in Table 1 and broken down into their respective jurisdictions. Eighty-six percent (86.0%) of dosing points (ie where pharmacotherapy treatment is supplied to a client) are in the community pharmacy setting, whilst the remaining 14% are in public, state or private clinics.

Table 1: Number of clients and pharmacies receiving/providing pharmacotherapies for opioid dependence

Jurisdiction Number

pharmacotherapy clients

1

Number of clients receiving

pharmacotherapy from pharmacies

2

Number of pharmacies

3

Pharmacies supplying

pharmacotherapies4

New South Wales (NSW) 17,168 7,535 1733 558

Victoria (VIC) 11,821 11,106 1181 483

Queensland (QLD) 4,899 3,873 1011 353

Western Australia (WA) 2,908 2,608 515 233

South Australia (SA) 3,052 2,440 409 189

Tasmania (TAS) 588 544 137 48

Australian Capital Territory (ACT)

786 509 61 26

Northern Territory (NT) 125 97 28 10

Australia 41,347 28,712 5,075 1,900

1 Australian Institute of Health and Welfare, 2008

2 Australian Institute of Health and Welfare, 2008

3 Department of Health and Ageing, 2007-2008

4 Australian Institute of Health and Welfare, 2008

12

FULL FINAL REPORT

Disincentives for community pharmacies to participate in pharmacotherapy treatment

The disinclination of community pharmacists to be involved in pharmacotherapy treatment has been noted (Matheson, Bond, & Mollison, 1999). The notion of injecting drug use, plus the difficulties posed by some of these clients in terms of difficult interactions, can make this a hard group manage. More recently in Australia, research has shown that there is a high rate of bad debt amongst clients, with over 70.0% of pharmacies reporting that they provide credit to clients and one-quarter of the 407 community pharmacies surveyed indicating that they had clients in debt (A. Winstock, Lea, & Molan, 2008). Additional problems, associated with behavioural disinhibition and aggression among clients, shop-lifting and general dislike of the clientele, also reduce pharmacists’ enthusiasm for involvement in the pharmacotherapy program.

Recent research in NSW shows that the community pharmacy sector has the capacity to increase client numbers, however the factors identified that would encourage the pharmacists to take on additional clients included the stability of clients, the ability to return unstable clients to a clinic for dosing and increased financial return per client (A. Winstock et al., 2008).

Barriers and facilitators to treatment entry and retention

Based on a comprehensive review of the literature, (Ward et al., 1998) reported that the following factors were predictors of or were associated with, better outcomes in pharmacotherapy treatment:

• Higher daily doses of Methadone are positively associated with better outcomes and longer time in treatment. Higher buprenorphine doses are also likely to enhance retention and compliance (Mattick, Kimber et al., 2008).

• The provision of ancillary services (e.g. medical, psychological, financial services) is associated with increased retention in treatment and better outcomes (Ball & Ross, 1991).

• Accessibility of treatment is associated with better outcomes. Having a dosing point within reasonable travelling time is important; similarly, take-away Methadone doses have been associated with higher retention rates, but can lead to the diversion of Methadone.

• The imposition of treatment fees decreases retention and has a detrimental impact on outcomes (Ward et al., 1998).

Client fees for pharmacotherapy treatment

Among users entering treatment in Australia, only one in six clients (17%) listed a wage as their main source of income, most existing on social security payments (Ross et al., 2005). Generally, income data indicate that most illicit opioid users are poorly placed to pay any significant amounts towards the costs of their treatment.

A number of studies have shown that the reduction and elimination of fees for pharmacotherapy treatment increases retention and improves outcomes and that the introduction of fees may result in clients leaving the program and return to illicit drug use (Murphy & Rosenbaum, 1988; Rosenbaum, Irwin, & Murphy, 1988; Rosenbaum, UIrwin, & Murphy, 1987; Breen et al., 2003; Maddux et al., 1994). Only in the ACT is the client fee subsidised by the ACT Government, where $20 is paid by ACT Health directly to the pharmacy. Other States and Territories may provide limited funding, to some extent, to compensate the pharmacists; however they vary in their respective models. Often this funding is only for the pharmacist to incentivise supplying pharmacotherapy and none of this funding reduces the fee for the client.

Incentivising pharmacotherapy treatment

As mentioned above, a number of jurisdictions have differing funding models to try and incentivise pharmacies to deliver the service and overcome some of the negatives associated with offering the service in the community pharmacy.

13

FULL FINAL REPORT

A survey conducted in 2002 by the Pharmacy Guild of Australia (the Guild) found that the arrangement with the ACT Government was effective and of benefit to the clients and subsequently the pharmacy. The subsidy was found to assist the client’s finances and had an overall positive effect on their treatment. The funding was found to be beneficial for the pharmacies delivering the program by increasing their overall satisfaction in providing this service.

It was concluded from this survey that there was a need to develop “best practice funding options” for the supply of pharmacotherapies for opioid dependence in community pharmacies. This was investigated in a study commissioned by the Guild from Healthcare Management Advisors which was published in June 2007.

This report presented information on the efficiency, effectiveness and sustainability of the funding models and client outcomes. The study reported client improvement in the areas of health, wellbeing, social, treatment and economic outcomes, along with improved client satisfaction with the treatment. From a pharmacy perspective, the trial indicated improved service and economic outcomes, and improved relationships and communication between pharmacists/staff and clients. Although the study provided insight into these important investigative areas, it was hampered by the small sample size and short duration of the trial amongst other things; however it did show that further investment in investigating the funding models and client outcomes was warranted.

1.2. Rationale for the research

The differing funding models for pharmacotherapy treatment across the states and territories (including those with zero funding) lead to a lack of consistency in the:

• variable cost of treatment for the clients

• variable levels of incentives for the pharmacist.

While the efficacy of pharmacotherapy treatment is well demonstrated, accessibility and attractiveness to the target group and the ability to retain these clients in treatment to provide long-term care, depends on a number of factors. Affordability of the treatment is thought to be key in this regard.

The role of community pharmacy in providing access to pharmacotherapy has been significant and the potential to grow their role in this area has been identified. The Guild reported that 80% of consumers receiving treatment in public clinics have indicated that they would prefer to receive their treatment in community pharmacies

5.

There is currently no national system in place to support community pharmacies for their role in the treatment of this client group.

Given these factors, the Guild commissioned PricewaterhouseCoopers (PwC) and NDARC to undertake a research project to trial and evaluate the impact of a funding model as part of pharmacotherapy treatment for opioid dependence through community pharmacy.

5 The Pharmacy Guild of Australia, NSW Branch http://www.guild.org.au/nsw/content.asp?id=1081 (last accessed

28/03/08)

14

FULL FINAL REPORT

1.3. Objectives and hypotheses

The focus of the research project was to consider key elements of a national funding model for pharmacotherapy treatment in community pharmacy. Specifically, the study focused on two key objectives:

1. To evaluate the benefits of the implementation of a nationally consistent client subsidy and to consider the impacts of this arrangement on the following key outcomes:

- client retention rates in the program

- client compliance / adherence with doses

- client health and social outcomes

- the use of illicit opioids.

2. To consider the perspective of the community pharmacists who supply pharmacotherapy, within a nationally consistent subsidised environment, and to consider the impact of this arrangement in terms of:

- the pharmacists’ satisfaction with the trialled funding arrangements

- the way in which pharmacists deliver the service

- the costs associated with the delivery of the service.

1.4. Research project outline

The pharmacotherapy treatment for opioid dependence research project consisted of five stages:

• Stage 1: Project initiation – included liaison with the Guild, Advisory Panel and Reference Group and development of the literature review.

• Stage 2: Funding model selection and trial design – involved the development of the funding model, trial design, evaluation framework and trial tools in consultation with the Guild, Advisory Panel and Reference Group. In addition, ethics approval was obtained for the research.

• Stage 3: Trial implementation – the implementation of the trial included recruitment of pharmacies and clients, implementation of the intervention and the collection of data.

• Stage 4: Trial analysis – included the evaluation of the trial and identification of recommendations.

• Stage 5: Draft and final reporting - included the development of considerations for national implementation.

An outline of the research project is illustrated in Figure 1.

15

FULL FINAL REPORT

Figure 1: Research project outline

Trial design Ethics approval

Analysis of trial results

Initial consultations and literature review

Recruitment of pharmacies

Final report and key findings

Recruitment of clientsand treatment allocation (2:1)

Stage 1: Project initiation

(5 weeks)

Stage 2: Trial design

(3 months)

Stage 3: Trial implementation

(11 months)

Stage 4: Analyse trial results(2 months)

Stage 5: Final report

(2 months)

September 2008

October 2008

February 2009

January 2010

April 2010

Pharmacy level data collection• Pharmacy demographics• Client payment per dose

• Satisfaction (baseline & 6 month)• Activity based costing study

Client level data collection• Monthly dosing data (provided by

pharmacist)• Client survey (baseline & 6 month)

1.5. Purpose of the report

This is the Full Final Report for the National Funding Model for Pharmacotherapy Treatment for Opioid Dependence in the Community Pharmacy research project. The primary purpose of the Full Final Report is to present a comprehensive and detailed description of the approach and findings of the research project, along with all necessary academic details to substantiate the rigour of the project.

People who receive pharmacotherapy treatment are variously referred to as patients, clients or consumers. We have chosen to use the term ‘client’ throughout this report.

1.5.1. Roadmap to the report

The report comprises six sections:

• Section 1: provides the background to the project and rationale for the research, including the objectives and hypotheses of the study.

• Section 2: provides the approach to the design of the trial, including the committees and literature review undertaken to inform the development of the trial.

• Section 3: provides an overview of the trial design including the randomised controlled trial approach, the funding model and data collection.

• Section 4: provides the detailed description of the trial methodology, including the approach to sampling and recruitment, data collection and data analysis.

• Section 5: provides the key findings and results from the trial relating to the client.

16

FULL FINAL REPORT

• Section 6: provides the key findings and results from the trial relating to the pharmacy.

• Section 7: provides a discussion of the key findings from the trial along with the implications of the trial and recommendations.

Appendix A outlines the pharmacies who participated in the trial.

17

FULL FINAL REPORT

2. Approach to the design of the trial

Stage 1 and Stage 2 of the research project consisted of a number of key activities which focused on the development of the trial. To assist the research team in designing the trial, a number of methods were used to gather information and advice. These methods included:

• consultation with the Advisory Panel and Reference Group

• a comprehensive literature review.

Each of these methods is described in greater detail below.

2.1. Advisory Panel and Reference Committee for the research project

The recognition of the importance of stakeholder engagement was established at the beginning of the trial, which would involve regular communication with the Guild and the formation of an Advisory Panel and a Reference Group. Each of the two specialist groups was an essential part of the overall trial.

2.1.1. Advisory Panel

The purpose of the Advisory Panel was to provide a representative group of key stakeholders to oversee the design and implementation of the trial, endorse and guide the project and review key aspects of the project. Membership of the Advisory Panel comprised the Guild, the Department of Health and Ageing, the Pharmaceutical Society of Australia, the Intergovernmental Committee on Drugs, community pharmacists and consumer representative organisations. Terms of Reference were developed for Advisory Panel members and outlined their functions as:

• overseeing the project from all aspects of representation

• acting as an advisor to the project on behalf of the Guild

• reviewing aspects of the project e.g. methodology, data collection and final report as required.

Four meetings of the Advisory Panel were held on:

• 2 October 2008

• 6 April 2009

• 23 September 2009

• 25 March 2010.

Members of the Advisory Panel are detailed at the front of the report.

2.1.2. Reference Group

The Reference Group was established to convene a representative group of practising community pharmacists, who supply pharmacotherapy, to provide practical and operational advice for the project (members are outlined in Appendix B). The role of the role of the Reference Group was to:

• provide operational advice based on their knowledge and experience

• provide operational advice for recruitment and implementation of the trial in a community pharmacy setting

18

FULL FINAL REPORT

• act as local champions in each state to provide local advice, encourage participation and support to the participating pharmacists.

2.2. Literature review

The literature review was undertaken by Professor Richard Mattick (NDARC) to identify:

• trials undertaken in a similar environment, to help inform trial design

• alternative funding models for consideration

• evidence of the health and societal impacts of opioid treatment, particularly opioid treatment provided through pharmacies and through alternative delivery mechanisms, such as clinics

• evidence of the cost of provision of opioid treatment through pharmacies and alternative delivery mechanisms

• evidence of the impact on health and societal outcomes of alternative funding models.

The literature search strategy included examination of both peer reviewed journals and other published literature. The complete literature review is included in Appendix C.

19

FULL FINAL REPORT

3. Trial design

The trial design is a national randomised controlled trial (RCT). In devising a suitable trial design, a number of features relevant to the pharmacotherapy and pharmacy environment were considered. For example:

• Pharmacotherapy treatment is, for many, a long term treatment regime with some clients staying in treatment over a number of years. Consequently, the data collection period of six months was needed to enable a large majority of clients who were in treatment at the commencement of the trial to still be in treatment at the end of the trial.

• Over the trial period, there was likely to be some clients who finished treatment and others who commenced treatment. A cut-off date was needed to be agreed after which no clients would be recruited. Clients finishing treatment during the trial were asked to agree to the follow-up questionnaire at the end of the six-month trial.

• Clients may move treatment pharmacy during the trial for a variety of reasons. Clients who moved from an intervention pharmacy to a control group or a non-participating pharmacy will cease to have their payments subsidised but will be asked to provide follow up data. Clients who moved from a control group pharmacy to an intervention group pharmacy did not receive subsidy payments to prevent gaming.

Features of the trial design, discussed in further detail, include:

• development of the randomised controlled trial approach

• design of the funding model

• collection of data from pharmacies, including the activity based costing study

• collection of data from clients

• collection of data from jurisdictional health authorities

• timeframes for the research project

• ethics approval.

3.1. Development of the randomised controlled trial approach

RCTs are the methodology of choice in healthcare research as they provide the highest level of evidence for treatment efficacy; they are an invaluable support to the development of evidence based policy. RCTs involve the random allocation of participants in treatment to different intervention groups. The most important advantage of proper randomisation is that it eliminates selection bias, balancing both known and unknown prognostic (ie outcome influencing) factors, in the assignment of participants to different treatment groups. The methodology for achieving appropriate and statistically robust randomisation is well documented and was used in this study.

Randomisation to the intervention and control groups was based on pharmacy rather than the client. This design choice was based on the following considerations:

• the pharmacies needed to agree to changes in the payments they receive

• the method of recruiting clients was through pharmacies that had agreed to participate in the trial

• if randomisation were based on clients, there would have been the potential for clients who were randomised to the control group attending the same community pharmacy as those randomised to the intervention group, which could have raise tensions and issues of fairness

20

FULL FINAL REPORT

• both clients and pharmacies could have been randomised into intervention and control groups, however this would necessitate clients to move to a pharmacy matching their allocation; this would raise tensions and issues of fairness (given clients are often allocated to a community pharmacy for convenience near to their work or residence)

• if randomisation were based on clients, it would need to occur at the commencement of that client’s treatment at a community pharmacy which had agreed to participate in the trial, however, such a model would require a longer time to recruit a suitable sized sample than was available for this project.

Randomisation occurred at a 2:1 ratio, intervention: control. To achieve 80% power of detecting a difference between the intervention and control groups, a total sample size of around 700 clients was required to be recruited to detect a small difference (effect size = 0.25).

3.2. Design of the funding model for the trial

The trial tested a specified funding model nationally. The funding model was collaboratively developed and agreed by the Guild, the Advisory Panel and the research team at the outset of the trial. The funding model comprised two components, a client subsidy and a pharmacy incentive, as outlined in Table 2.

Table 2: Funding model for the trial

Intervention group Control group

Pharmacists • Pharmacists received an incentive of $15 per consented client per week:

- to subsidise the service

- for the provision of monthly client dosing data for the research.

• Pharmacists received an additional $15 per consented client per week:

- to subsidise the client fees by $15 per week

- pharmacists signed a declaration stating that the subsidy would be passed on to the client

• Pharmacists received an incentive of $50 per month for the provision of monthly data for the research.

Clients • Clients received a reduction of $15 per week in their payments.

• The trial aimed to maintain a co-payment focus; where the client was to pay a minimum of $5 per week.

• Clients received no reduction in their payments and received usual care.

Table 3 provides further detail pertaining to the business rules for the funding model.

21

FULL FINAL REPORT

Table 3 Business rules for the funding model

Type of payment

Intervention group: pharmacy incentive

Intervention group: client subsidy

Control group: pharmacy incentive

All participating pharmacies: pharmacy survey

All participating clients: baseline and follow up-client survey

Activity that generates the payment

Payment per intervention client per week for provision of service and intervention client dosing data

Payment per intervention client per week for client fee subsidy

Payment for the provision of control client dosing data

Pharmacists completes survey at baseline and end of trial

Consented clients complete telephone interview at baseline and end of trial

Value of payment

$15 per client per week

$15 per client per week

$50 per month $60 $35

Frequency and timing of payment

Six payments: on a monthly basis for the duration of the trial

Six payments: on a monthly basis for the duration of the trial

Six payments: on a monthly basis for the duration of the trial

Twice: baseline and end of trial

Twice: baseline and end of trial

Proof of service to generate payment

Client data to confirm invoiced amount for intervention group pharmacies

Client data to confirm invoiced amount for intervention group pharmacies

Client data to confirm invoiced amount for control group pharmacies

Completed paper based survey

Completed telephone interview

3.3. Collection of data from pharmacies

The collection of data from pharmacies included initial data on the characteristics of their practice and satisfaction with the funding arrangements for the pharmacotherapy program and the trial. Pharmacies were also asked to provide data to the trial on the attendance and compliance rates of the individual clients that were recruited into the trial through the provision of dosing data on a monthly basis.

The project team tracked receipt of the monthly data through a purpose built Access database which enabled monitoring and follow-up with pharmacies. The research team also developed a number of communication strategies to keep the pharmacies informed and engaged over the duration of the trial. This included:

• regular newsletters

• a trial specific 1800 phone number, in operation between 9:00am to 5:00pm Monday to Friday

• a trial specific fax number

• a trial specific email address.

Throughout the trial, pharmacists were asked to provide reasons for clients dropping out from the trial and the program, if known. This data was then substantiated by the data obtained from the health authorities outlined in Section 3.5. Further detail on the method of data collection from pharmacies is outlined in Section 4.5.

22

FULL FINAL REPORT

3.3.1. Activity based costing study

As part of the trial design, an activity based costing study was developed to cost the provision of a pharmacotherapy service in community pharmacies. The key component of the costing study was an observational study undertaken in a mix of 20 pharmacies across three states; this included the collection of data on the cost of consumables and infrastructure specifically related to the delivery of the program. Further detail on the activity based costing study methodology is outlined in Section 4.5.3.

3.4. Collection of data from clients

All clients (both intervention and control) were asked to provide information at the beginning of the trial and agree to participate in a six-month follow-up of the survey. The client survey was based on the Brief Treatment Outcome Measure – Concise (BTOM-C) and included a quality of life scale, financial strain and satisfaction questions. More detail on the client survey methodology can be found in Section 4.6.1.

3.4.1. Brief Treatment Outcome Measure – Concise

The BTOM-C is an outcome measure, incorporating the NSW Minimum Data Set for Alcohol and other Drug Treatment Services, developed for use across NSW for the consistent measurement of treatment outcomes for clients of Methadone and Buprenorphine treatment programs. The BTOM-C is intended to demonstrate the effectiveness of Methadone and Buprenorphine treatment by measuring the changes it makes to client’s lives. Further detail on the items collected in the BTOM-C is in Section 4.6.1.

3.4.2. Quality of life

A quality of life scale was included in the client survey to look at the impact of the trial on the clients overall quality of life. The World Health Organisation’s Quality of Life-8 scale (WHOQoL-8) was used and was designed for use where researches needed a short and concise quality of life instrument. The eight items were empirically derived from the WHOQoL-Brief.

3.4.3. Financial strain and satisfaction

Client satisfaction with the program and the trial was measured at baseline and the end of the trial. Based on advice from the Advisory Panel, additional financial strain questions were included in the six-month follow-up survey.

3.4.4. Administration of the client survey

Following consultation with the Guild and Advisory Panel, it was decided that the client survey would be administered via telephone. Telephone interviewing was the preferred method considering:

• the time and cost efficiency given the sample size and spread across Australia

• it is reliable and valid

• it is generally well accepted by subjects in studies of this nature.

Two possible threats to the use of telephone surveys were considered:

1. Different information may be obtained via the telephone as opposed to face-to-face – evidence in the literature indicates that the nature of the information collected does not change and agreement to participate is likely to be higher for telephone methods due to greater anonymity.

2. Greater loss to follow-up using telephone interview rather than with face to face methods – evidence suggests a very high follow-up success using a telephone survey method in a range of settings especially over a six-month period. In particular, NDARC reports a greater than 90% follow up rate using this method specifically with pharmacotherapy clients. Success with telephone interviews is based on:

23

FULL FINAL REPORT

- Assiduous collection of contact details when recruiting clients to the project – the trial design included the collection of multiple sources of contact methods e.g. phone numbers of the client and key family members (including mobile numbers), email address, postal address etc to facilitate follow-up.

- Payment for both interviews (baseline and follow-up) – the trial design included payments to all clients who completed the interview for their time and inconvenience via a $35 gift card.

Roy Morgan Research was engaged to administer the client survey. Roy Morgan is an established market research and public opinion survey company, specialising in all aspects of market research information gathering, including by telephone.

3.5. Collection of data from health authorities

A key outcome indicator for the trial was client retention in treatment. When a participating client was identified as no longer attending the participating community pharmacy, the reason for that discontinuation of treatment was an important outcome indicator. The trial sought to collect data to determine if the client had dropped out of treatment or changed treatment location.

Retention rates were refined to distinguish between the client remaining in treatment versus remaining in the trial. Retention can be defined in the following ways, where the client:

1. remains in treatment and the trial at the same pharmacy

2. remains in treatment and the trial but moves to a different pharmacy who is participating in the trial

3. remains in treatment but moves to a non-participating provider and therefore is lost to the trial

4. can not be located and is assumed to have dropped out of treatment and therefore is lost to the trial and to treatment.

The trial design included the collection of data from the relevant jurisdictional health authority in each state and territory. This data aimed to understand client movements from participating pharmacies and was to be collected for clients moving during the trial and those lost to follow-up. The data collection aimed to both verify the reasons for change provided by the pharmacist and determine the overall number of clients retained in treatment.

3.6. Timeframes for the research project

The research project commenced in August 2008 and concluded in April 2010. The trial was implemented over a period of eleven months from February 2009 to January 2010, with the intervention implemented over a period of six months, reflecting the staggered commencement in the trial of three cohorts.

24

FULL FINAL REPORT

Figure 2: Timeline for key research project activities

Stage 2 Stage 3 Stage 5Stage 4Stage 1

Draft reportDraft report

1 ½ months1 ½ months 3 months3 months 11 months11 months 2 months2 months 2 months2 months

Oct 2008

Feb2009

Jan 2010

Feb 2010

April 2010

1 Sep 2008

Literature ReviewLiterature Review

Start-up meeting

Start-up meeting

Activity 1

Activity 2

Activity 12

Final reportFinal report

Activity 13

Collect baseline dataCollect baseline data

Staggered recruitment of pharmacies and

individuals

Staggered recruitment of pharmacies and

individuals

Activity 7

Activity 8

Trial implementationTrial implementation

Activity 9

Trial evaluationTrial evaluation

Activity 11

Ethics committee approval

Ethics committee approval

Design trial tools &

infrastructure

Design trial tools &

infrastructure

Activity 5

Activity 6

Select funding model for trial

Select funding model for trial

Select trial design & evaluation framework

Select trial design & evaluation framework

Activity 3

Activity 4

Collection of post-trial data

Collection of post-trial data

Activity 10

Draft report and final report

Trial analysisTrial implementationFunding model selection

and trial designProject

initiation

3.7. Ethics approval

The national funding model for pharmacotherapy treatment for opioid dependence in the community pharmacy research project was conducted with ethics approval from the University of New South Wales Human Research Ethics Committee, granted on 10 November 2008 (HREC: 08268).

25

FULL FINAL REPORT

4. Trial methodology

The specific objectives of the research project were to develop, implement and evaluate a funding model for opioid dependence in community pharmacy, which was trialled nationally.

This section describes the methodology for the trial including:

• overview of key trial activities and timelines (Section 4.1)

• development of the sampling framework (Section 4.2)

• the recruitment process for pharmacies and clients (Section 4.3)

• allocation to the intervention and control group (Section 4.4)

• pharmacy and client data collection (Sections 4.5 and 4.6)

• data management processes (Section 4.7).

4.1. Overview of trial activities and timelines

Figure 3 provides an overview of the timing of the key activities undertaken during the eleven-month trial period, each of which is described in further detail throughout this section.

Figure 3: Timeline for key trial activities

Jan 09 Feb 09 March 09 April 09 May 09 June 09 July 09 Aug 09 Sept 09 Oct 09 Nov 09 Dec 09

Key trial activities

Pharmacy recruitment and baseline data collection

Pharmacy selection

Client recruitment

Pharmacy allocation to control and intervention

groups

Baseline client survey

Monthly dosing data

Activity based costing study

Pharmacy commencement in

trial by cohort

Six-month client survey

Pharmacy satisfaction

survey

Jurisdictional drop out

data

4.2. Sampling framework

The aim of the sampling framework was to identify a nationally representative sample of pharmacies and clients across all jurisdictions (except the ACT) and Pharmacy Access/Remoteness Index of Australia (PhARIA) categories. The ACT was excluded from the trial given that they have a funding model in place which subsidises the client fee.

26

FULL FINAL REPORT

To achieve approximately 80% power of detecting a small to moderate difference (effect size =0.25) between groups (intervention and control) a total sample size of 700 clients, with complete follow-up data, was required.

The PhARIA index provides the specific measurement of remoteness of any pharmacy within Australia. PhARIA was developed by the National Centre for Social Applications of Geographic Information Systems. PhARIA was used to minimise sampling bias from pharmacy location and to ensure the inclusion of pharmacies from rural and remote locations. The PhARIA index incorporates a ‘professional isolation’ component which measures the road distance to the nearest five pharmacies. There are six categories under the definitions of the index:

1. PhARIA 1 – Highly Accessible

2. PhARIA 2 – Accessible (Group A)

3. PhARIA 3 – Accessible (Group B)

4. PhARIA 4 – Moderately Accessible

5. PhARIA 5 – Remote

6. PhARIA 6 – Very Remote.

To ensure complete representation across all jurisdictions and PhARIA categories, a comprehensive sampling framework was developed. The first step of developing the sampling framework was to understand the total number of pharmacies who supply pharmacotherapy treatment and the number of clients receiving treatment, by jurisdiction (Table 4).

Table 4: Number of pharmacies and clients registered for pharmacotherapy treatment

State Number of pharmacies (≥1 client)6 Number of registered clients receiving

treatment in community pharmacies7

NSW 490 6,626

VIC 403 11,018

QLD 356 5,517

WA 215 Unknown8

SA9 77 1,383

TAS 46 532

NT 10 102

Total 1,597 25,178

6 Data based on jurisdictional data obtained during the trial period.

7 Data based on jurisdictional data obtained during the trial period.

8 WA data could not be accurately broken down into their respective categories as WA Health chose not to release

these figures. Instead WA Health asked all community pharmacies supplying pharmacotherapy for approval to have their data passed onto the researchers and to indicate their interest in participation in the trial.

9 SA data was based on the number of pharmacies that consented to have their details provided to us, not a

representation of the total number of pharmacies (172) which supply pharmacotherapies.

27

FULL FINAL REPORT

The breakdown of pharmacotherapy registered pharmacies by PhARIA category was obtained by matching up the locality of the pharmacies in Table 4 with the localities in the 2008-2009 PhARIA index, Table 5.

Table 5: Pharmacy location broken down by jurisdiction and PhARIA category

State PhARIA 1 PhARIA 2 PhARIA 3 PhARIA 4 PhARIA 5 PhARIA 6

NSW 491 39 45 26 11 5

VIC 328 24 37 11 2 1

QLD 260 26 30 10 17 7

WA Unknown Unknown Unknown Unknown Unknown Unknown

SA 57 2 10 6 0 2

TAS 25 9 9 1 2 0

NT 8 0 1 0 1 0

Total 1169 100 132 54 33 15

Utilising the information in Table 4 and Table 5, a sample of 81 pharmacies across jurisdictions was identified as providing a sufficient pool of clients from which to recruit the required sample of 700 clients (Table 6). The distribution of pharmacies selected was based on the distribution of registered pharmacies in each jurisdiction.

Table 6: Number of pharmacies assigned for the sampling framework

State Number of pharmacies

(≥1 client)

Representative percentage of total pharmacotherapy

pharmacies

Total number of pharmacies selected for sampling

framework

NSW 490 31% 25

VIC 468 25% 20

QLD 356 22% 18

WA 215 13% 11

SA 77 5% 4

TAS 46 3% 2

NT 10 1% 1

Total 1,97 100% 81

The only exclusion criterion for pharmacies to participate was based on the size of their client group. Those pharmacies with more than 50 clients participating in the pharmacotherapy program were excluded from participating in order to prevent any single pharmacy from contributing too great a proportion of the overall sample.

In order to avoid creating tension between clients in a single pharmacy, it was deemed advisable that all pharmacotherapy clients in a pharmacy should be invited to participate.

Pharmacies were further broken down into two bands, those with one to 20 clients and those with 21 to 50 clients, by PhARIA category for each jurisdiction. The final sampling framework is described in Table 7, and yielded a pool of over 900 registered clients from which to recruit the study sample of 700, based on the recruitment of a nationally representative sample of pharmacies.

28

FULL FINAL REPORT

Table 7: Sampling framework

NSW VIC QLD WA SA TAS NT Total

Pharm

acie

s

Clie

nts

Pharm

acie

s

Clie

nts

Pharm

acie

s

Clie

nts

Pharm

acie

s

Clie

nts

Pharm

acie

s

Clie

nts

Pharm

acie

s

Clie

nts

Pharm

acie

s

Clie

nts

Pharm

acie

s

Clie

nts

PhARIA 1 16 232 13 258 9 118 9 125 3 20 1 11 1 6 52 770

PhARIA 2 2 18 2 16 3 14 0 0 0 0 1 7 0 0 8 55

PhARIA 3 4 33 4 19 3 12 2 26 1 5 0 0 0 0 14 95

PhARIA 4 2 5 1 5 1 4 0 0 0 0 0 0 0 0 4 14

PhARIA 5 1 2 0 0 1 5 0 0 0 0 0 0 0 0 2 7

PhARIA 6 0 0 0 0 1 2 0 0 0 0 0 0 0 0 1 2

Pharmacies <20 clients

21 158 16 164 16 91 9 83 4 25 2 18 1 6 69 545

Pharmacies >20 clients

4 132 4 134 2 64 2 68 0 0 0 0 0 0 12 398

Total 25 290 20 298 18 155 11 151 4 25 2 18 1 6 81 943

4.3. Recruitment process

The aim of the recruitment process was to have a nationally representative sample of pharmacies and clients across all jurisdictions (except the ACT) and PhARIA categories.

4.3.1. Pharmacy recruitment

The sampling framework (described in Table 7) formed the basis for selecting and recruiting pharmacies for the trial. The steps for recruiting the target sample of 81 pharmacies were:

1. obtaining expressions of interest

2. selection of pharmacies to meet the sampling framework

3. sending a ‘consent bundle’ to pharmacies.

Further detail on each of these steps is described below.

Step 1: Expressions of interest

Pharmacists were recruited via an expression of interest (EOI) process. The EOI was sent to all pharmacies currently supplying pharmacotherapy treatment to invite them to participate in the trial. Addresses of the pharmacies were obtained from each relevant jurisdictional department managing the registration of participating pharmacies, with the exception of WA and SA. WA Health and SA Health elected to directly contact all registered pharmacies supplying pharmacotherapy treatment in their respective jurisdictions, on behalf of the research team.

29

FULL FINAL REPORT

For all other jurisdictions, the project team mailed the EOI. The mail-out was staggered over a number of cohorts broken down by jurisdiction:

• Cohort 1: NSW

• Cohort 2: QLD, SA and NT

• Cohort 3: VIC and TAS.

The EOI cut-off was approximately three weeks after expressions of interest were mailed out, response rates to the EOI are listed in Table 8.

Table 8: Response rate from EOI

State Total number of eligible

pharmacies (with clients ≥1)10

Total number of

responses to the EOI11

EOI response rate (%)

NSW 490 81 17%

WA 215 51 24%

QLD 356 49 14%

SA12

77 14 18%

NT 10 1 10%

VIC 403 75 19%

TAS 46 11 24%

Total 1597 282 18%

Step 2: Pharmacy selection

From the EOI, pharmacies were randomly selected. That is, all pharmacies responding to the EOI were placed in alphabetical order in Excel, broken down by jurisdiction, client band and PhARIA category. The Excel function RAND() was then used to select the pharmacies from the list in such a way as to not allow the same pharmacy to be chosen twice. That is, once a pharmacy had been selected it was removed from the list and the total number of pharmacies in the RAND() function was reduced by one until the sampling framework was filled, as described in Table 7.

If a gap existed in meeting the sampling framework, the project team called pharmacies directly to gauge their interest. Limited calling was required as the EOI response was adequate in almost all cells of the sampling framework. When direct calling was initiated, it was undertaken based on random selection from the list of the pharmacies fitting the characteristics of the gap in the sampling framework. That is, the names of all the pharmacies eligible to participate, which met the sampling framework, were placed in alphabetical order in Excel and a random number generator was created to place the pharmacies in a call list.

10

Data obtained from the jurisdictions.

11 Total number of EOI responses was recorded up to the recruitment cut-off point, pharmacies registering their interest

after this time period were not recorded and therefore the numbers reported are a slightly under the total number of pharmacies that ended up expressing interest.

12 Data was based on the number of pharmacies that consented to have their details provided to us, not a

representation of the total number of pharmacies (172) which supply pharmacotherapy, it was assumed that 77 was the total number of pharmacies supplying pharmacotherapy in SA for the purposes of the trial.

30

FULL FINAL REPORT

Step 3: Consent bundle

Once the 81 pharmacies were selected, a pharmacy ‘consent bundle’ was sent. The consent bundle included the following (see Appendix D):

• Letter of invitation: informed pharmacies of their selection to the trial.

• Next steps information sheet: informed pharmacies of the actions they were required to undertake prior to client recruitment.

• Pharmacist frequently asked questions: provided answers to common questions regarding the trial.

• Information statement and consent form: outlined the key aspects of the trial which the pharmacists were consenting to along with the consent form. Pharmacists were provided with an additional copy for their records.

• Pharmacy information form: collected the core information on pharmacies required for the trial.

• Pharmacy satisfaction survey: collected information on pharmacist’s satisfaction with the current funding arrangements for the pharmacotherapy program.

A total of 79 pharmacies were recruited to the trial, of which 74 actively participated (ie consented clients and/or provided data to the trial).

4.3.2. Client recruitment

As part of the trial design, it was recommended that pharmacists approach all of their eligible clients on the pharmacotherapy program to participate in the trial. To be considered eligible for the trial, a client needed to have met the following inclusion criteria:

• be prescribed Methadone, Subutex or Suboxone which was supplied by a participating community pharmacy

• be aged 18 years and over.

Exclusion criteria for participation in the trial included those clients who did not speak English well enough to communicate with the pharmacist or sign the consent form or they did not complete and sign the client consent form.

A client ‘consent bundle’ was supplied to the pharmacist. The purpose of this bundle was to provide the pharmacist with the tools and documents needed to recruit their clients to the trial. It included the following documents (see Appendix E):

• Client recruitment and next steps sheet: outlined to the pharmacist the process for client recruitment, participation in the trial and the documents in the bundle.

• Trial advert: for the pharmacist to display in the pharmacy to signal their participation in the trial.

• How to gain client consent: outlined the process to follow to gain informed consent form the clients.

• Client frequently asked questions: provided answers to common questions regarding the trial.

• Client information and consent form: outlined the information clients needed to know regarding the trial to provide informed consent in addition to the consent form. Clients were provided with an additional copy for their records.

• Client locator form: documented the clients contact details.

31

FULL FINAL REPORT

• Payment information sheet: outlined the business payment rules which governed the trial payment structure.

In addition, pharmacies were also provided with:

• Trial invoices and invoicing guide: provided advice on the invoicing process for the trial and invoicing templates.

• Subsidiary pharmacotherapy treatment registers: provided for pharmacists to record the dosing data of consented clients throughout the trial.

• CD-ROM: contained electronic versions of the above-mentioned data collection tools.

4.4. Allocation to intervention and control groups

Based on the trial design, pharmacies were randomly allocated to either the control group or the intervention group based on the process described below. Randomisation was managed by PwC and NDARC staff who were not directly involved in the research project to maintain independence in the process. Pharmacies were collated into the following four categories, as outlined in Table 9.

Table 9: Categories for RCT allocation

Category Category description Number of pharmacies in category

1 PhARIA 1, with one to 20 clients 48

2 PhARIA 1, with 21 to 50 clients 18

3 PhARIA 2 to 6, with one to 20 clients 36

4 PhARIA 2 to 6, with 21 to 50 clients 0

Pharmacies could only be allocated into three of the four categories, as there was no representation from pharmacies with 21 to 50 clients in PhARIA 2 to 6. Randomisation into the control (“usual care”) or intervention group was set up using these broad categories in a computer-generated random number allocation program (“Ralloc” in STATA).

STATA was the most appropriate software to use in this situation given its in-built options specific to epidemiological studies and clinical trials. Ralloc provides a sequence of treatments randomly permuted in blocks, in which size and order are also random; to prevent the prediction of future treatment allocation based on past allocation patterns. In this study, the number of subjects to be potentially randomised was set at 102 according to the numbers in the brackets above to allow for extra recruitment if needed over and above the 81 pharmacies proposed in the sampling framework as described in Table 7.

Block sizes of three and six were chosen as variables and the number of each block size was randomised to improve concealment. A stratified randomisation was undertaken based on rural remoteness of the participating pharmacies (PhARIA) and balance was also sought in allocation between the numbers of potential client participants in each pharmacy. The randomisation list was created by NDARC personnel and PwC custodians were appointed to manage the randomisation process, none of whom were associated with the study.

A set of rules were drafted up for the research team to ensure the randomisation list was kept confidential. Two custodians outside of the research team managed the randomisation of pharmacies into the two groups using the list compiled from STATA.

A protocol for the custodians was drafted for reliability and integrity of the randomisation. Additional information on the randomisation process can be found in Appendix F. For the participating 74 pharmacies, Table 10 outlines the allocation into the intervention and control groups; where a ratio of one control to two intervention group pharmacies was achieved in all categories and overall.

32

FULL FINAL REPORT

Table 10: Pharmacy allocation into control and intervention groups

PhARIA 1

<21 clients

Control

PhARIA 1

<21 clients

Intervention

PhARIA 1

>20 clients

Control

PhARIA 1

>20 clients

Intervention

PhARIA 2-6

<21 clients

Control

PhARIA 2-6

<21 clients

Intervention

Total control group

Total intervention

group

NSW 1 10 2 2 2 7 5 19

VIC 4 4 0 4 4 1 8 9

QLD 4 3 1 1 2 7 7 11

WA 1 6 1 0 0 1 2 7

SA 2 0 0 0 0 1 2 1

TAS 0 1 0 0 1 0 1 1

NT 0 0 0 1 0 0 0 1

Total 12 24 4 8 9 17 25 49

4.5. Pharmacy data collection

Data collection at the pharmacy level consisted of the following, each of which is described in further detail below:

Type of data collection Timing Section

Pharmacy information form Baseline 4.5.1

Pharmacy satisfaction survey Baseline, end of trial 4.5.2

Activity based costing study Mid trial 4.5.3

4.5.1. Pharmacy information form

The pharmacy information form (Appendix D) collected core information on the pharmacies at baseline and included:

• pharmacy and pharmacist contact details

• type of pharmacy (eg banner, independent etc)

• pharmacy location

• number of clients the pharmacy is authorised to supply to

• number of clients on the pharmacotherapy program

• daily fee charged to clients for Methadone, Subutex and Suboxone.

The response rate for the pharmacy information form was 100%.

33

FULL FINAL REPORT

4.5.2. Pharmacy satisfaction survey

The pharmacy satisfaction survey (Appendix D) was designed to ascertain the pharmacist’s satisfaction with the pharmacotherapy program and the trial. The survey was administered at baseline and at the end of the trial. The survey measured pharmacist satisfaction with the following:

• financial arrangements for the supply of pharmacotherapy treatment

• client contribution / co-payment

• interactions and relationship with clients

• trial subsidy and pharmacy incentive fee

• participation in the trial.

Response rate for the pharmacy satisfaction survey was 100%.

4.5.3. Activity based costing study

As part of the trial, an activity based costing study was undertaken. The aim of this data collection was to obtain a comprehensive snapshot of the the way in which pharmacists deliver the service and the costs associated with the delivery of the service. There were two components to the study, further detail on each of these are outlined below:

1. An observational study in a mix of 20 pharmacies in NSW, QLD and VIC.

2. Data on the cost of consumables and investments specific to the program.

Observational study

The observational component of the study was conducted in the pharmacy during, and around, peak dispensing times over a three day period, for approximately three hours per day. A minimum of three days for the observation study was considered necessary to collect valid and reliable data on the time spent by pharmacists and pharmacy staff on pharmacotherapy treatment activities.

A research team member observed every pharmacotherapy treatment client who presented for their dose during the observation period in the pharmacy. No data collected in the study could be attributed to any individual.

The researcher recorded the time for both the activities directly associated with each client who attended for dosing and those activities not directly associated with an individual client, but relevant to the overall delivery of the pharmacotherapy program. The key aactivities associated with delivering the pharmacotherapy program were defined as:

• preparation of doses

• pre-screening the client

• recording of the doses

• dosing the client

• processing payments

• counselling / talking with the client

• administration activities

• other activities eg inducting new clients, communication with other providers etc

34

FULL FINAL REPORT

The study also collected data regarding variables that may have an impact on the time taken and resources used to deliver the service, including:

• workload of the pharmacy

• time of day

• day of the week

• experience of the pharmacist

• type of pharmacotherapy treatment

• type of dose (eg take away vs. daily dose)

• client variables (eg age, time in treatment)

• characteristics of the pharmacy / layout

• location of the pharmacy eg strip or shopping centre

• investment in the delivery of the program eg security, equipment, software, consumables.

An observational data collection tool was developed to ensure that the researchers were recording information in a consistent manner. The tool was piloted and refined based on observations in two volunteer pharmacies who were not participating in the trial.

Pharmacies were selected for the observation study based on a random sample of 20 pharmacies already consented to the trial from the east coast states of Australia (NSW, VIC and QLD). The aim was to represent both PhARIA and client numbers per pharmacy in a similar fashion to what was represented in the overall trial. It was decided there would not be any additional value in including the ACT in the study, given that previous research has shown that apart from the subsidy scheme, the rest of the service in the ACT is comparable to the other states. The sampling framework for the observational study is detailed in Table 11. Pharmacies were randomly selected to fill the sample; a convenience sample was then selected based on location from a metropolitan centre.

Table 11: Sampling framework for the observational study

NSW VIC QLD Total

PhARIA 1 6 5 4 15

PhARIA 2 0 1 1 2

PhARIA 3 2 0 1 3

PhARIA 4 0 0 0 0

PhARIA 5 0 0 0 0

PhARIA 6 0 0 0 0

Total 8 6 6 20

Pharmacies with <20 clients

7 4 5 16

Pharmacies with >20 Clients

1 2 1 4

35

FULL FINAL REPORT

The researchers also mapped the processes they observed for supplying and handling pharmacotherapies to better understand the way in which the service is delivered in the community pharmacy.

Costing data

Costing data was based on time spent delivering the defined activities (from the observational study), salary costs, along with the cost of consumables and infrastructure specific to the delivery of the program. Salary costs were based on the average base hourly rates, as outlined in the Community and Hospital Pharmacists’ Remuneration Survey Report (Association of Professional Engineers, Scientists & Managers, Australia, 2009).

4.6. Client data collection

Data collection at the client level consisted of the following, each of which is described in further detail below:

Type of data collection Timing Section

Client survey Baseline, end of trial 4.6.1

Client dosing data Monthly 4.6.2

Client drop-out data At point of drop-out, end of trial 4.6.3

4.6.1. Client survey

As described in Section 3.4, the client survey was made up of four components: the BTOM-C; a quality of life scale (WHOQoL-8); financial strain questions; and satisfaction questions. The survey was administered at baseline and at the end of the trial to clients via a telephone survey, which was undertaken by Roy Morgan Research. Table 12 provides a description of the four components of the client survey; the BTOM-C and WHOQoL-8 can be found in (Appendix G).

Table 12: Description of the components in the client survey

Survey component Type of information measured

BTOM-C • Client demographics

• Information on the clients recent drug and alcohol use and risk behaviours

• Information on the social aspects of the client’s life and their perception of their health

• Information about the client’s current and previous drug and alcohol treatment

WHOQoL-8 • How the clients feels about their quality of life, health, relationships, financial situation and other areas of life

Satisfaction questions • The client’s satisfaction with the costs associated with the program and the program in general.

Financial strain questions • The financial impact (if any) on the lives of the client, resulting from the fees associated with participating in the pharmacotherapy program

36

FULL FINAL REPORT

Figure 15 provides an overview of the response rates for the baseline and six-month client interviews.

Table 13: Response rate and statistics from client interview

Baseline Six-month follow-up

Total completed interviews 626 485

Total refusals 7 1

Total unreachable / no longer on program 7 31

Average interview length 21.81 minutes 16.81 minutes

4.6.2. Client dosing data

Pharmacies were required to provide client dosing data on a monthly basis and all dosing data included the following information:

• client name

• number of days in a month (some months were defined as a five week period)

• whether the client was supplied regular dose, takeaway dose, double/triple dose or missed their dose

• dosage strength.

The purpose of collecting the clients dosing information was to ascertain the number of missed doses in a given month to measure the client’s retention and compliance to their treatment. Other information, such as medication changes and monitoring any changes in their dosage over the six-month period, could also be identified. As part of the business payment rules it was important to establish that the client dosed at least once in any given week during the trial period to ensure eligibility for payment for an intervention pharmacy.

4.6.3. Client drop-out data

Clients were able to be identified as having dropped-out from the trial from the monthly dosing data. There were three methods utilised for following up client drop-outs:

1. Pharmacies – when a client was identified as being a drop-out from the trial, pharmacists were contacted to verify this was not missing data and were asked to provide a reason for trial drop-out, if known.

2. Clients – those clients who participated in the end of trial survey were asked whether they are still on the program. If they were no longer on the program, information was collected as to the reason why.

3. Jurisdictional health authorities – for those clients who were identified as a drop-out from the trial, data was requested from the jurisdictional health authorities to identify whether the client had moved pharmacy or had left the program.

4.7. Data management processes

4.7.1. Access database

Due to the large number of forms received from each pharmacy and client, there was a need for an efficient and confidential data management system to be constructed. A user restricted database was built so all trial specific information could be stored in a secure and coherent manner for analysis.

The database was developed using Microsoft Access, and using split database architecture, ie with a front end user interface in Access and all back end datasets stored in a Microsoft Structured Query Language

37

FULL FINAL REPORT

(SQL) Server. This approach enabled multiple users to enter data at the same time and improved data security. Using the functionality within MS Access and Access Visual Basic, a series of Access data entry forms were created based on the data collection forms. The following principles were followed in designing the database:

• Data validations were applied to ensure the user could only enter certain information (eg Yes/No dropdowns containing “Yes” and “No”, but no other variation). This approach ensured the data was fit for analysis.

• A series of business rules were developed to minimise errors in data entry (for example pharmacy start dates on the trial could not be after a certain date, the same form could not be entered more than once).

• The database was engineered to produce a series of automated reports, which enabled the research team to keep track of how many hard copy forms had been received, how many pharmacies and clients were in the trial and how many had dropped out.

• The database had built in safety components to give the user the option to cancel or save information entered. The architecture was built in such a way that the user could drill down to each data entry forms for each pharmacy by selecting from a dropdown menu.

• The datasets containing information entered were created and stored using Microsoft SQL Server. These tables corresponded to the level of information entered – for example some information was collected at the pharmacy level, with single records for each pharmacy demographic. So a “Pharmacy” dataset was created. Other information was collected on a client level, feeding into a “client” dataset. These datasets could then be remerged later for the purpose of analysis.

4.7.2. Data quality

Data quality assurance checks were conducted, which involved comparison between hard copy data received and the data entered into the database. Percentages of data checked were:

• Monthly patient dosing data - 20% (1 in 5 patients)

• Pharmacy surveys - 20% (1 in 5 pharmacies)

• Data from activity based costing observational study - 50% (1 in 2 pharmacies).

4.8. Data analysis design

All analyses were undertaken using Statistical Analysis Software (SAS) Version 9.1 for windows, with descriptives compiled for all variables. The following analyses were used:

• Baseline and six-month values were compared using paired t-tests.

• Client retention to the trial was analysed using a Kaplan-Meier survival estimate and a log-rank test.

• A Generalised Estimating Equation (GEE) model was used to analyse drop-out by month and compliance of doses, taking into account repeated measures.

• Multivariate analysis was undertaken to test the association between time to drop-out and client / pharmacy characteristics using the Cox proportional hazard regression model, which were confirmed by a sensitivity analysis which adjusted for pharmacy clustering.

38

FULL FINAL REPORT

5. Trial results – Client

This section presents the trial results for the client, including:

• Client recruitment and retention (see Section 5.1)

• Client demographics (see Section 5.2)

• Client outcomes:

- Retention in treatment (see Section 5.3)

- Compliance with treatment (see Section 5.4)

- Drug and alcohol use and related behaviours (see Section 5.5)

- Health and social outcomes (see Section 5.6)

• Client financial strain (see Section 5.7)

• Client satisfaction (see Section 5.8)

• Costs of the pharmacotherapy program to clients (see Section 5.9)

• Summary of results (see Section 5.10).

5.1. Client recruitment and retention

A total of 749 clients were recruited to the trial, of which 259 were allocated to the control group and 490 to the intervention group. A total of 639 clients remained in the trial at six-months. A total of 110 clients had dropped-out of the trial at six-months; 44 (18.3%) were control clients and 66 (14.0%) were intervention clients. Of the 110 clients who dropped out, follow-up data was obtained on 83 and it was identified that:

• forty-nine clients had transferred to another pharmacy and remained in treatment

• fifteen clients had finished on the program

• eleven clients had been incarcerated

• six clients had been banned from the pharmacy

• two clients had died.

This indicates that of the 749 clients who started in the trial, 703 (93.9%) remained in treatment at the end of the trial. Section 5.3 provides further detail on client retention.

It was identified that 37 clients had insufficient data at baseline to be included in the analysis of trial data. This resulted in 712 clients being included in the trial for analysis purposes (240 in the control group and 472 in the intervention group). Figure 4 illustrates the process for recruitment and the follow-up of clients during each stage of the trial.

39

FULL FINAL REPORT

Figure 4: Process for the recruitment and follow-up of clients

749 clients consented

712 clients (for analysis purposes)

231 control clients month 1

463 intervention clients month 1

448 intervention clients month 2

438 intervention clients month 3

430 intervention clients month 4

417 intervention clients month 5

406 intervention clients month 6

226 control clients month 2

214 control clients month 3

207 control clients month 4

201 control clients month 5

196 control clients month 6

37 clients had insufficient baseline data to be included in trial analysis

259 clients in control group

490 clients in intervention group

15 clients dropped out5 clients dropped out

10 clients dropped out

8 clients dropped out

13 clients dropped out

11 clients dropped out

12 clients dropped out

7 clients dropped out

6 clients dropped out

5 clients dropped out

240 control clients 472 intervention clients

9 clients dropped out9 clients dropped out

5.2. Client demographics

A summary of the demographic characteristics of the clients are presented in Table 14. Of the participating clients, 59.3% were male and the large majority were born in Australia (85.88%) with English as their preferred language (98.34%). Most clients were aged between 30-50 years in age (71.3%), with the mean age of the control group being 39 and the intervention group 40.

For approximately half of the cohort (52.5%), their main source of income was a pension, while 15.1% had full-time employment and 6.5% had part-time employment. Only 1.5% of clients had no source of income. Approximately 40.3% of all clients earned between $250-$399 per week and 20.6% earned between $150-$249 per week. The usual place of residence for a large proportion of clients was in a rented house or flat, public or private (70.9%), 22.1% lived in a privately owned house or flat and 1.3% had no usual residence or were homeless.

Overall, the baseline characteristics of the control and intervention groups were well matched. There were no differences in the characteristics of the two groups.

40

FULL FINAL REPORT

Table 14: Demographic characteristics of clients

Variable description All clients

N (%)

Control group

N (%)

Intervention group

N (%)

Gender Male

Female

Unspecified

357

244

1

(59.3%)

(40.5%)

(0.2%)

122

91

0

(57.3%)

(42.7%)

(0.0%)

235

153

1

(60.4%)

(39.3%)

(0.3%)

Age 18 to 25 years

25 to <30 years

30 to <40 years

40 to <50 years

50 to <60 years

60 to <70 years

14

73

236

193

82

4

(2.3%)

(12.1%)

(39.2%)

(32.1%)

(13.6%)

(0.7%)

7

28

86

65

27

0

(3.3%)

(13.1%)

(40.4%)

(30.5%)

(12.7%)

(0.0%)

7

45

150

128

55

4

(1.8%)

(11.6%)

(38.6%)

(32.9%)

(14.1%)

(1.0%)

Country of birth Australia

Other

517

85

(85.88%)

(14.12%)

187

26

(87.79%)

(12.21%)

330

59

(84.83%)

(15.17%)

Preferred language English

Other

592

10

(98.34%)

(1.66%)

212

1

(99.53%)

(0.47%)

380

9

(97.69%)

(2.31%)

Main source of income Full-time employment

Part-time employment

Temporary benefit (eg sickness, unemployed)

Pension (eg aged, disability)

Student allowance

Dependant on others

Retirement fund

No income

Other

91

39

136

316

1

5

1

9

4

(15.1%)

(6.5%)

(22.6%)

(52.5%)

(0.2%)

(0.8%)

(0.2%)

(1.5%)

(0.7%)

34

15

44

115

0

2

0

2

1

(16.0%)

(7.0%)

(20.7%)

(54.0%)

(0.0%)

(0.9%)

(0.0%)

(0.9%)

(0.5%)

57

24

92

201

1

3

1

7

3

(14.7%)

(6.2%)

(23.7%)

(51.7%)

(0.3%)

(0.8%)

(0.3%)

(1.8%)

(0.8%)

Usual place of residence Rented house or flat (public or private)

Privately owned house or flat

Boarding house

Hostel / supported accommodation services

Psychiatric home / hospital

Shelter / refuge

Caravan on serviced site

No usual residence / homeless

Other

427

133

11

14

1

1

2

8

5

(70.9%)

(22.1%)

(1.8%)

(2.3%)

(0.2%)

(0.2%)

(0.3%)

(1.3%)

(0.8%)

154

45

4

5

1

0

0

3

1

(72.3%)

(21.1%)

(1.9%)

(2.3%)

(0.5%)

(0.0%)

(0.0%)

(1.4%)

(0.5%)

273

88

7

9

0

1

2

5

4

(70.2%)

(22.6%)

(1.8%)

(2.3%)

(0.0%)

(0.3%)

(0.5%)

(1.3%)

(1.0%)

41

FULL FINAL REPORT

Variable description All clients

N (%)

Control group

N (%)

Intervention group

N (%)

Weekly income $0

$1-$149

$150-$249

$250-$399

$400-$599

$600-$799

$800-$999

$1,000-$1,299

$1,300-$1,599

$1,600-$1,999

$2,000 or more

Refused

3

9

97

190

81

38

15

16

7

4

7

4

(0.64%)

(1.91%)

(20.59%)

(40.34%)

(17.20%)

(8.07%)

(3.18%)

(3.40%)

(1.49%)

(0.85%)

(1.49%)

(0.85%)

1

1

31

64

31

19

4

4

2

1

3

1

(0.62%)

(0.62%)

(19.14%)

(39.51%)

(19.14%)

(11.73%)

(2.47%)

(2.47%)

(1.23%)

(0.62%)

(1.85%)

(0.62%)

2

8

66

126

50

19

11

12

5

3

4

3

(0.65%)

(2.59%)

(21.36%)

(40.78%)

(16.18%)

(6.15%)

(3.56%)

(3.88%)

(1.62%)

(0.97%)

(1.29%)

(0.97%)

42

FULL FINAL REPORT

5.3. Retention in treatment

Length of time in current treatment

Clients were asked at baseline, how long they had been in their current treatment. Table 15 provides an overview of how long the cohort had been in their current treatment. Approximately 30.0% of the cohort had been in treatment between one to three years, while 23.94% of clients had been in treatment between five to ten years and 21.7% of clients had been in treatment for more than 10 years. This indicates that clients in the trial were relatively stable in terms of their treatment, with only a small proportion having been in treatment for less than a year.

Table 15 Number of years client’s had been in their current treatment

Number of years in current treatment All clients

N* %

Less than 1 year 34 6.90%

1 to 3 years 147 29.82%

3 to 5 years 87 17.65%

5 to 10 years 118 23.94%

10 to 15 years 57 11.56%

15 to 20 years 33 6.69%

More than 20 years 17 3.45%

*Missing = 219

Retention in treatment

Client retention was measured using the monthly client dosing data provided by pharmacists, which was substantiated by the drop-out data provided by the jurisdictions and clients. Retention over the six-months of the trial was high for this cohort of clients, with only 44 (18.3%) clients dropping-out of the control group 66 (14.0%) from the intervention group.

From the follow-up data it was identified that of the clients who had dropped out of the trial, 15 (14.6%) had successfully come off the program over the six-month period and 49 (47.6%) were identified has having transferred to another pharmacy. This indicates that of the 749 clients who started in the trial, 703 (93.9%) actually remained in treatment at six-months.

A Kaplan-Meier survival estimate was used to plot client retention for both the control and intervention groups (Figure 5). The mean number of days in treatment for the control group was 156.3 days and 164.5 for the intervention group.

43

FULL FINAL REPORT

Figure 5: Survival curve to time of drop-out

controlintervention

Kapla

n-M

eie

r E

stim

ate

0.70

0.75

0.80

0.85

0.90

0.95

1.00

Retention Time in days

0 20 40 60 80 100 120 140 160 180

The log-rank test of the difference between the two survival curves was found to be not significant (p=0.1192), indicating that there was no significant difference in retention between the control and intervention groups, although there was a clear trend in the predicted direction.

A GEE model was used to analyse client drop-out by month, taking into account repeated measures of client drop-out status. Table 16 indicates that the observed retention rate was slightly higher in the intervention group; however the overall difference in retention (approximately 5%) was not statistically significant (p=0.0830 from GEE model).

Table 16: Number of clients who stayed in the trial by month and treatment allocation

Month

Control group drop-outs

(baseline n=240)

Intervention group drop-outs

(baseline n=472) P-value

N (%) N (%)

Month 1 231 (96.25%) 463 (98.09%) 0.1455

Month 2 226 (94.17%) 448 (94.92%) 0.6743

Month 3 214 (89.17%) 438 (92.80%) 0.1001

Month 4 207 (86.25%) 430 (91.10%) 0.0466

Month 5 201 (83.75%) 417 (88.35%) 0.0861

Month 6 196 (81.67%) 406 (86.02%) 0.1275

Assuming a constant linear trend in drop-out rate (as presented in Figure 6), at a twelve-month follow-up with the cohort there would be a small increase in the difference in retention rate between the two groups; with 62% retention for the control group (95%CI

13=(51.9%, 73.2%)) and 72% retention for the intervention group

(95%CI=(65.5%, 78.6%)). However the difference is not statistically significant.

13

Confidence Interval (CI)

44

FULL FINAL REPORT

Figure 6 Retention rate per month

7 8 9 121110

These findings regarding retention are not surprising. While the literature identifies that client fees (ie the cost of participating in the program at a given pharmacy) can be a barrier to uptake of treatment, the literature does not unambiguously show a strong relationship between fees and retention in treatment. One example is research undertaken by Bammer et al., (2000) which followed a major expansion in treatment places for pharmacotherapy treatment in the ACT; over the same period client fees were introduced. Over the following two-year period, no evidence was found of a fall in the number of new admissions to the program, or a decrease in treatment retention due to the introduction of client fees.

Characteristics of clients associated with retention

Multivariate analyses were undertaken to test the association between time to drop-out and client characteristics, using the Cox proportional hazard regression model. The results of which were confirmed by a sensitivity analysis which was adjusted for pharmacy clustering.

Results of the multivariate analysis revealed that only age was found to be significantly associated with retention (p=0.0020), indicating that the older the client, the more likely they were to stay in treatment.

Research supports that older age is associated with treatment success. In addition, a review by Ward et al., (1998) found the following characteristics to be associated with improved outcomes: older age, absence of criminal behaviour, shorter duration of opioid use, less severe psychopathology, being employed, being married, less polydrug use, and higher expressed desire in getting help with a drug problem.

Reasons for client drop-out

A total of 44 (18.3%) control clients had dropped-out of the trial at six-months, while 66 (14.0%) of intervention clients had dropped-out (110 clients in total). Of the 749 clients who started in the trial, 639 remained in the trial at six-months.

Of the 110 drop-outs, follow-up data was obtained on 83. Table 17 provides an overview of the drop out reasons for those clients for whom follow-up data was obtained. For 27 clients, follow-up data was not received due to the pharmacist not knowing what happened to the client, the client was unreachable in the follow-up survey or jurisdictional data was not received. Of the 749 clients who started in the trial, 703 (93.9%) remained in treatment at the end of the trial.

45

FULL FINAL REPORT

Table 17 Client drop-out reason

Client drop-out reason All clients

N (%)

Control group

N (%)

Intervention group

N (%)

Finished program successfully 15 (14.56%) 7 (17.07%) 8 (12.90%)

Transferred to another pharmacy/ program 49 (47.57%) 21 (51.22%) 28 (45.16%)

Banned from pharmacy 6 (5.83%) 2 (4.88%) 4 (6.45%)

Incarcerated 11 (10.68%) 1 (2.44%) 10 (16.13%)

Died 2 (1.94%) 2 (4.88%) 0 (0.00%)

5.4. Compliance with treatment

Compliance with the treatment regimen was measured using the monthly client dosing data provided by pharmacists and examined the frequency of doses provided by the pharmacy, from which a missed dose could be identified.

Paired sample t-tests were undertaken to test the significance of compliance with doses, by month, in the control and intervention groups. Table 18 provides the results of each of the paired sample t-tests and shows that overall, the observed compliance with doses was high for both the intervention and control groups, with a small increase in observed compliance from month one to month six over the six-month period for both groups. Overall, no statistically significant differences were found between the control and intervention groups over time.

Table 18: Compliance of doses supplied by month among all available clients over the six-month trial

Month Control Intervention Control vs. Intervention

N Mean SE N Mean SE Mean

difference 95%CI P-value

Month 1 219 94.3 1.0581 457 94.5 0.7852 -0.24 -2.89 2.41 0.8579

Month 2 215 94.2 1.0765 456 94.5 0.7868 -0.34 -3.01 2.34 0.8055

Month 3 211 94.5 1.0605 444 95.2 0.7354 -0.69 -3.23 1.85 0.5922

Month 4 203 96.0 0.8338 436 95.8 0.6772 0.16 -2.08 2.41 0.8784

Month 5 186 97.0 0.7603 430 96.4 0.5908 0.46 -1.55 2.48 0.6291

Month 6 167 97.6 0.6811 399 96.7 0.5691 0.85 -1.08 2.79 0.3370

A GEE model was used to analyse compliance by the cost of pharmacotherapies for opioid dependence, taking into account repeated measures of client compliance and pharmacy clustering. The cost of pharmacotherapies was based on information provided by pharmacists on the fees that they charge their clients (ie the cost of participating in the program at their pharmacy), excluding the client subsidy. The total average cost of Methadone, Suboxone and Subutex per pharmacy, per week were calculated and categorised into three groups based on tertiles of cost:

Cost group Tertile of cost

Low cost group Up to $29.80 per week

Medium cost group Between $29.81 to $42.50 per week

High cost group Greater than $42.50 per week

46

FULL FINAL REPORT

By considering the compliance data as a function of absolute level of client payments, a significant difference was found between the low cost group and high cost group (p=0.005) for compliance with doses.

More specifically, the level of costs associated with client fees was found to have a significant impact on a client’s compliance with their doses. Those clients who were paying less than an average of $29.80 per week were more likely to be compliant than those who were paying more than $42.50 per week in fees, regardless of whether they were in the intervention or control group.

This finding indicates three things very clearly:

• First, a nationally consistent client subsidy, like that implemented in the trial, was unlikely to have had a uniform impact on affordability for all clients.

• Second, the specific level of client subsidy may not have been sufficient to improve compliance among clients at the higher end of the cost spectrum.

• Third, the cost actually incurred by clients participating in the trial, quite apart from whether they were in the intervention or control group, had a significant impact on their compliance with treatment, which is in line with previous research.

Taken together, these three findings point to the significant impact of the cost of treatment on the client.

5.5. Drug and alcohol use and related behaviours

Drug and alcohol use was measured using the BTOM-C which formed part of the client survey administered at baseline and the six-month follow-up. The key results at baseline and six-months are discussed in further detail below.

5.5.1. Baseline drug and alcohol use and related behaviours

Drug use and related behaviours

Figure 7 provides an overview of the main drug which led the client to seek treatment. For the majority of clients (74.6% control group and 82.5% intervention group) this was heroin. A total of 9.5% of clients described morphine as being the main drug which led them to seek treatment and for 2.7% it was oxycodone.

47

FULL FINAL REPORT

Figure 7: Main drug which led the client to seek treatment

0

10

20

30

40

50

60

70

80

90

Heroin Morphine Oxycodone Codene Methadone Other

All clients

Control

Intervention

Figure 8 describes how the clients usually took the main drug of concern. Injection of the drug was the most common method reported by all clients (86.5%), followed by ingestion (9.5%) and smoking (8.8%) of the drug.

Figure 8 How the main drug of concern was usually taken

0

10

20

30

40

50

60

70

80

90

Ingest Smoke Inject/

intravenous

Sniff/ snort Inhale Other

All clients

ControlIntervention

48

FULL FINAL REPORT

Approximately half of the client cohort (46.8%) reported that it had been longer than twelve months since they last injected any drug. While 23.5% of control clients and 20.6% of intervention clients reported that they had injected a drug within the previous month. Approximately 6% of all clients reported that they have never injected a drug.

Of the client cohort, 11.6% reported that in the previous three months, they had used a needle and syringe after someone else had already used it (12.7% control group and 11.0% intervention group). Twenty-eight percent (28.0%) reported that they had shared spoons, filters, water, tourniquets, drug solution or mix, or swabs in the previous three months (26.0% control group and 22.0% intervention group).

Figure 9 provides an overview of the other drugs of concern for the client cohort. Nearly half of the cohort (41.1%) reported that cannabis was a drug of concern for them. Speed (27.4%), sleeping pills (14.7%), alcohol (12.6%) and tobacco (10.5%) were also identified as drugs of concern for the client cohort.

Figure 9: Other drugs of concern

0

5

10

15

20

25

30

35

40

45

Cannabis Tobacco Cocaine Ecstasy Ice/ Crystal

meth

Speed Sleeping

pills/

depressants

GHB Prescription

drugs

Alcohol Other

All clients

Control

Intervention

Approximately 98% of the client cohort reported that they had not had an overdose in the previous three months, while:

• five clients reported that they had overdosed once (three from the control group)

• two reported that they had overdosed twice (one from the control group and one from the intervention group)

• one client from the intervention group reported they had overdosed twelve times in the previous three months.

49

FULL FINAL REPORT

Table 19 provides an overview of polydrug use at baseline, looking at the average number of days in the pervious month the drug was used. Again, the control and intervention groups were quite similar, with low drug use for heroin, other opioids, cocaine and amphetamines.

Table 19: Baseline polydrug use - average number of drug using days in the previous month

Variable Control Intervention

N Mean SE N Mean SE

Heroin 213 0.845 0.2872 389 1.1697 0.2243

Other opioid drugs 213 1.4131 0.3084 389 0.7969 0.1819

Cocaine 213 0.0751 0.0491 389 0.1028 0.0565

Amphetamines 213 0.3005 0.1057 389 0.3676 0.0962

Tranquilisers 213 5.7183 0.7417 388 6.0928 0.5834

Marijuana / cannabis 213 8.4977 0.8344 389 8.0283 0.6046

Alcohol and tobacco use

Table 20 provides an overview of the reported baseline cigarette and alcohol use over the previous month, for control and intervention clients. Overall, the control and intervention clients used cigarettes, on average, 26 days of the month, with an average of 16 cigarettes smoked per day. The control group drank alcohol, on average, 6.25 days of the month and the intervention group 5.28 days per month, both groups drinking approximately six drinks per day.

Table 20: Baseline tobacco and alcohol use over the previous month

Variable Control Intervention

N Mean SE N Mean SE

Days using cigarettes 213 26.68 0.64 389 26.17 0.5

Number of cigarettes per day 192 16.70 0.69 347 15.35 0.42

Total number of cigarettes for month 192 501.34 20.92 347 457.7 13.01

Days drinking alcohol 213 6.25 0.69 389 5.28 0.48

Number of drinks per day 105 6.18 0.56 169 6.31 0.52

Total number of drinks for month 105 85.32 12.56 169 87.91 13.44

5.5.2. Six-month drug and alcohol use and related behaviours

Drug use and related behaviours

Paired sample t-tests were undertaken to test the significance of treatment group on drug using days. Overall, there were no significant differences in drug use over time between the control and intervention groups, for any drug.

Table 21 provides the results of the paired sample t-test for heroin drug using days. Overall, the number of drug using days for this cohort of clients remained very low. From Table 21, it can be seen that the observed mean number of drug using days did slightly increase from baseline to six-months for the control group, while the intervention group remained approximately the same. However, the difference in heroin use over time between the control and intervention groups was not significant (p=0.1805).

50

FULL FINAL REPORT

Table 21: Mean number of heroin drug using days in the previous month – six-month follow-up

Control Intervention Control vs. Intervention

N Mean SE N Mean SE Mean

difference 95%CI P-value

Baseline 213 1.0845 0.2872 389 1.1697 0.2243 -0.085 -0.812 0.6419 0.8182

6 months 162 1.7593 0.4658 309 1.1683 0.2578 0.591 -0.373 1.5546 0.268

Difference 162 0.7407 0.387 309 0.1424 0.2513 0.5983 -0.278 1.4749 0.1805

The results of the paired sample t-tests for other drugs (ie other opioid based drugs, cannabis, cocaine, amphetamines, tranquilisers and other drugs) can be found in Appendix H.

From the BTOM-C, a polydrug use scale can be compiled which looks the number of days in the previous month, which the client used illicit drugs including heroin, other opioid based drugs, cannabis, cocaine, amphetamines, tranquilisers and other drugs.

Table 22 provides the results of the paired sample t-test for the polydrug use scale. Overall, polydrug use was low in this cohort of clients. There were slight decreases between baseline and six-months in the observed mean number of polydrug using days for both the control and intervention groups. However, the difference in polydrug use over time between the control and intervention groups was not significant (p=0.4348).

Table 22: Polydrug use in the previous month – six-month follow-up

Control Intervention Control vs. Intervention

N Mean SE N Mean SE Mean

difference 95%CI P-value

Baseline 213 1.4366 0.0837 389 1.3239 0.0628 0.1127 -0.094 0.3191 0.2838

6 months 162 1.2963 0.0998 309 1.2039 0.0666 0.0924 -0.137 0.3223 0.4299

Difference 162 -0.025 0.0975 309 -0.117 0.068 0.0918 -0.139 0.3226 0.4348

A Chi-square test was undertaken to test the significance of group difference on overdose frequency. The difference in overdoses at six-months, between the control and intervention groups, was not significant (p=0.1752).

Table 23 provides the results of the paired sample t-test for needle sharing frequency. The observed mean frequency of needle sharing in the previous month, decreased from baseline to six-months for the control group, while the intervention group remained the same. However, the difference in needle sharing frequency over time between the control and intervention groups was not significant (p=0.6686).

Table 23: Needle sharing frequency in the previous month - six-month follow-up

Control Intervention Control vs. Intervention

N Mean SE N Mean SE Mean

difference 95%CI P-value

Baseline 9 7.1111 2.6164 13 4.1538 2.1834 2.9573 -4.155 10.07 0.3961

6 months 7 2.1429 0.5533 4 4.25 2.136 -2.107 -5.993 1.779 0.4023

Difference 3 -5 5.0332 1 0 . -5 -48.31 38.312 0.6686

51

FULL FINAL REPORT

Tobacco and alcohol use

Paired sample t-tests were undertaken to test the significance of treatment group on tobacco and alcohol use. At six-months, there was a significant difference in the average number of cigarettes smoked per day (in the previous month) between the control and intervention group (p=0.0052) and the total number of cigarettes smoked in the month (p=0.0082), with higher use in the control group.

Table 24: Six-month tobacco and alcohol use over the previous month

Variable Control Intervention Control vs. Intervention

N Mean SE N Mean SE Mean dif. 95%CI P-value

Days using cigarettes

162 26.68 0.75 307 26.21 0.57 0.47 -1.413 2.3609 0.622

Number of cigarettes per day

147 17.70 0.74 274 15.12 0.54 2.57 0.7747 4.3775 0.0052

Total number of cigarettes

147 529.39 22.56 274 455.13 16.57 74.26 19.275 129.24 0.0082

Days drinking alcohol

168 5.77 0.77 308 5.63 0.57 0.14 -1.751 2.0286 0.8853

Number of drinks per day

77 5.29 0.58 142 5.80 0.40 -0.52 -1.88 0.846 0.4555

Total number of drinks

77 82.62 18.36 142 70.09 7.93 12.53 -21.51 46.573 0.5323

5.6. Health and social outcomes

Health and social outcomes were measured using the BTOM-C and the WHOQoL-8, which formed part of the client survey administered at baseline and the six-month follow-up. The key results at baseline and six-months are discussed in further detail below.

5.6.1. Baseline health and wellbeing

Social functioning questions from the BTOM-C were used to measure the client’s recent (ie the past three months) levels of: financial hardship; conflict in relationships with spouses/partners; other relatives and employers/school staff and students; time spent living with a drug user and time spent with non-drug users.

Figure 10 provides an overview of how the client cohort reported their financial situations and conflict in relationships over the previous three months. Overall, there were no notable differences between the control and intervention groups. The key findings include:

• Just under one third of clients from both the control and intervention groups reported that they had ‘never or almost never’ had any money problems (including not having enough for food or housing). While approximately one fifth of the cohort reported that they ‘always or nearly always’ had money problems.

• Approximately one third of each cohort reported that they ‘never or almost never’ had conflict with their partner/spouse or their employee over the previous three months.

• Over half the cohort reported that they ‘never or almost never’ had conflict with relatives over the previous three months (62.9% control group and 61.2% intervention group).

52

FULL FINAL REPORT

Figure 10 Baseline client social functioning over the past three months

0

10

20

30

40

50

60

70

Money

problems

CONTROL

Money

problems

INTERVENTION

Conflict with

partner/

spouse

CONTROL

Conflict with

partner/

spouse

INTERVENTION

Conflict with

relatives

CONTROL

Conflict with

relatives

INTERVENTION

Conflict with

employer/

colleagues

CONTROL

Conflict with

employer/

colleagues

INTERVENTION

Never or almost never Sometimes Often Always or nearly always

Approximately 80.0% of the cohort reported that they had not lived with anyone who used heroin or other illicit drugs over the previous three months. While 9.4% of the control group and 5.9% of the intervention group reported that they lived with someone who used heroin/other illicit drugs.

Of client cohort, 31.5% (control) and 25.7% (intervention) clients reported that they spent all of their time (over the previous three months) with people who did not use heroin or other illicit opioids. The proportion of the client cohort who spent their time with friends who used heroin/other illicit opioids was 13.6% from the control group and 18.5% from the intervention group.

From Figure 11 it can be seen that the majority of clients had not had any arrests in the previous three months (92.0% control group and 92.3% intervention). Approximately 5.0% of the cohort had been arrested once and one client from the intervention group had been arrested between seven to ten times.

53

FULL FINAL REPORT

Figure 11 Arrests in the previous three months

0

10

20

30

40

50

60

70

80

90

100

None 1 2 3 to 4 5 to 6 7 to 10

All clients

ControlIntervention

5.6.2. Six-month health and wellbeing

A paired sample t-test was undertaken to test the significance of treatment group on social functioning (using the social functioning scale from the BTOM-C). Overall, there were no significant differences in social functioning over time between the control and intervention groups (p=0.1374), with social functioning remaining high for both groups (Figure 12). The results of each individual item of the scale are included in Appendix H.

54

FULL FINAL REPORT

Figure 12: Six-month client social functioning over the past three months

0

10

20

30

40

50

60

70

Money

problems

CONTROL

Money

problems

INTERVENTION

Conflict with

partner/

spouse

CONTROL

Conflict with

partner/

spouse

INTERVENTION

Conflict with

relatives

CONTROL

Conflict with

relatives

INTERVENTION

Conflict with

employer/

colleagues

CONTROL

Conflict with

employer/

colleagues

INTERVENTION

Never or almost never Sometimes Often Always or nearly always

At six-months, approximately 84.0% of the cohort reported that they had not lived with anyone who used heroin or other illicit drugs over the previous three months (compared to 80.0% at baseline). While 6.8% of the control group and 7.1% of the intervention group reported that they lived with someone who used heroin or other illicit drugs in the previous three months.

Of client cohort, 32.3% (control) and 34.4% (intervention) clients reported that they spent all of their time (over the previous three months) with people who did not use heroin or other illicit opioids at the six-month follow-up. The proportion of the client cohort who spent time with friends who used heroin or other illicit opioids was approximately 15.0% from both the control group and intervention groups.

A Chi-square test was undertaken to test the significance of treatment group on arresting frequency. The difference in arresting frequency at six-months, between the control and intervention groups, was not significant (p=0.0928). A high proportion of clients (90.1% control group and 95.5% intervention group) reported they had no arrests in the last three months.

5.6.3. Baseline quality of life

The WHOQoL-8 was used to measure the clients’ perceived quality of life. There were no notable differences between the control and intervention groups with respect to perceived quality of life, using standardised instruments. At the outset of the trial 40.3% of the control group and 40.9% of the intervention group described their overall quality of life as being ‘good’, while 15.6% of the control group and 17.5% of the intervention group described their overall quality of life as being ‘poor’ (Figure 13).

55

FULL FINAL REPORT

Figure 13: Baseline overall quality of life

0

5

10

15

20

25

30

35

40

45

50

Very poor Poor Neither poor or

good

Good Very good

Control

Intervention

Figure 14 and Figure 15 provide an overview of how the control and intervention clients rated themselves on other items of the quality of life scale. Overall, the control group and intervention groups were quite similar:

• thirty-nine percent (39.0%) of control and 36.3% of intervention clients were ‘satisfied’ with their health, while 25.4% and 31.3% of control and intervention clients were ‘dissatisfied’

• forty-one percent (41.0%) of control and 41.0% of intervention clients were ‘satisfied’ with themselves, while 22.3% and 24.5% of control and intervention clients were ‘dissatisfied’

• seventy three percent (73.0%)of control and 76.9% of intervention clients were ‘satisfied’ or ‘very satisfied’ with the condition of the place in which they live.

Figure 14: Baseline quality of life items for control clients

0

5

10

15

20

25

30

35

40

45

Satisfaction with

health

Satisfaction with

ability to perform

daily living

activities

Self satisfaction Satisfaction with

personal

relationships

Satisfaction with

living conditions

Control

Very dissatisfied

Dissatisfied

Neither satisfied/dissatisfied

Satisfied

Very satisfied

56

FULL FINAL REPORT

Figure 15: Baseline quality of life items for intervention clients

0

5

10

15

20

25

30

35

40

45

Satisfaction with

health

Satisfaction with

ability to perform

daily living

activities

Self satisfaction Satisfaction with

personal

relationships

Satisfaction with

living conditions

Intervention

Very dissatisfied

Dissatisfied

Neither satisfied/dissatisfied

Satisfied

Very satisfied

5.6.4. Six-month quality of life

A paired sample t-test was undertaken to test the significance of treatment group on quality of life (using the WHOQoL-8). Overall, there were no significant differences in quality of life over time between the control and intervention groups (p=0.5798), with quality of life remaining reasonability high for both groups.

The results of each individual item of the scale are included in Appendix H.

5.7. Client financial strain

The six-month client survey included questions which examined the financial impact of participating in the pharmacotherapy program on their lives (over the previous six-months). Discussed below is the variation in the financial impact between the control and intervention groups. None of the observed differences between the two groups approached statistical significance.

Overall, 41.6% of the client cohort reported that the program had a ‘large impact’ on their financial circumstances (36.7% control group and 44.1% intervention). A total of 8.1% of clients reported that the program had ‘no impact’ on their financial circumstances.

Figure 16 provides an overview of the reported financial impact on aspects of their life; overall the control and intervention groups were extremely similar. Key points from these data include:

• Seventy-one percent (71.0%) of control clients and 67.7% of intervention clients reported that they have not had to go without prescribed medications over the previous six-months, as a result of the financial impact of the program.

• Ninety-five percent (95.0%) of all clients reported that they have not had to sleep on the street or in a squat, refuge or other type of homeless shelter over the previous six-months, as a result of the financial impact of the program.

• Sixty-six percent (66.0%) of control clients and 70.2% of intervention clients reported that they have not had to use supportive services over the previous six-months, as a result of the financial impact of the program.

57

FULL FINAL REPORT

• Eighty-eight percent (88.0%) of control clients and 92.3% of intervention clients reported that they have not had an argument with their pharmacist about client fees over the previous six-months.

• Ninety-four percent (94.0%) of all clients reported that they have not sought money from sources such as crime, asking strangers for small change or sex work, over the previous six-months, as a result of the financial impact of the program.

• Sixty-nine percent (69.0%) of all clients reported that they have not incurred debts or taken up loans over the previous six-months, as a result of the financial impact of the program.

• Forty-five percent (45.0%) of the debts / loans incurred equated to less than $500 and 18.0% were for $2000 or more.

Figure 16: Financial impact on the client of participation in the program

0

10

20

30

40

50

60

70

80

90

100

Gone

without

food

Not gone

without

food

Gone

without

prescribed

medicines

Not gone

without

prescribed

medicines

Slept on

street,

squat,

refuge

Not slept on

street,

squat,

refuge

Used

supportive

services

Not used

supportive

services

Money from

crime,

begging,

sex work

No money

from crime,

begging,

sex work

Debts and

loans

No debts

and loans

Control Intervention

Figure 17 provides an overview of the level of subsidy reported by clients that would increase the likelihood of staying in treatment. A total of 65.0% of clients reported that a reduction of $5 per week in their fees would increase the likelihood of them staying on the program and 85.5% reported that a $20 per week reduction would increase the likelihood of staying in treatment.

58

FULL FINAL REPORT

Figure 17: Client perception of the level of subsidy that would increase likelihood of staying on program

0

10

20

30

40

50

60

70

80

90

$5 reduction per/wk $10 reduction per week $15 reduction per week $20 reduction per week

Yes

Maybe

No

5.8. Client satisfaction

Client satisfaction was measured at baseline and at six-months, using a five-point likert scale, as part of the client survey. The key results at baseline and six-months are discussed in further detail below.

5.8.1. Baseline client satisfaction

There were no notable differences between the control and intervention groups’ levels of satisfaction. At baseline, the majority of clients (59.2% of control clients and 53.9% of intervention clients) reported being ‘dissatisfied’ or ‘very dissatisfied’ with the fees they were charged for their doses. Just under one third of clients (27.2% control and 31.9% intervention clients) reported being ‘satisfied’ or ‘very satisfied’ with the fees (Figure 18).

Figure 18: Baseline client satisfaction with fees

0

5

10

15

20

25

30

35

Very

dissatisfied

Dissatisfied Neither

satisfied or

dissatisfied

Satisfied Very satisfied

Control

Intervention

59

FULL FINAL REPORT

5.8.2. Six-month client satisfaction

A paired sample t-test was undertaken to test the significance of treatment group on client satisfaction. There was a significant difference between the control and intervention groups in overall client satisfaction with the fees charged for their doses (p=<.0001); with 30.9% of the control group and 53.1% of the intervention group reporting that they were ‘satisfied’ or ‘very satisfied’ with the fees (Figure 19).

Figure 19: Six-month client satisfaction with fees

0

5

10

15

20

25

30

35

40

Dispensing fees CONTROL Dispensing fees

INTERVENTION

Very disatisfied

Dissatisfied

Neutral

Satisfied

Very satisfied

5.9. Costs of pharmacotherapy treatment to clients

5.9.1. Background and context

The Australian Government funds the provision of pharmacotherapies via pharmaceutical benefits arrangements, through clinics and pharmacies approved by state and territory governments. That is, the federal government provides pharmacotherapies free of charge to those responsible for supplying them (eg public/private clinic, community pharmacy, prison or hospital); however the costs associated with supplying the pharmacotherapies to clients is not funded.

Australia’s official statistics on health expenditure do not identify the proportion allocated to drug and alcohol treatment, nor specifically to pharmacotherapy treatment (Australian Institute of Health and Welfare 2008).

State/Territory governments subsidise service providers in some instances; however they vary in their models. For example, the NSW Government pays a once-only incentive payment to pharmacies new to supplying pharmacotherapy and pays pharmacies ongoing incentives for continuous clients. The ACT Government subsidises community pharmacies to the value of $20 per week per client. The Tasmanian Government also provides incentive payments to pharmacists (Ritter & Chalmers 2009).

In public clinics in NSW, QLD, SA, WA and NT, the State/Territory government covers the client fee and in the other jurisdictions (VIC, TAS and the ACT) the clients pay some proportion of the fee, while the supply of pharmacotherapy treatment in prison is funded by State/Territory governments. In community pharmacies

60

FULL FINAL REPORT

and private clinics, across Australia, clients pay the cost of fees. Adding a further level of complexity is the different policies that are implemented at the pharmacy level (such as the provision of credit).

Currently State, Territory and federal governments, in addition to clients, all contribute to the funding of opioid pharmacotherapy treatment services. A recent study has estimated that the overall cost for the provision of pharmacotherapy treatment is $11.73 million per month (Chalmers et al., 2008). Forty-three percent (43.0%) of this cost is borne by State/Territory governments ($5.06m per month), 33% is borne by the clients ($3.90m per month), and the remaining 24% is borne by the federal government ($2.77m per month).

The following section examines the costs of participating in the program for clients.

5.9.2. The costs of participating on the program for clients

Cost of program participation, prior to the application of the client subsidy

Information was collected from trial pharmacies at baseline on the daily fees that they charge their clients for participating in the program at their pharmacy. These data were collected prior to the recruitment of clients to the trial and therefore do not include the client subsidy. The following tables outline the average cost per day for Methadone (Table 25), Subutex (Table 26) and Suboxone (Table 27).

Overall, the amount that pharmacies reported clients being charged per day, varied considerably within the three types of pharmacotherapies (Methadone, Subutex and Suboxone); ranging from $1.43 to $10 per dose. There was little variation in cost across the type of pharmacotherapy and the variation by jurisdiction did not appear to be particularly systematic. The overall median cost per day to the client ranged from $4.65 to $5.00 across Methadone, Subutex and Suboxone.

Table 25: Daily average cost of Methadone to the client

State Pharmacies (N) Mean Std deviation Minimum Maximum Median

NSW 22 4.99 1.58 1.43 10.00 5.00

NT 1 5.00 . 5.00 5.00 5.00

QLD 16 3.82 1.05 2.15 6.00 3.79

TAS 2 4.90 0.14 4.80 5.00 4.90

VIC 16 4.66 0.79 3.57 6.00 5.00

WA 9 4.52 0.44 4.00 5.00 4.50

SA 3 2.90 0.66 2.20 3.50 3.00

Total 69 4.49 1.23 1.43 10.00 4.65

61

FULL FINAL REPORT

Table 26: Daily average cost of Subutex to the client

State Pharmacies (N) Mean Std deviation Minimum Maximum Median

NSW 10 5.46 2.17 1.79 10.00 5.00

NT 1 5.00 . 5.00 5.00 5.00

QLD 8 4.51 1.30 3.00 7.00 4.40

TAS 1 5.00 . 5.00 5.00 5.00

VIC 14 4.64 0.70 3.57 6.00 5.00

WA 7 5.65 0.75 5.00 7.00 5.57

SA 3 2.90 0.66 2.20 3.50 3.00

Total 44 4.86 1.42 1.79 10.00 5.00

Table 27: Daily average cost of Suboxone to the client

State Pharmacies (N) Mean Std deviation Minimum Maximum Median

NSW 14 5.24 1.91 1.79 10.00 5.00

NT 1 5.00 . 5.00 5.00 5.00

QLD 15 4.17 1.23 2.15 7.00 4.29

TAS 1 5.00 . 5.00 5.00 5.00

VIC 16 4.66 0.79 3.57 6.00 5.00

WA 6 5.76 0.75 5.00 7.00 5.79

SA 3 2.90 0.66 2.20 3.50 3.00

Total 56 4.71 1.39 1.79 10.00 5.00

Cost of program participation, following the application of the client subsidy

Pharmacies in the intervention group were asked to provide detail on what they charged their clients per week (on average), after the $15 per week client subsidy had been applied. All pharmacies had been informed at baseline that their clients should be charged a minimum of $5 per week. Three pharmacies reported that they did not charge their clients anything for their doses; the three pharmacies had a total of six clients participating in the trial. These pharmacies tended to be located in rural and remote areas.

62

FULL FINAL REPORT

Table 28 Average cost per week of client fees for intervention clients – after the application of the $15 client subsidy

State Pharmacies (N) Mean Std deviation Minimum Maximum

NSW 15 $16.27 $6.81 $0 $27

NT 1 $15.00 . $15 $15

QLD 9 $9.11 $9.39 $0 $20

TAS 1 $18.00 . $18 $18

VIC 8 $12.75 $4.83 $5 $20

WA 4 $21.75 $11.35 $10 $35

SA 1 $0.00 . $0 $0

Total 39 $3.66 $7.75 $0 $35

5.10. Summary of client results

The focus of the present study was to consider key elements of a national funding model for pharmacotherapy treatment. A key plank of the funding model was improved affordability for the client group, through a nationally consistent subsidy of the treatment cost for clients. The benefits of the implementation of a nationally consistent client subsidy, in terms of impacts on key client outcomes, need to be considered against the background of the variable cost of participating in the pharmacotherapy program in Australia.

Overall, the findings of this national trial lend weight to the impact of affordability as a determinant of treatment outcome. A clinically non-trivial improvement in retention (around 5% improvement) was observed as a consequence of receiving the client subsidy. The trend towards improved retention was corroborated by significantly higher satisfaction with the program among intervention clients, compared with control clients. The level of cost associated with client payments for treatment had a significant impact on client compliance with their dosing regime. Taken together, the findings indicate that not all clients can afford the costs associated with pharmacotherapy treatment and that a subsidised arrangement as part of a national funding model is likely to reap greater benefits from the program.

63

FULL FINAL REPORT

6. Trial results – Pharmacy

This section presents the results of the trial at the pharmacy level, including:

• Pharmacy recruitment and retention (see Section 6.1)

• Pharmacy demographics (see Section 6.2)

• Pharmacy satisfaction (see Section 6.3)

• The service delivered by the community pharmacist and the associated costs (see Section 6.4)

• Summary of results (see Section 6.5).

6.1. Pharmacy recruitment and retention

Based on the recruitment process described in Section 4.3.1, a total of 79 pharmacies were consented to the trial. Of these pharmacies, 74 actively participated in the trial (ie recruited clients and/or provided data to the trial) and 72 pharmacies completed the trial. There were 25 pharmacies in the control group and 49 in the intervention group at the outset of the trial, with 24 control and 48 intervention pharmacies completing the trial. Figure 20 illustrates the process for recruitment and the retention of pharmacies during each stage of the trial.

Figure 20: Process for the recruitment and retention of pharmacies

282 pharmacies expressed interest

79 pharmacies consented

74 pharmacies participated in the trial

81 pharmacies selected based on sampling

framework

25 pharmacies in control group

49 pharmacies in intervention group

25 pharmacies at month 1

48 pharmacies at month 1

48 pharmacies at month 2

48 pharmacies at month 3

48 pharmacies at month 4

48 pharmacies at month 5

48 pharmacies at month 6

24 pharmacies at month 2

24 pharmacies at month 3

24 pharmacies at month 4

24 pharmacies at month 5

24 pharmacies at month 6

1 pharmacy withdrew 1 pharmacy withdrew

64

FULL FINAL REPORT

6.2. Pharmacy demographics

A summary of the demographic characteristics of the participating pharmacies are presented in Table 29. As described in Section 4.2, pharmacies were selected to fill a sampling framework based on PhARIA categories The highest proportion of participating pharmacies were from NSW (32.43%), QLD (24.32%) and VIC (22.97%). Of the participating pharmacies, 64.86% were in PhARIA category 1 (ie are highly accessible) and 57.53% were independent pharmacies.

Table 29: Demographic characteristics of pharmacies

Variable description All pharmacies

N (%)

Control group

N (%)

Intervention group

N (%)

State NSW

NT

QLD

TAS

VIC

WA

SA

24

1

18

2

17

9

3

(32.43%)

(1.35%)

(24.32%)

(2.70%)

(22.97%)

(12.16%)

(4.05%)

5

0

7

1

8

2

2

(20.00%)

(0.00%)

(28.00%)

(4.00%)

(32.00%)

(8.00%)

(8.00%)

19

1

11

1

9

7

1

(38.78%)

(2.04%)

(22.45%)

(2.04%)

(18.37%)

(14.29%)

(2.04%)

PhARIA category Category 1

Category 2-6

48

26

(64.86%)

(35.14%)

16

9

(64.00%)

(36.00%)

32

17

(65.31%)

(34.69%)

Pharmacy type Brand

Independent

31

42

(42.47%)

(57.53%)

13

11

(54.17%)

(45.83%)

18

31

(36.73%)

(63.27%)

Table 30 provides an overview of the numbers of clients that pharmacies had on the pharmacotherapy program at baseline for Methadone, Subutex and Suboxone. Overall, the pharmacies were quite similar in terms of average client numbers for each of the pharmacotherapies.

Table 30: Average number of clients on the program

Pharmacies (N) Mean Std deviation Minimum Maximum

Control Methadone

Subutex

Suboxone

24

24

24

9.71

1.17

4.42

7.88

1.46

3.96

1

0

0

30

4

14

Intervention Methadone

Subutex

Suboxone

49

49

49

9.47

1.53

3.59

10.35

4.40

5.83

0

0

0

50

30

30

6.3. Pharmacy satisfaction

Pharmacy satisfaction was measured at baseline and at six-months, using a five-point likert scale. The key results of the pharmacy satisfaction survey are outlined below for baseline and six-months.

65

FULL FINAL REPORT

6.3.1. Baseline pharmacy satisfaction

At baseline, 29.2% of control pharmacies and 42.9% of intervention pharmacies reported being ‘satisfied’ overall with the pharmacotherapy program (Figure 21). In addition:

• The majority of both control and intervention pharmacies (69.6% and 57.1%, respectively) reported that they were ‘satisfied’ with their current interactions and relationships with their clients, or very satisfied (13.0% and 12.2%, respectively) with these relationships.

• Thirty-nine percent (39.0%) of control group pharmacies and 38.8% of intervention pharmacies reported that they were ‘dissatisfied’ with the financial arrangements for the provision of pharmacotherapy treatment.

• Forty-six percent (46.0%) of control group pharmacies and 42.9% of intervention pharmacies reported that they were ‘dissatisfied’ with the client contribution / co-payment for pharmacotherapy treatment.

Figure 21: Baseline pharmacy satisfaction for control and intervention groups

0

10

20

30

40

50

60

70

Current

financial

arrangements

CONTROL

Current

financial

arrangements

INTERVENTION

Consumer

contribution

CONTROL

Consumer

contribution

INTERVENTION

Client

relationships

CONTROL

Client

relationships

INTERVENTION

Overall

satisfaction

CONTROL

Overall

satisfaction

INTERVENTION

Very dissatisfied Dissatisfied Neither satisfied/dissatisfied Satisfied Very satisfied

6.3.2. Six-month pharmacy satisfaction

A Chi-square test was undertaken to test the significance of treatment group on pharmacy satisfaction at the completion of the trial. Key findings were:

• There was a significant difference between the control and intervention pharmacies in overall pharmacy satisfaction with the pharmacotherapy program at six-months (p=0.0002), with 55.5% of control pharmacies and 92.5% of intervention pharmacies reporting that they were ‘satisfied’ or ‘very satisfied’ (Figure 22).

66

FULL FINAL REPORT

• There was a significant difference between the control and intervention pharmacies in satisfaction with the funding arrangements for the supply of pharmacotherapies for opioid dependence over the six-months (p=0.0001), with 40.0% of control pharmacies and 92.9% of intervention pharmacies reporting that they were ‘satisfied’ or ‘very satisfied’ (Figure 22).

• There was a significant difference between the control and intervention pharmacies in satisfaction the client contribution / co-payment over the six-month period (p=0.0026), with 45.0% of control pharmacies and 83.3% of intervention pharmacies reporting that they were ‘satisfied’ or ‘very satisfied’ (Figure 22).

• There was a significant difference in satisfaction with overall satisfaction with the trial between the intervention and control groups six-months (p=0.0009), with 75.5% of control pharmacies reporting that they were ‘satisfied’ and 92.4% of intervention pharmacies reporting that they were ‘satisfied’ or ‘very satisfied’ (Figure 22).

Figure 22: Six-month pharmacy satisfaction

0

10

20

30

40

50

60

70

80

Financial

arrangements

CONTROL

Financial

arrangements

INTERVENTION

Client

contribution/

co-payment

CONTROL

Client

contribution/

co-payment

INTERVENTION

Interactions &

relationships

with clients

CONTROL

Interactions &

relationships

with clients

INTERVENTION

Program

satisfaction for

last 6 months

CONTROL

Program

satisfaction for

last 6 months

INTERVENTION

Overall

satisfaction

with the trial

CONTROL

Overall

satisfaction

with the trial

INTERVENTION

Very disatisfied Dissatisfied Neutral Satisfied Very satisfied

The vast majority of intervention pharmacies reported being satisfied with the client subsidy. A total of 51.3% of intervention pharmacies reported that they were ‘very satisfied’ with the client subsidy that was provided in the trial and 35.9% reported that they were ‘satisfied’.

Of the intervention pharmacies, 38.8% reported that they were ‘very dissatisfied’ and 34.7% ‘dissatisfied’ with the reimbursement that they were provided with as part of the trial.

67

FULL FINAL REPORT

6.4. The service delivered by community pharmacists and the associated costs

As described in Section 4.5.3, 20 pharmacies across NSW, QLD and VIC participated in the observational study. Three observation periods of approximately three hours each were undertaken in each pharmacy and all activities associated with each client presenting in that time were timed and recorded. A total of 218 client presentations were observed in the study.

The purpose of the observational study was to obtain comprehensive snapshot of the way in which pharmacists deliver the service and the costs associated with the delivery of the service.

6.4.1. The service delivered by the community pharmacist

Figure 23 provides an overview, in general terms, of the way in which pharmacotherapy treatment was found to be delivered across the observed 20 pharmacies. The process map demonstrates the diversity of the service that the community pharmacist provides and highlights that the service is user driven and there is a need for flexibility on the part of the pharmacist.

While there were variations in the way in which the activities outlined in Figure 23 were undertaken it was clear that it was possible to identify broad categories of activities – those involving direct client interaction, those that were client related but not direct face-to-face interactions, and business related back-end and administration activities that varied depending on the pharmacy’s business model.

From Figure 23, it can be seen that back-end business related activities includes things like ordering, storage, record keeping and the like. Direct client interaction activities include pre-screening of the client to assess their eligibility for treatment on the day, or perhaps counselling the client. Indirect client related activities include pre-preparation of doses, inducting new clients and communication with other service providers. Perhaps most obviously from the process map, it can be seen that the clinical judgement of the pharmacist is key determinant of the interaction with the client on any given occasion.

The process map in Figure 4 is also the basis of the observational data capture tool used for the observational study. For the purposes of data collection, analysis and reporting, the schema in the process map was classified into eight basic categories. All but one of the categories, administration, are client related activities:

• preparation of doses

• pre-screening the client

• recording of the doses

• dosing the client

• processing payments

• counselling / talking with the client

• administration activities

• other activities eg inducting new clients, communication with other providers etc

68

FULL FINAL REPORT

Figure 23 Overview of the way in which pharmacotherapy treatment is delivered in community pharmacy

69

FULL FINAL REPORT

6.4.2. Average time spent by pharmacists on regularly occurring activities

As described previously, the specific activities which were agreed with the Advisory Panel as being regularly occurring activities and included as part of the observational component of the activity based costing study, were defined as:

• preparation of doses

• pre-screening the client

• recording of the doses

• dosing the client

• processing payments

• counselling / talking with the client

• administration activities

• other activities eg inducting new clients, communication with other providers etc

The average observed time spent on each of the categorised activities, per client, is described in Table 31.

Overall, the greatest proportion of pharmacist time was devoted to clinical care delivery associated with the treatment Approximately one fifth of the pharmacists’ time was spent interacting and conversing with the client during the pre-screening of the client and the provision of counselling to the client. In total, clinical activities - those relating to the delivery of the pharmacotherapy treatment (preparation of the dose, dosing the client), together with counselling and pre-screening - accounted for around three quarters of the time spent by pharmacists in delivering the service.

The single most time consuming activity from the observational study was the ‘other activities’ category, reflecting the ad hoc nature of the requirements that might occur in the community pharmacy setting. This finding is well corroborated anecdotally by pharmacists who report the occasional, but essential, activities that they may need to be involved in when a client has a crisis. A consistently reported anecdotal example of this is assisting clients in their dealings with other healthcare and social welfare providers.

Table 31: Average time spent by pharmacist on regularly recurring activities, expressed as time per observation (client)

Mean (mins)

Std deviation

(mins)

Minimum

(mins)

Maximum

(mins)

Median

(mins)

Proportion of

pharmacist time based on mean

Preparation of doses 1.26 0.90 0.25 4.0 1.04 36.04%

Pre-screening the client

0.35 0.32 0.25 1.5 1.25 5.96%

Recording of doses 0.85 1.14 0.25 5.0 0.48 16.69%

Dosing the client 0.71 0.70 0.25 3.0 0.51 19.77%

Processing payments 0.67 0.55 0.25 2.4 0.50 4.08%

Counselling / talking with the client

0.89 0.55 0.25 2.5 0.77 11.50%

Administration activities

1.31 0.82 0.25 2.5 1.00 2.15%

Other activities* 1.43 1.55 0.38 5.38 0.25 3.81%

*Includes occasional but client related activities such as inducting new clients, communication with other providers etc

70

FULL FINAL REPORT

While the data in Table 31 provide robust average data, it also documents the significant variability attached to each of the main classes of activity. Thus, while on average, the time spent (per client) by pharmacists on administration activities per client was among the highest, the quantum of time spent on this activity was reasonably small (the maximum being 2.5 minutes). In contrast, other (and often ad hoc) client related activities ranged from less than a minute to over five minutes in duration. From the perspective of pharmacy practice, this suggests an imperative need for flexibility in the provision of pharmacotherapy treatment. Variation due to changing client needs is an inherent part of the service delivery model.

The results of the observational study demonstrate that the time/cost drivers for delivery of pharmacotherapy treatment in the community pharmacy setting are comprised of two separate variables that would need to be part of a national funding model - 'service delivery models' and 'levels of complexity' – both of which are driven by the needs of the client:

• Service delivery models: six service delivery models were identified, reflecting three models of care which are underpinned by pharmacy business models (these are described in Section 6.4.3).

• Levels of complexity: it was also identified that there was unpredictable variation in the time per client interaction, related to the variable and unpredictable time that pharmacists needed for counselling the client and coordinating with other healthcare and social welfare providers. We have called this a ‘level of complexity' because it is analogous to the urgency/complexity features that characterise Emergency Care classification systems.

The data document that the time spent by the pharmacist engaging with, supporting and interacting with this client group is a fundamental part of the service offered by community pharmacists.

A robust national funding model, one that explains the variation in the time spent / cost to deliver a pharmacotherapy service in the community pharmacy, will need to be an algorithm comprised of both a 'service delivery model' component and 'level of complexity' component.

6.4.3. The time and cost drivers of the provision of pharmacotherapy treatment in the community pharmacy

A key purpose of the observational study was to obtain robust estimates of the costs associated with providing the pharmacotherapy program in community pharmacies, though costing: (a) the provision of an occasion of service by the pharmacy (ie each client interaction); and (b) consumables and investments made specifically for the purposes of the program. The costs obtained in the study are based on the following factors:

• the amount of time involved in specific activities directly related to each client interaction – as described in Table 31

• the hourly cost of the pharmacist

• investments made and ongoing costs in delivering the program.

Service delivery models

From the observational study it was apparent that the fundamental processes between each pharmacy did not greatly differ. However two main dimensions resulted in systematic variation in the processes undertaken. These variations occurred regularly within individual pharmacies; rather than systematically varying by pharmacy.

The first dimension driving variation of process was clearly based on client health need, ie based on the type of treatment required. These can be best thought of as different treatment ‘models of care’, analogous for instance to home dialysis vs hospital dialysis. The data from the observational study revealed that there are three such discernable models of care for the delivery of pharmacotherapy treatment to clients in the community pharmacy.

71

FULL FINAL REPORT

These models of care were based on whether:

• the client’s dose was taken in the pharmacy

• the client’s dose was a take away, or

• the client had had a dose in the pharmacy and also received a take-away dose.

These models of care are underpinned by the second dimension driving variation of process, the pharmacist driven business model, ie whether the doses were pre-prepared or not. While a range of other occasional variations were observed in business processes, these were not consistently observed across the pharmacies and therefore were not included in this analysis. For example: the use of a syringe or manually measuring doses, whether medication was stored back in the safe after each dosing event, or whether the client signed individual dosing sheets to record each dose.

Figure 24 shows that the combination of models of care with business models resulted in six possible service delivery models, which together capture the observational data in a form suitable for cost analysis.

Figure 24: Observed models of care, business models and resulting service delivery models

Take away dose

In-pharmacy + take away

dose

Dose pre-prepared

Dose not pre-prepared

Dose pre-prepared

Dose not pre-prepared

Dose pre-prepared

Dose not pre-prepared

In-pharmacy dose

Service delivery model 1

Service delivery model 2

Service delivery model 3

Service delivery model 4

Service delivery model 5

Service delivery model 6

Model of care:

Business model:

Service delivery model:

Hourly costs

During the observational study, it became apparent that in most pharmacies the Pharmacist in Charge was responsible for the regular activities that were observed as part of the pharmacotherapy program. This is unsurprising, given the sensitive clinical nature of the pharmacotherapy program. The implication for the current analysis is that the rate used for the calculation of costs was based on the average base hourly rates for the Pharmacist-in-Charge level (Association of Professional Engineers, Scientists & Managers, Australia, 2009). The rate used was $47.53 per hour which included 25% on-costs to cover annual holidays, public holidays, sick leave, long-service leave and superannuation contributions.

The time spent on the service delivery models

Table 32 provides an overview of the average time spent by pharmacists in the delivery of each service delivery model. The models include the supply of all three types of pharmacotherapies: Methadone, Suboxone and Subutex.

Of the six models, Model 3 was not observed as part of the study. While not all the models were observed in all the pharmacies, when they did occur, they did not systematically vary by pharmacy. In other words, all models could potentially occur in all pharmacies participating in the observational study, and the same client could receive their dose in several different ways depending on the point at which their treatment was observed.

72

FULL FINAL REPORT

Moreover, data from the observational study indicated that quantum’s of time are logically and systematically associated with different care and business models. That is not to say that the time associated with service delivery models is invariant, or even similar across or within pharmacies. As outlined in Table 32, substantial variation was evident. However logical patterns of time to undertake different service delivery models were clear.

For example, on average, service delivery models involving a combination of take away and in-pharmacy doses were the most time consuming. This is as expected, given the multiplicative clinical and administrative effort. Similarly, it would be expected that a model involving pre-preparation of doses would result in less time spent than a model involving preparation of doses at the time of service; but that the total time would be a function of the business approach in combination with the model of care required by the client at the time of service. These expectations were confirmed in the observations reported in Table 32. For example, Model 6, the delivery service delivery model including an intensive model of care (in-pharmacy) and an intensive business model (take-away dose that was not pre-prepared), was the most time consuming model observed.

The observational study found that there were discernable service delivery models for delivery of pharmacotherapy treatment in community pharmacy. The implication of this finding is central. The data indicate that it is possible to descriptively classify systematic service delivery models in community pharmacies participating in the pharmacotherapy program. Understanding service delivery functions in this way is central to developing a national funding model and provides a platform for associating systematic variation in delivery with a costing. The observational study demonstrates the complexity that will be required for a robust national funding model, ie there is wide variation between the costs associated with service delivery models.

Table 32: Average time spent by pharmacists in the delivery of each service delivery model, per occasion of service - based on time observed from the observational study

Service delivery model

Number of pharmacies in

which the model was observed

Number of observations on which this model was

based*

Average time per client / occasion of service (mins)

Mean Std

deviation Minimum Maximum Median

Model 1 6 9 5.32 6.13 1.25 17.42 3.38

Model 2 19 99 14.57 9.79 2.25 32.00 12.50

Model 3 Model not observed

Model 4 4 7 7.25 1.24 6.00 8.75 7.13

Model 5 6 34 12.87 7.90 2.25 23.75 13.99

Model 6 11 65 19.86 16.70 2.25 63.50 18.00

* Four observations were excluded from the analysis

The cost of the service delivery models

Table 33 provides the average cost for each of the service delivery models. The cost of each service delivery model was calculated using the observed time spent on the regularly recurring activities (Table 31) and the average base hourly rates outlined earlier in this section (Association of Professional Engineers, Scientists & Managers, Australia, 2009).

The cost of each service delivery model clearly parallels the observed time, given that none of the service delivery models involves devolution of the activities to different staff and is therefore a multiple of senior pharmacist time. The greatest average cost was associated with Model 6 at $16.60 per client/observation and Model 1 was the least expensive, which was on average, $4.21 per client/observation.

On further investigation of the characteristics of the pharmacies who provided Model 1 and Model 2, which provided similar models of care - the only difference being whether the doses were pre-prepared or not, yet

73

FULL FINAL REPORT

there was a substantial difference in cost – it was found that there were no key differences in the characteristics of these pharmacies (eg numbers of clients, PhARIA, State, pharmacotherapy type etc).

Table 33: Average cost, per client, of the provision of each service delivery model per occasion of service – based on time observed in the observational study and average base hourly rates for Pharmacist-in-Charge

Service delivery models

Number of pharmacies in

which the model was observed

Number of observations on which this model was

based*

Average cost per client per occasion of service ($)

Mean Std

deviation Minimum Maximum Median

Model 1 6 9 4.21 4.85 0.99 13.80 2.67

Model 2 19 99 11.81 7.48 2.57 25.35 9.90

Model 3 Model not observed

Model 4 4 7 5.74 0.98 4.75 6.93 5.64

Model 5 6 34 10.20 6.26 1.78 18.81 11.08

Model 6 11 65 16.60 13.38 1.78 50.30 14.26

* Four observations were excluded from the analysis

The cost of consumables for pharmacotherapy treatment

Table 34 provides an overview of the average cost, per week, of consumables to the pharmacy for the provision of the pharmacotherapy program, as reported by the pharmacists who participated in the activity based costing study.

The highest weekly consumable cost was for bottles which pharmacists reported as costing the pharmacy, on average, $31.36 per week. The lowest cost was related to cordial, which cost the pharmacy, on average, $2.71 per week. In total, the average weekly cost of consumables to the pharmacy was $44.92.

Table 34: Average cost, per week, of consumables to the pharmacy – based on pharmacist self-reported data

Consumable Pharmacies

(N)

Average cost per week to the pharmacy ($)

Mean Std deviation Minimum Maximum Median

Bottles 18 21.36 19.05 2.31 69.23 16.50

Labels 20 7.65 15.41 0.06 57.69 1.14

Mugs 18 2.95 2.81 0.50 9.74 1.95

Cordial 8 2.71 3.04 0.23 10.00 2.00

Others 20 6.01 17.21 0.14 80.77 2.90

Total 20 44.92 51.62 4.62 228.46 30.24

Table 35 provides an overview of the average cost, per client, for consumables. These costs are based on the weekly consumable costs presented in Table 31 and the total number of clients in the pharmacies who participated in the activity based costing study and identifies the cost of consumables, per client, for the provision of a dose. For the purposes of building the cost of consumables in the costing, it has been assumed that on average, a client will have five dosing points or occasions of service per week (with two take-away doses).

Overall, the average total cost of consumables per client was $0.62 per occasion of service. Again the cost of consumables varied widely.

74

FULL FINAL REPORT

Table 35: Average cost, per client, of consumables per occasion of service – based on pharmacist self-reported data

Consumable Pharmacies

(N)

Average cost, per client, of consumables per occasion of service ($)

Mean Minimum Maximum

Bottles 18 0.31 0.028 0.6

Labels 20 0.092 0.002 0.708

Mugs 18 0.052 0.01 0.14

Cordial 8 0.03 0.002 0.066

Others 20 0.082 0.002 0.462

Total 20 0.618 0.098 1.462

The total cost of the service delivery models based on pharmacist time and the cost of consumables

Table 36 provides the total average cost, per client, for each of the service delivery models, including the cost of consumables. The costs presented in Table 36 are based on the average cost, per client, of the delivery each service delivery model plus the average cost of consumables per client.

These costs represent what it actually costs the pharmacist to deliver the service to clients in the community pharmacy (based on time spent by the pharmacist and the cost of consumables) and would need to be factored into any future national funding model. Overall, the total average costs outlined in Table 36 indicate that the pharmacist incentive trialled in this study, of $15 per client per week, does not cover the costs of providing any one of the service delivery models on a weekly basis.

Table 36: Total average cost, per client, of the service delivery models and consumables per occasion of service – based on time observed in the observational study, average base hourly rates for Pharmacist-in-Charge and the average cost of consumables per dose

Service delivery model

Number of pharmacies in

which the model was observed

Number of observations on which this model was

based

Average cost per client of service delivery model and consumable, per occasion of service ($)

Mean Std

deviation Minimum Maximum Mean

Model 1 6 9 5.94 6.75 1.87 18.04 4.00

Model 2 19 92 15.19 10.41 2.87 32.62 13.12

Model 3 Model not observed

Model 4 4 7 7.87 1.86 6.62 9.37 7.75

Model 5 6 34 13.49 8.52 2.87 24.73 14.61

Model 6 11 59 20.48 17.32 2.87 64.12 18.62

* Four observations were excluded from the analysis

75

FULL FINAL REPORT

The cost of investments for pharmacotherapy treatment

As part of the activity based costing study, pharmacists reported on the investments directly related to the delivery of the pharmacotherapy program that had been made in their pharmacy. These costs included, for example, a shop fit-out, equipment, software, security etc.

They are noted for interest, although their place in a nationally consistent funding model is unclear and would need to be considered further. Pharmacists reported that:

• 50% had made investment costs of <$500

• 25% had made investment costs of $500-$2,000

• 25% had made investment costs of >$2,000.

6.5. Summary of pharmacist results

A key objective of this trial was to collect data about the way in which pharmacotherapy services are delivered, the costs associated with the delivery of the service and the satisfaction of community pharmacists with a nationally consistent subsidised environment. In the first instance, it was seen as desirable to develop an understanding of the service as delivered by community pharmacists, along with the true costs of the service provision.

First and foremost, the observational study indicated that the nationally consistent pharmacy incentive implemented for the trial, $15 per consented client per week, was unlikely to meet the costs of the provision of the service, even at its least costly estimate. Once the cost of consumables and investment in infrastructure was factored in, the likely shortfall of the implemented incentive was further highlighted.

The other key implication from the observational study is the ability to identify and describe relatively invariant service delivery models within community pharmacies participating in the pharmacotherapy program. These service delivery models varied by their nature, rather than by pharmacy habit. Gaining an understanding that service delivery functions in this way is central to developing a nationally consistent funding model. In addition, it was identified that there was unpredictable variation in the time taken per client interaction, which was a result of the responsiveness of the pharmacist to the needs of the client on any given day. This adds a level of complexity to the service which is analogous to the urgency/complexity features that characterise Emergency Care classification systems.

The findings from the observational study contribute significantly to the key considerations for a future funding model. Fundamentally, the observational study highlights the extent and the flexibility of client care delivered by community pharmacy, in terms of responsiveness to client need, client interaction and engagement. The high levels of retention and treatment compliance demonstrated in the trial attests to the value brought to the treatment environment, for this cohort, by community pharmacy.

Perhaps most importantly for consideration in a funding model is the impact on satisfaction of the subsidised and incentivised environment. Despite the fact that the pharmacy incentive provided in the trial was shown to be unlikely to cover the costs of providing the service, the satisfaction of pharmacists in the intervention group improved dramatically. One interpretation of this finding might be that, in the first instance, the positive impact on the dynamic of the care setting provided by the subsidised and incentivised environment (for example improved affordability), was in fact a key positive result.

76

FULL FINAL REPORT

7. Key elements of a national funding model for pharmacotherapy treatment in community pharmacy: Implications of the national trial

The present study trialled a nationally consistent funding model for the provision of pharmacotherapy treatment in community pharmacy. The funding model reflected two fundamental elements: improved affordability for clients through a subsidy of their treatment costs and improved incentives for pharmacists. The trial yielded three areas of findings central to providing recommendations concerning the key elements of a national funding model.

7.1. Implementation of a nationally consistent funding model

Contrary to previous research, the present study found that a reduction in client payments, through the nationally consistent subsidy of the treatment cost, did not result in a statistically significantly improvement in compliance with the treatment regimen, retention in treatment or broader health outcomes. The data showed that there was a high level of compliance and retention overall (a ‘ceiling effect’).

The study also found that the fees paid by clients participating in the pharmacotherapy program in a community pharmacy setting are highly variable – by jurisdiction and by pharmacy. Accordingly, a uniform client subsidy ($15 per week for each client in this instance) had a highly variable offsetting impact - in line with the variable level of cost to individual clients (‘affordability’).

Thus, a key finding of the study was the inherent difficulty of implementing a nationally consistent funding model in the Australian context. Given that context, the impact of any nationally consistent client subsidy on affordability is likely to be inconsistent because of the variable cost borne by the client to remain in treatment. The findings suggest that a key element of any future national funding model should be a client subsidy that takes into account the base cost of client payments, in order to achieve equitable financial impact across the client group.

7.2. Subsidisation of client payments: the impact on affordability for treatment for clients

The efficacy of pharmacotherapy treatment is very well documented and was not the primary focus of the present study. The primary focus was harnessing the full benefits of what is clearly a very effective treatment regimen. It is well accepted that the foundation of good treatment outcomes in pharmacotherapy treatment is the ability to retain clients in treatment and provide long-term care. The central construct underpinning the thinking behind subsidised treatment is that such arrangements increase affordability, and through that, increase access to and improve retention in treatment.

The study found that affordability is an important determinant of treatment outcome. While the uniform client subsidy did not significantly impact client outcome, the absolute level of client payments borne by the client did significantly impact outcome. The net weekly treatment cost (prior to the provision of the client subsidy) provided a way of differentiating clients based on the cost of treatment borne by them at an individual level. The level of cost associated with client payments for treatment had a significant impact on client compliance with their dosing regime, with higher cost being associated with significantly poorer compliance.

While the study did not set out to specifically examine the impact of treatment affordability on client outcomes, the tipping point for net cost of treatment to have an impact, was evident in the findings. Those clients who were paying less than an average of $29.80 per week were significantly more likely to be compliant with the treatment regimen than those who were paying more than $42.50 per week in fees.

Taken together, the findings indicate that affordability significantly impacts the ability of clients to maintain their treatment regimen in the pharmacotherapy program and that a subsidised arrangement, as part of a national funding model, is likely to reap greater benefits from the program. More specifically, the findings suggest that a key element of a national funding model would be to consider the real impact of any flat client

77

FULL FINAL REPORT

subsidy and/or possibly the case for means testing low income earners. The present results suggest that those in the top two thirds of the cost spectrum, that is where cost is greater than $29.80, are at increased risk of being unable to maintain treatment.

7.3. Reimbursement of pharmacies: how the service is delivered and the associated costs

The observational study identified relatively invariant service delivery models in community pharmacies participating in the pharmacotherapy program. In all, six service delivery models were identified, with two main dimensions driving the precise characteristics of the service delivery model.

The first dimension reflected client health need, based on the type of treatment required. These can be best thought of as different treatment ‘models of care’, based on whether the treatment regimen required in-pharmacy, take-away or some combination of treatment approaches. These models of care were found to be underpinned by a core component of the pharmacist driven business model, namely whether the doses were pre-prepared or not. The combination of models of care with business models resulted in six possible service delivery models. Moreover, data from the observational study indicated that the quantum of time for service delivery was logically and systematically associated with different care and business models.

A second key finding of the observational study was the extent and the flexibility of client care delivered by community pharmacy in terms of responsiveness to client need, client interaction and engagement. In particular, it was identified that there was unpredictable variation in the time taken per client interaction, which was a result of the responsiveness of the pharmacist to the needs of the client on any given day. This adds a level of complexity to the service which is analogous to the urgency/complexity features that characterise Emergency Care classification systems. The high levels of retention and treatment compliance obtained attests to the value brought to the treatment environment, for this cohort, by community pharmacy. The observational study corroborated much anecdotal evidence concerning the need for flexibility in practice to meet the needs of this high need client group.

Gaining an understanding that service delivery functions this way (ie there are both service delivery models and levels of complexity) is central to developing a national funding model. The data indicate that it is possible to descriptively classify systematic service delivery models in community pharmacies participating in the pharmacotherapy program. These data provide a platform for associating systematic variation in service delivery with a costing, analogous to more or less complex procedures, as the basis for costing as in other areas of healthcare delivery, eg emergency care. The observational study demonstrated the complexity that will be required for a robust national funding model. That is, there is wide variation between the time/costs associated with the service delivery models and the complexity of the client group across the sector, which will need to be considered further.

The observational study also suggests that the pharmacy incentive implemented in the trial, $15 per consented client per week, was unlikely to meet the costs of the provision of the service, even at its least costly estimate. Once the cost of consumables, and investment in infrastructure was factored in, the likely shortfall of the implemented incentive was further highlighted.

A premise for the present study was that a national funding model needs to incorporate some level of incentive for pharmacists as a key element, to retain existing service providers and to attract new pharmacies into the program. Despite the fact that the incentive provided in the trial was shown to be unlikely to cover the costs of providing the service, the satisfaction of pharmacists in the intervention group improved dramatically. From a national funding model’s perspective, this suggests that incentives for participating pharmacies do not entirely depend on precise reimbursement, and that the improved environment under the subsided and incentivised arrangements may also contribute to service provider satisfaction.

78

FULL FINAL REPORT

7.4. Strengths and limitations of the study

The great strength of this study was its design and the successful implementation of that design. The randomised controlled trial methodology provides the gold standard in health care research because of the ability to attribute influence specifically to the study factors – in this instance the impact of being in the intervention or the control group. The strength of the study design was supported by extremely complete data being collected from all participating pharmacies and clients.

The main limitations of the study emanate from the observational study. While the study was entirely robust within the parameters of the data collected, the sampling of observations was potentially incomplete. A decision, in collaboration with the Advisory Group, was taken to include three observation periods in each pharmacy at peak supply periods. The impact on pharmacy practice of different times of the day, different days of the week and simply having more observation periods are all unknown. However, the likely impact of these limitations relate to the distribution of different service delivery models: Which are most common? When are they most common? These and similar questions would be answered by more complete sampling.

From the perspective of the study objectives however, to evaluate key elements of a national funding model for pharmacotherapy treatment, the findings remain robust. They demonstrate that several discernable and systematic service delivery models are in place to provide the service and that differing levels of pharmacy effort and cost are associated with the different models.

7.5. Recommendations for key elements of a national funding model

This study has reported new data that are critical to the design of a national funding methodology for the provision of pharmacotherapy treatment in the community pharmacy. It has identified the key elements of a national funding model, as outlined below.

Client subsidy element

• Affordability is a significant determinant of treatment outcome and a subsidised arrangement, as part of a national funding model, is likely to reap greater benefits from the program.

• The precise level of the client subsidy will need to consider ‘real’ financial impact, given the variability between jurisdictions in client payment costs.

• The high frequency of financial disadvantage in this client cohort suggests a case for some sort of means testing for low income earners.

• The study results suggested that net treatment costs greater than $29.80 are associated with increased risk of being unable to maintain treatment.

Pharmacist service incentive element

• The funding model needs to incorporate some level of incentive for pharmacists, to improve their satisfaction with the program and attract new pharmacies into the program.

• In developing the service incentive, the ‘true’ costs of the delivery of the pharmacotherapy program in the community pharmacy needs to be factored in, recognising that discernable service delivery models exist and that these have systematically varying costs associated with them.

• In developing the service incentive, the flexibility and variable time commitment required to meet the complexity of and health needs of the pharmacotherapy consumer population needs to be factored in.

Pharmacy consumables reimbursement element

• Consumables were found to be an important cost element in providing the service and should be factored into a national funding model.

79

FULL FINAL REPORT

References

Amato, L., Davoli, M., Perucci, C. A., Ferri, M., Faggiano, F., & Mattick, R. P. (2005). An overview

of systematic reviews of the effectiveness of opiate maintenance therapies: Available evidence

to inform clinical practice and research. Journal of Substance Abuse Treatment, 28, 321-329.

American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental Disorders

(Fourth Edition, Text Revision ed.). Washington, DC: American Psychiatric Association.

Association of Professional Engineers, Scientists & Managers, Australia (2009). Community and

Hospital Pharmacists’ Remuneration Survey Report.

Australian Institute of Health and Welfare. (2008). National opioid pharmacotherapy statistics

annual data collection: 2007 report. Canberra: Author.

Ball, J. C., & Ross, A. (1991). The effectiveness of methadone maintenance treatment: Patients,

programs, services, and outcome. Vienna: Springer-Verlag.

Ball, J. C., Shaffer, J. W., & Nurco, D. N. (1983). The day-to-day criminality of heroin addicts in

Baltimore - a study in the continuity of offence rates. Drug and Alcohol Dependence, 12, 119-

142.

Bammer, G., Battisson, L., Ward, J. & Wilson, S. (2000). The impact on retention of expansion of

an Australian public methadone program. Drug and Alcohol Dependence, 58(1–2): 173–180.

Barnett, P. G., & Hui, S. S. (2000). The cost-effectiveness of methadone maintenance. Mount Sinai

Journal of Medicine, 67, 365-374.

Barnett, P. G., Zaric, G. S., & Brandeau, M. L. (2001). The cost-effectiveness of buprenorphine

maintenance therapy for opiate addiction in the United States. Addiction, 96, 1267-1278.

Breen, C., Harris, S., Hawken, L., Mattick, R., Lintzeris, N., Ritter, A., & Mendoza, E. (2003).

Cessation of methadone maintenance treatment using buprenorphine: Transfer from

methadone to buprenorphine and subsequent buprenorphine reductions. Drug and Alcohol

Dependence, 71(1), 49-55.

Caplehorn, J. R. M., Dalton, M. S. Y. N., Cluff, M. C., & Petrenas, A. M. (1994). Retention in

methadone maintenance and heroin addicts' risk of death. Addiction, 89, 203-207.

Caplehorn, J. R. M., Dalton, M. S. Y. N., Halder, F., Petrenas, A. M., & Nisbet, J. G. (1996).

Methadone maintenance and addicts' risk of fatal heroin overdose. Substance Use and Misuse,

31, 177-196.

80

FULL FINAL REPORT

Chalmers, J., Ritter, A., Heffernan, M., & McDonnell, G. (2008). Pharmacotherapy maintenance in

Australia: exploring affordability, availability, accessibility and quality using a systems dynamic

modelling approach. Sydney: National Drug and Alcohol Research Centre.

Commonwealth Department of Health and Ageing. (2004). National Evaluation of

Pharmacotherapies for Opioid Dependence. Canberra: Author.

Darke, S., Degenhardt, L., & Mattick, R. (2007). Mortality amongst illicit drug users. Melbourne:

Cambridge University Press.

Degenhardt, L., Rendle, V., Hall, W., Gilmour, S., & Law, M. (2004). Estimating the number of

current regular heroin users in NSW and Australia 1997-2002. Sydney: National Drug and

Alcohol Research Centre.

Dole, V. P., & Joseph, H. J. (1978). Long-term outcome of patients treated with methadone

maintenance. Annals of the New York Academy of Sciences, 311, 181-189.

Fleming, G. F., McElnay, J. C., Hughes, C. M., Sheridan, J., & Strang, J. (2001). The role of

community pharmacist in drug abuse: A comparision of servce provision between Northern

Ireland and England/Wales. Pharmacy World and Science, 23, 13-16.

Hall, W., & Wodak, A. (1999). Is naltrexone a cure for heroin dependence? The evidence so far is

not promising. Medical Journal of Australia, 171, 9-10.

Harwood HJ, Hubbard RL, Collins JJ, Rachal JV. The costs of crime and the benefits of drug

abuse treatment: a cost-benefit analysis using TOPS data. In: Leukefeld CG, Time FM (eds.).

Compulsory Treatment of Drug Abuse: Research and Clinical Practice. NIDA Research

Monograph Series 86. Rockville, MD: National Institute on Drug Abuse, 1988.

Harwood, H. J., Malhotra, D., Villarivera, C., Liu, C., Chong, U., & J., G. (2002). Cost effectiveness

and cost benefit analysis of substance abuse treatment: an annotated bibliography. Falls

Church Va.: Centre for Substance Abuse Treatment.

Lind, B., Chen, S., Weatherburn, D., & Mattick, R. (2005). The effectiveness of methadone

maintenance treatment in controlling crime. British Journal of Criminology, 45, 201-211.

Maddux, J. F., Prihoda, T. J., & Desmond, D. P. (1994). Treatment fees and retention on

methadone maintenance. Journal of Drug Issues, 24, 429-443.

Matheson, C., Bond, C. M., & Mollison, J. (1999). Attitudinal factors associated with community

pharmacists' involvement in services for drug misusers. Addiction, 94, 1349-1359.

81

FULL FINAL REPORT

Matheson, C., Bond, C. M., & Tinelli, M. (2007). Community pharmacy harm reduction services for

drug misusers: National service delivery and professional attitude development over a decade

in Scotland. Journal of Public Health, 29, 350-357.

Mattick, R. P., Ali, R., White, J. M., O'Brien, S., Wolk, S., & Danz, C. (2003). Buprenorphine versus

methadone maintenance therapy: a randomized double-blind trial with 405 opioid-dependent

patients. Addiction, 98, 441-452.

Mattick, R. P., Breen, C., Kimber, J., & Davoli, M. (2004). Methadone maintenance therapy versus

no opioid replacement therapy for opioid dependence (Cochrane Review). The Cochrane

Library Issue 3.

Mattick, R. P., Breen, C., Kimber, J., & Davoli, M. (2008). Methadone maintenance therapy versus

no opioid replacement therapy for opioid dependence. The Cochrane Database of Systematic

Reviews(Issue 2).

Mattick, R. P., Kimber, J., Breen, C., & Davoli, M. (2008). Buprenorphine maintenance versus

placebo or methadone maintenance for opioid dependence. The Cochrane Database of

Systematic Reviews, Issue 2.

Murphy, S., & Rosenbaum, M. (1988). Money for methadone: II. Unintended consequences of

limited-duration methadone maintenance. Journal of Psychoactive Drugs, 20, 397-402.

New York State Committee of Methadone Program Administrators. COMPA’s Five Year Plan for

the Methadone Treatment System in New York State. New York: New York State Committee of

Methadone Program Administrators, 1991.

Ritter, A., & Chalmers, J., (2009). Polygon: the many sides to the Australian opioid

pharmacotherapy maintenance system. Sydney: National Drug and Alcohol Research Centre.

Rosenbaum, M., Irwin, J., & Murphy, S. (1988). De facto destabilization as policy: The impact of

short-term methadone maintenance. Contemporary Drug Problems, 15, 491-517.

Rosenbaum, M., UIrwin, J., & Murphy, S. (1987). Money for methadone: Preliminary findings from

a study of Alameda County's new maintenance policy. Journal of Psychoactive Drugs, 19,

13-19.

Ross, J., Teesson, M., Darke, S., Lynskey, M., Ali, R., Ritter, A., & Cooke, R. (2005). The

characteristics of heroin users entering treatment: Findings from the Australian Treatment

Outcome Study (ATOS). Drug and Alcohol Dependence, 24, 411-418.

Rowe, J. (2008). A raw deal? Impact on the health of consumers relative to the cost of

pharmacotherapy. Melbourne: Centre for Applied Social Research (RMIT).

82

FULL FINAL REPORT

Rufener BL, Rachal JV, Cruz AM. Management Effectiveness Measures for NIDA Drug Abuse

Treatment Programs. Cost Benefit Analysis. Rockville, MD: National Institute on Drug Abuse,

1977.

Shanahan, M., Hetherington, K., Mattick, R. P., & Weatherburn, D. (2007). Estimating the cost-

savings of reduced crime while in methadone treatment. Sydney, NSW: National Drug and

Alcohol Research Centre, University of New South Wales.

Sheridan, J., Manning, V., Ridge, G., Mayet, S., & Strang, J. (2007). Community pharmacies and

the provision of opioid substitution services for drug misusers: Changes in activity and attitudes

of community pharmacists across England 1995-2005. Addiction, 102, 1824-1830.

Ward, J., Hall, W., & Mattick, R. P. (1999). Role of maintenance treatment in opioid dependence.

Lancet, 353, 221-226.

Ward, J., Mattick, R. P., & Hall, W. (1992). Key issues in methadone maintenance treatment.

Sydney: New South Wales University Press.

Ward, J., Mattick, R. P., & Hall, W. (Eds.). (1998). Methadone maintenance treatment and other

opioid replacement therapies. London: Harwood Press.

Winstock, A., Lea, T., & Molan, J. (2008). NSW community pharmacy survey: SWAT report 2 (Vol.

297). Sydney: National Drug and Alcohol Research Centre.