20
1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23 rd -25 th of June 2014 Increasing the personalisation of social care in England: factors associated with variations in local patterns 12 Authors: Jose-Luis Fernandez, Valentina Zigante © PSSRU, London School of Economics and Political Science Keywords: social care, personalisation, direct payments Abstract Personalisation has been a defining trend in English social care provision, exemplified through the increasing focus on personal budgets and direct payments. This paper explores the demand, supply and other factors associated with patterns of local variability in coverage and intensity of direct payments in English local authorities. We use local authority data on extensiveness of services, expenditure/cost and local circumstances such as age structure, health and deprivation, for England from 2005 to 2012. Statistical analysis was conducted for four client groups; older people, people with physical disabilities, people with learning disabilities and people with mental health needs. We find a steady increase in the number direct payments relative to other service types, however with considerable variation between local authorities, partly explained by the local needs profile. Direct payments are further found to be driven by the cost of alternative services and overall rationing of services. 1 Work in progress. Please do not quote, cite or distribute. 2 The work was funded by the Department of Health to the Policy Research Unit in the Economics of Social and Health Care. The views expressed are those of the authors and not necessarily those of the funders.

Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

1

Paper for Health Economists' Study Group (HESG) Meeting

Glasgow Caledonian University 23rd

-25th

of June 2014

Increasing the personalisation of social

care in England: factors associated with

variations in local patterns12

Authors: Jose-Luis Fernandez, Valentina Zigante ©

PSSRU, London School of Economics and Political Science

Keywords: social care, personalisation, direct payments

Abstract

Personalisation has been a defining trend in English social care provision, exemplified

through the increasing focus on personal budgets and direct payments. This paper

explores the demand, supply and other factors associated with patterns of local

variability in coverage and intensity of direct payments in English local authorities.

We use local authority data on extensiveness of services, expenditure/cost and local

circumstances such as age structure, health and deprivation, for England from 2005 to

2012. Statistical analysis was conducted for four client groups; older people, people

with physical disabilities, people with learning disabilities and people with mental

health needs. We find a steady increase in the number direct payments relative to

other service types, however with considerable variation between local authorities,

partly explained by the local needs profile. Direct payments are further found to be

driven by the cost of alternative services and overall rationing of services.

1 Work in progress. Please do not quote, cite or distribute.

2 The work was funded by the Department of Health to the Policy Research Unit in the Economics of

Social and Health Care. The views expressed are those of the authors and not necessarily those of the

funders.

Page 2: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

2

Introduction

Personalisation is a denoting feature across public services, exemplified by the

emphasis on individual decision making in aspects of health care such as choice of

hospital and treatment. More recently this has extended through the introduction of

personal health budgets (Forder et al. 2012). Social care is one of the public services

areas where personalisation is most relevant due to the importance of individual

preferences in the design of ideal care packages. Direct payments forms a key

component in the application of the broader ‘personalisation agenda’ governing the

reform of public services. This agenda has been promoted by successive governments

to encourage service users themselves to take greater control of commissioning

(Glendinning et al. 2008). Care services in England can be thought to have been

transformed by two phases of marketisation, with a shift in the provision of services

from local authorities to private organisations and, more recently, a shift from local

authority to individual purchasing of care through the allocation of personal budgets3

and direct payments instead of services in kind (Glendinning 2012).

Direct payments legislation was implemented in April 1997 in England, Wales and

Scotland and a year later in Northern Ireland. The policy framework enabled local

authorities4 and health and social service trusts to make cash payments to users to

purchase services on their own behalf rather than using those supplied and organised

by the local authority. The user has virtually full autonomy in what is purchased for

the direct payment; for example a hiring a personal assistant, paying an informal

caregiver (with restrictions) or purchasing food services. In order to qualify for a

direct payment, an individual has to be assessed as being eligible for community care

services (Department of Health 2000). Access was initially restricted to users between

the ages of 18 and 65 but was later extended to other client groups.5 The emphasis on

3 Direct payments are closely related to personal budgets, however while direct payments are monetary

payments made by local authorities directly to service users, in lieu of social service provisions, a

personal budget, on the other hand, is a notional amount of social care funding for an individual’s

support, some or all of which may be taken as a direct payment. 4 In England social care is organised at the local authority or council area level by Councils with Adult

Social Services Responsibilities (‘CASSR’). As is the common practice we refer to these as ‘local

authorities’ (LA) throughout this paper. 5 The government policy and guidance documents on Direct Payments have been continually amended

and altered over the past ten years in an attempt to widen the eligibility criteria for Direct Payments and

increase take up (see for example Department of Health 2000).

Page 3: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

3

personalisation has continued since the implementation of the Direct Payments act –

in later years through the introduction of personal budgets.

Personalisation is valued for its benefits in terms of improved outcomes for users, a

more efficient market structure and an attractive way of managing services for local

authorities. For users the main advantages are the opportunity to have care

preferences met, gain autonomy and gain access to services which may not be

available through local authority provision. Similarly, more active and ‘mobile’ users

create impetus for more responsive and efficient behaviour among providers, in order

to remain competitive and attract users who are more able to ‘vote with their feet’.

The care market is further expanded, beyond the traditional conception of a mixed

economy of care – including outsourced provision and a strong role for voluntary

provision (Glendinning et al. 2008) – to include new types of care provision such as

friends and family and local pubs instead of meals on wheels. This can be particularly

beneficial in areas where formal supply is limited, such as in sparsely populated, or

expensive (e.g. in London) areas. The advantages for local authorities lie in the

prospective for rationing services – i.e. offering smaller care packages, compared to

market value, with the expectation that individuals are able to identify more cost-

efficient sources of care (for example paying an informal caregiver). This is facilitated

by the substantial discretion in the allocation of DPs on the part of care managers

(Ellis 2007; Ellis 2011).

Even though direct payments are generally seen as improving the user experience and

service outcomes, some concerns are worth noting. Individual users may struggle with

lacking information about the market in order to make appropriate commissioning

decisions. Similarly, individuals have less market power than local authorities which

implies a change in the balance of power between providers and purchasers (in this

case the individual). The hiring of personal assistants can place significant transaction

costs on service users due to the legal and administrative duties associated with

becoming an employer. The severity of these concerns is linked to the relative ability

among users to cope with the management and commissioning of their own care.

Partly as a result of differences in their motivations and abilities and the availability of

support for the management of care package from family and friends, research has

Page 4: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

4

found substantial differences in take-up between client groups (Priestley et al. 2006;

Glendinning 2008).

At an individual level, a range of factors are likely to drive the success of the direct

payment model (for a general discussion see the Individual Budgets evaluation,

largely applicable also to DPs: Glendinning et al. 2008). Firstly, a larger care package

may be more attractive in terms of reducing the relative impact of the transaction

costs related to setting up and learning the system. Likewise, it is important that

market structures are in place which support the individuals as commissioners, in

dealing with care providers, and that the allocation and sign-off processes are not

overly long or bureaucratic (Holloway et al. 2011). As mentioned above, individual

ability to manage a payment matters (depending on informal support, cognitive

impairment etc.), but also the wishes and preferences of service users and their

families. Self-determination is unlikely to be equally attractive to all service users,

particularly in terms of being involved in the commissioning process. Different

individuals will also require different levels of support, which targeting of models will

be key to alleviate (Glendinning et al. 2008).

The use of direct payments has increased steadily across the country since the

introduction. The overall number of people who receive services on the other hand,

has contracted since 2005 and the rate of the fall increased after 2010 (Fernandez et

al. 2013). The proportion of the population covered by services dropped by just over

30% between 2005 and 2012 and it is clear that this is not due to decreasing need; in

fact, research indicates increasing need due to generally acknowledged demographic

and social change (Wittenberg et al. 2011). Figure 1 illustrates the change in number

of users of community and nursing home services.

Page 5: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

5

Figure 1 Number of community care clients and total number of clients 2005-2012, 1000s

Source: data from RAP collection from the health and social care information centre.

Direct payments on the other hand exhibit a rapid growth since 2005. DP’s as a

proportion of all services has increased from 3% of services to 17% between 2005 and

2012. It is important to remember that although DPs is the policy type increasing the

most; it only services a minority of users. This is particularly evident for certain user

groups; among older people and users with mental health needs on average around

10% of users receive a direct payment (see figure 2). Users in the group ‘Physical

disability’ on the other hand use DP to the highest extent. The difference between the

groups has remained fairly constant over the period, with no evidence of MH and OP

usage catching up. The growth in DP usage among older people has flattened out

somewhat between 2011 and 2012.

Previous work found distinct variation between local authorities in the extent to which

DP:s are utilised and also considerable variation between client groups; older people,

people with physical disabilities, people with mental health issues and finally, people

with learning disabilities (Fernandez et al. 2007; Leece & Leece 2006). The

differences between the user groups are also well known in the literature. For example

the particular role of the young disabled, in driving the introduction of DPs as

discussed above, resulted in the high usage rates also in our time period (Glasby &

Littlechild 2009).

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

2005 2006 2007 2008 2009 2010 2011 2012

Community care clientsCommunity care, residential and nursing clients

Page 6: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

6

Figure 2 Number of direct payments - proportion of community care users, by client group

Note: RAP data 2005-2012. Client groups: OP ‘older people’, MH ‘mental health’, PD ‘physical

disability’, LD ‘learning disability’.

This paper explores the factors associated with direct payments usage in the four

client groups between 2005 and 2012. The period is denoted by direct payment being

fairly well established (past the main implementation costs and early glitches) but also

by fiscal austerity being a key concern for the provision of social care in England. We

attempt to identify the role direct payments play as a part of the care offer of local

authorities, in relation to local need and economic and social characteristics, service

supply and rationing of services. We take into account how the allocation of services

takes place in a complex system of care provision, in which care managers play a key

gatekeeping role, with decisions constrained by local supply and commissioning

processes.

The paper is structured as follows; a brief background section first elaborates the

conceptual model of local care provision followed by a discussion of the data used,

the hypothesised relationships and econometric methods. The results are then outlined

followed by a discussion and conclusions.

Utilisation of direct payments – conceptual model

In the context of the general contraction of services we find an increase in DP

utilisation, however, there is as shown in figure 2, a wide disparity between client

groups, and as shown in figure 3, between local authorities. Each line in figure 3

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

2005 2006 2007 2008 2009 2010 2011 2012

OP MH PD LD

Page 7: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

7

represents the 149 local authorities in terms of DP utilisation as a proportion of total

services. The local authorities are ranked by proportion for each client group, hence

the order varies accordingly. For example the client group ‘physical disability’ the

proportion varies from 10% to 90% of community services in 2012. The client group

‘older people’ has the least variation between local authorities, from 3 to 35%.

Figure 3 Variation in local authority utilisation of direct payments packages 2012

Note: Direct payments packages, as a proportion of community care services from RAP 2012/2013, by

client group.

The policy architecture underlying the distribution of DP clients relative to other

services as a whole is complex (see Fernandez et al. 2007) Given these complexities

and the remarkable difference is DP utilisation we now move to conceptually explore

the underlying structures and hypothesised influential variables which we will later

test empirically. By law, all users should be offered the option to receive their care

package as a direct payment (Glasby & Littlechild 2009). In practice however, the

process of deciding on a care package is likely to be a function of a wide range of

factors influencing both the user’s and the care manager’s preferences and constraints.

The decision making process can be likened to a bargaining process, in which the

negotiation between the user and the LA can result in a variety of care outcomes. In

figure 4 the decision making process is illustrated; from the initial contact to an

implemented care package. The various decisions come with a specific set of

considerations and negotiation between user and local authority representative

determines the final care allocation.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

OP MH PD LD

Page 8: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

8

Figure 4 Model of care decision making – stages and options for individuals with care need

Source: author’s own

At the first stage: the initial contact between individual and the LA, the individual is

often in a situation where the need for (additional) care has become acute. This

contact often takes place at a time of crisis, for example after the individual suffering

from an adverse health event, such as a fall. Family and friends tend to be

instrumental in making this contact. Starting from the time of the user with care need

being assessed by the local authority, the first decision is whether the individual is

eligible – i.e. meeting the need and means tests.6 In many cases the individual him or

herself will fund all or part of the care needs (henceforth self-funded). This paper is

focused on the proportion of services being provided through the local authority;

hence excluding self-funded care for which to date little data is available. In terms of

the care choice, nursing or residential care will seem attractive if needs are substantial

or it is difficult for the user to stay at home for other reasons. This is commonly the

6 The specifics of the means-test is set at the national level (currently at £23,250 in capital for home

care) whereas the eligibility (needs test) varies significantly. The Fair Access to Care Services (FACS)

guidelines in 2003 set up a common framework against which to assess needs and set local eligibly

thresholds for the provision of care. Many authorities have however tightened eligibility thresholds

such that only those with critical or substantial levels of need are entitled to receive publicly-funded

care (Fernandez et al. 2013).

Page 9: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

9

case for users with dementia or with extensive health care needs, possibly in

combination with a lack of informal support.

If the user is able to remain at home – which is the preferred option in line with the

‘ageing in place’ (Hillcoat-Nalletamby & Ogg 2013) discourse – numerous care

options become available; for example home care (directly provided), day care and

meals.7 The choice between direct payments and directly provided services (services

in kind) is at the focal point of this paper. Following this, we define our dependent

variable as the number of clients receiving their allocated care as a direct payment,

relative to the overall total number of clients receiving LA provided community care.

Through this we attempt to standardise for the needs profile of each LA, as we are

considering the care allocation for users who have already been deemed eligible for

care, and equally for DPs.

In terms of the choice between DP and directly provided services a few aspects

beyond those alluded to in the introduction are worth exploring further. In terms of

care users’ preference and abilities it has been found that relatively articulate and

‘forceful’ users, or those with a ‘champion’, such as a family member or engaged care

manager are more likely to use DPs. This may be reinforced by uncertainties held by

front line staff about the suitability of direct payments particularly impacting on users

with mental health needs, dementia or learning difficulties (Priestley et al. 2006).

Further, the implementation of direct payments involves numerous practical,

organisational and cultural challenges for staff at all levels in local authorities. The

care manager has a key role a gatekeeper (Ellis 2011) and makes an assessment of the

risk of relinquishing control to the user – whether the user has the adequate support

and ability to suitably manage his or her care (Milner & O’Byrne 1998). Foster et al.

found that practitioners’ awareness of the constraints and limitations of the resource

context influenced their assessment process and their decision-making. At times

constraints and limitations stood above the needs of service users and could be a

barrier to inclusive decision making and users receiving the appropriate care (2006).

7 Community-based services are services provided to people who live in their own homes (this includes

supported/sheltered accommodation and extra care housing). Clients living in a home setting can

receive a number of community-based services. Subsets of these services or components of service

include: Home care, Day care, Meals, Short-term residential – not respite, Direct payments and

personal budgets, Equipment and adaptations, Professional input and support, Other.

Page 10: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

10

For our analysis this is especially relevant if there are systematic differences in

preferences and behaviour of care managers between local authorities.

Data and methods

This paper uses panel data from 2005 to 2012, obtained from official sources at local

authority level (n=147 2005-2008, n=149 2009-2012). Table 1 outlines the variables

used and sources of the data. These are discussed in detail below – outlining the

hypothesised relationships.

Table 1 Description of variables

Category Variable Source

Local need Health status Census

Limiting long-standing illness Census

Deprivations scores Census

Mortality ONS

Receipt of Social Security benefits DWP

Population profiles ONS

Informal support Census

Service provision Number of direct payments RAP forms, DoH

Number of clients of other services RAP forms, DoH

Residential care utilisation RS forms, DoH

Home care utilisation RS forms, DoH

Expenditure in social care services RS forms, DoH

Supply (Unit costs) Home care unit costs PAF forms, DoH

Local wages New Earnings Survey

House prices Land Registry

Area, Density ONS

Local performance Satisfaction with services

PAF forms, DoH/ASCS

survey 2011

Political control Electoral commission

The set of local need controls included in the models aim, in various ways, to control

for the characteristics of the user population and the way that they are likely to

interact with social care service provision.

We assume indicators of population health to be correlated with the need for social

care. This includes indicators of deprivation which is known to be closely related to

poor health among the population. On the other hand, DPs come with an expectation

of service users to manage their care. Users with poor health, beyond their disability,

Page 11: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

11

may be less inclined to demand DPs, and likewise care managers may be less inclined

to offer. Our disability indicators, on the other hand, are population averaged limiting

long-standing illness and proportions receiving disability benefits (AA and DLA). We

expect a close link between service provision and indicators of disability. In terms of

the effect on DP utilisation, users with more extensive need may receive a larger care

packages. Theoretically this may cause care users to be more prone to accept a DP

instead of directly provided care due to the reduced impact of transaction costs

(economies of scale). Similarly, we expect care managers, facing budget pressures to

favour DPs for large care packages, which would otherwise be expensive to provide

directly.

It is also well known that provider supply and the broader care market conditions will

influence provision decisions. Unit costs for alternative care services, such as home

care, depending on local wages and population density, will influence the cost of

providing services. High costs are likely to shift utilisation in favour of DPs where it

is expected that the care user is able to source care in a more cost-effective manner,

potentially through paying informal care givers. Informal care is a cheap and efficient

way of utilising a DP – however subject to limitations: users are technically not

allowed to hire co-residential family members, but this regulation is implemented at

care managers’ discretion. Care managers may be reluctant to offer a DP when a well-

established informal caregiving arrangement is in place, but may in any case be able

to offer a smaller care package to allow respite for the caregiver. Particularly, for the

client group LD (learning disability) this may be a common practice as parents of LD

users provide informal care while also managing a DP.

Finally, we hypothesise that service provision variables are key for the relative

utilisation of DPs. One of the assumed key driving forces of the shift in provision

towards DPs, in the current economic climate, is rationing to meet shrinking budgets.

An LA which covers a relatively large proportion of its residents may face higher

need (which we can control for in models) but is likely to be wealthy and generous

and can provide services directly, instead of offering DP. This assuming that care

managers in wealthy LAs does not have a particular preference for DPs.

Our ability to draw conclusions from the analysis is hampered by the limitations of

the data. We estimate reduced form models, in which we are not able to disentangle

Page 12: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

12

demand and supply. As the care utilisation is a result of a negotiation, taking into

account the preferences of both service user and LA, our control variables affect both

actors, potentially in contradictory ways. Similarly, as the data is collected at LA

level, what we are considering are the effects of population average of for example

health status or disability capturing need. This does not account for the severity of the

need, simply that a certain proportion of the population has self-rated as suffering

from a limiting long-standing illness or poor health. The strength of our interpretation

and precision of our estimates is affected.

Data is collected yearly from 2005 to 2012. 2005 covers the year 2005/2006 in terms

of the RAP (Referrals, Assessments and Packages of care) collection including

detailed data on social care provision. The majority or variables are proportions of the

relevant population – for example poor health status for older people is entered into

the regressions as the proportion of the total number of older people who have rated

their health as ‘poor’ in the Census questionnaire. The Census variables (2001 and

2011) as well as deprivation (2004, 2007 and 2010) are only available for a limited

number of years throughout the period these have been linearly intrapolated. The

dataset has 1,184 observations distributed over the 8 years of observation. The panel

is unbalanced due to the split of LA Cheshire into Cheshire East and Cheshire West &

Chester, and Bedfordshire into Bedford and Central Bedfordshire in 2009. The RAP

data collection has some missing values which have where possible been estimated.

The RAP recording of number of direct payments users changed in 2009/2010 –

essentially the direct payments were merged with the recording of personal budgets –

but reverted back to the old practice in 2010/2011. This means that the figures for

2009/2010 are not comparable with prior or later data (Health and Social Care

Information Centre 2012). The 2009 data was intrapolated and all models have been

run without the year included. Further, excluded categories are the substance misuse

client group which is generally not available for direct payments and similarly, three

small and atypical local authorities; City of London, Isles of Scilly and Isles of Wight

have also been excluded.

To account for the non-normality in the distributions of our key dependent

variable (the number of clients receiving their allocated care as a direct payment,

relative to the total number of clients receiving LA provided community care) we

Page 13: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

13

employ generalised linear modelling (GLM) estimation methods taking into

account the panel structure of the data, with the appropriate variance function

chosen following the methodology proposed by Manning and Mullahy (2001).

We assume an AR1 correlation, i.e. that observations closer in time are more strongly

correlated. We observe collinearily between the aggregate area level indicators and

hence we approach the model building in steps by firstly focusing on need indicators,

followed by informal care and then social care variables. Particular attention is paid to

a range of LA level social care indicators of overall population coverage of care,

expenditure and cost of services (unit costs). The social care variables are closely

correlated and there are issues with endogeneity; the overall and variety of the care

offer is a function of the user demand, LA preferences and care supply as discussed

above.

Results

We initially consider a model controlling for need (demand) and supply exclusively –

the results are reported in table 2. Need is a key control variable as it will strongly

influence the coverage of the population each LA can afford (as the cost of care

packages will increase when needs are more severe). Moreover, the level of need

among the covered population will influence whether DPs are an appropriate care

option.

The proportion of limiting longstanding illness positively influences the usage of DPs

while the proportion reporting poor health status has a negative relation with DP

usage. These counterintuitive results – we would expect a higher proportion of poor

health would lead to a higher usage of DPs – may be explained by the historic role of

the disabled (PD client group) as the main usage group of DPs. It is not health status

that is driving DP usage – but rather disability. Similarly it is conceivable that self-

rated health status may differ qualitatively from (self-rated) limiting longstanding

illness. Life expectancy at age 65 is positive, which supports the idea that it is the

distinction between disability and health which is driving the need effect in the

models.

Page 14: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

14

Table 2 Panel glm regressions, dependent variable DP as proportion of community care users, by

client group – need and supply

Client group Older people Physical

disability Mental health

Learning

disability

Poor health -10.585***

Poor health -33.331*** -53.012*** -42.575***

Life expectancy at age 65 0.324*** 0.060 0.249*** 0.179***

Limiting longst. illness 3.775***

Limiting longst. illness 1.430** 1.415 3.028***

AA 10.333***

DLA 23.513*** 50.015*** 49.150***

Weekly earnings (average) 0.001** 0.000 0.000 0.001***

Population density 0.068** 0.081*** 0.105*** 0.092***

Weekly cost, home care

provision 0.188** 0.100*** 0.262** 0.120**

Constant -12.884*** -3.201*** -9.490*** -7.395***

Number of observations 1147 1172 1148 1172

Notes: gamma, AR1 correlation structure a) Proportion of population age 65+

b) Proportion of population age 18-64

AA= attendance allowance (for individuals over age 65)

DLA = disability living allowance (for individuals under age 65)

Need is further captured through the proportion of the population receiving attendance

allowance (AA) and disability living allowance (DLA), which both have a positive

impact on DP usage. These variables are also closely linked to disability and need for

services rather than overall health. Other indicators of need with weaker relationship

with DP usage are social security benefits – here pension credit for OP and income

support for the other client groups (under age 65). These variables are related to

deprivation (also tested in the models) in that they capture the expected coverage of

social care services overall (the proportion of the population the LA is required to

serve based on the means-test). Including deprivation, the variables are also correlated

with need (health status and disability). These variables however tend to be

insignificant in models where need is controlled for through the variables discussed

above – with disability indicators having the strongest predictive power.

Moving to the supply side, also reported in table 2, our key indicators of the cost and

ease of provision are wages; population density and home care unit costs. Wages and

population density more broadly capture the price level in the local authority, and

implicitly the cost of providing various types of care. The positive sign on weekly

earnings and home care unit cost indicate that DPs are substituted when other types of

care are likely to be relatively expensive. Density on the other hand indicates that

Page 15: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

15

there is a positive relation between DP utilisation and more urbanised areas. We

expect DPs to be common where the spread of care users implies that formal services

are relatively difficult to provide. On the other hand, this indicates that density

captures the cost of alternative services, which is higher in urban areas potentially

accounting for the positive effect of density.

Linked to the supply side and a key complement to formal social care provision is the

availability of informal care. Results from models incorporating informal care

variables are shown in table 3. Informal care is also (endogenously) related to demand

for formal care services and hence potentially DPs as a part of care services. We find

an overall negative effect when adding an indicator of informal care supply (as a

proportion of population) to our models – across the client groups.

Table 3 Panel glm regressions, dependent variable DP as proportion of community care users, by

client group – informal care

Client group Older people Physical

disability Mental health

Learning

disability

Poor health -15.648***

Poor health -30.152*** -50.005*** -40.299***

Life expectancy at age 65 0.236** 0.079** 0.272*** 0.154***

Limiting longst. illness 1.853**

Limiting longst. illness 1.662** 0.384 2.628**

AA 6.997**

DLA 14.708*** 19.993** 26.133***

Weekly earnings (average) 0.001*** 0.000 0.001 0.001***

Population density 0.036 0.017 0.058** -0.002

Weekly cost, home care

provision 0.046** 0.084** 0.271*** 0.108**

Informal care (1-19 hours) -35.334**

Informal care (20-49 hours) 153.763**

Informal care (50+ hours) -10.652

Informal care (1-19 hours) -13.954*** -14.760** -21.345***

Informal care (20-49 hours) 36.762** 27.221 83.319***

Informal care (50+ hours) 0.753 74.984** 14.648

Constant -7.578*** -2.518*** -9.690*** -5.594***

Number of observations 1147 1172 1148 1172

Notes: gamma, AR1 correlation structure a) Proportion of population age 65+

b) Proportion of population age 18-64

AA= attendance allowance (for individuals over age 65)

DLA = disability living allowance (for individuals under age 65)

As DP regulations prevent paying a co-residential informal caregiver this is the

expected outcome. The same results were found for client groups LD and MH in the

previous study (Fernandez et al. 2007). If informal care is already available, this is

taken into account in the care assessment, and DPs are less likely to be offered.

Page 16: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

16

However, when considering intensity of informal care (number of hours per week) we

find that the 20-49 hour per week bracket has a positive effect on DP uptake while the

19 hours and below drives the overall negative effect identified. This is puzzling as

the higher the number of hours; the more likely is care is to be co-residential and

hence not generally covered by DPs. It is however possible that where directly

provided services are expensive; DPs are offered also where informal care is present

and potentially co-residential. For mental health users the positive effect is at the 50

hours and above bracket which may be explained by the potentially limited ability of

MH users to manage a DP without intensive support for an informal carer.

Finally, we directly consider indicators of spending and provision in LAs. Table 4

shows the results of the analysis, again across user groups.

Table 4 Panel glm regressions, dependent variable DP as proportion of community care users, by

client group – coverage and unit costs

Client group Older people Physical

disability Mental health

Learning

disability

Poor health -13.282***

Poor health -30.573*** -59.285*** -42.539***

Life expectancy at age 65 0.237** 0.083** 0.169* 0.086*

Limiting longst. illness 1.692*

Limiting longst. illness 1.341* 0.342 2.288**

AA 3.756

DLA 13.128*** 16.226* 22.625***

Weekly earnings (average) 0.001** 0.000 0.001 0.001***

Population density 0.015 0.022 0.082** 0.003

Weekly cost, home care

provision 0.083* 0.019* 0.201* 0.087**

Informal care (1-19 hours) -36.016**

Informal care (20-49 hours) 164.067***

Informal care (50+ hours) -17.192

Informal care (1-19 hours) -12.523*** -12.139* -21.827***

Informal care (20-49 hours) 32.770* 15.001 72.859***

Informal care (50+ hours) 8.431 88.724** 23.938

Change in coverage

(proportion of population

receiving services) -0.003** -0.006*** -0.006*** -0.002***

Constant -7.375** -2.415*** -7.159*** -3.866***

Number of observations 1147 1172 1148 1172

Notes: gamma, AR1 correlation structure a) Proportion of population age 65+

b) Proportion of population age 18-64

AA= attendance allowance (for individuals over age 65)

DLA = disability living allowance (for individuals under age 65

Page 17: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

17

The variable year on year change in coverage is negative across the user groups –

indicating that DPs are more common in LAs which are contracting services and

providing to a smaller proportion of the population. The results are identical if

including the level of coverage or a lagged (one or more years) indicator of level of

coverage. These local authorities are thought to cut low-need users and focus on the

neediest as a method of reducing cost. This process may in itself incentivise more

direct payments as care packages are larger and therefore more attractive relative to

transaction costs. DP will further seem attractive when care managers are under

pressure to contain spending as the care packages are generally cheaper than the

equivalent value of direct services.

We note that population density is not significant when controlling for informal care,

which is linked to household size and in turn density. Further, the effect of unit costs

is weakened when introducing coverage, the positive sign is consistently maintained.

A range of variables, theoretically potentially related to DP usage, have been included

but found to be insignificant in the regressions. In particular, political party in power

and indicators of satisfaction with social care (albeit only cross-sectional data for this

variable) were found to be largely insignificant in models controlling for need and

supply effects. Adding year dummies causes most variables to go insignificant except

for the change in coverage indicator.

Discussion

The results indicate two main processes at play; DP utilisation being driven by

rationing of services and relatedly, by relative cost of directly provided services. This

is consistent across the four client groups, even though we find a certain variation in

relationships with need and supply between the groups. Compared to previous

research (see Fernandez et al. 2007) our set of predictors have evolved to match the

current policy environment. We find more uniform effects across the client groups,

indicating some convergence due to the increasing spread of DPs across the groups. In

the early years of the 2000s ‘physical disability’ users where still the only user group

where DPs were an inherent part to service provision. We are now also more

concerned with the effect of the current fiscal austerity impacting significantly on care

services. There is concern that the role of DPs as a tool for rationing services in the

Page 18: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

18

current austere fiscal climate will lead to negative effects on quality. This particularly

due to the importance of information and support (Glendinning et al. 2008) in

managing DPs which may be rationalised in LAs attempting to contain costs.

A particularly puzzling result is the positive effect of population density on the

utilisation of DPs. However, when introducing informal care and coverage in the

models population density tends to be an insignificant – only for mental health users

does it remain significant. This implies that the care market effect of more densely

populated areas outweighs the potential challenges to direct service provision which

may lead more rural LAs to favour DPs. There may further be a demand effect in

cities where a DP seems more attractive due to the relative abundance of care staff

and other services which allows for a care package which meets the user’s

preferences. The ambiguity of this interpretation highlights the challenge of

disentangling the effect of the two utility functions underlying the negotiation

between the care manager and the service user. Population density clearly influences

both, potentially in contradictory directions, however the current data does not allow

us to separate the effect.

Further, the relationship identified between direct payments and the availability of

informal care is worth noting. Medium to high intensity (over 20 hours per week)

exhibits a positive relationship with DP utilisation indicating that informal care may

be complementary to DP usage. This may imply that pre-existent informal caregiving

relationships are being supplemented by DPs. If this is the case, this is likely to be a

positive development for informal caregivers, as the respite and support needed is

often sporadic and meeting (highly) specific needs, which direct service may struggle

to, or be unable to provide. We must however caution that the prevalence of informal

care is population averaged and intrapolated, leading to less precision in our

estimates.

Conclusions

This paper has empirically investigated the extent to which local patterns can help

explain the expansion of direct payments utilisation in English local authorities. A

causal interpretation is hampered by the inability to disentangle the care supply from

the demand or service users and the substitute/complement of informal care. The

Page 19: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

19

reduced form modelling was necessary due to lack of data allowing us to instrument

at local authority level. However, our results still confirm that direct payments

utilisation is significantly shaped by attributes of local social care systems and both

supply and demand side influences.

The noticeable increase in direct payments utilisation across client groups, although

driven by a forceful lobby and consecutive governments, seems in later years to

correlate with issues of fiscal austerity and rationing of services. We find evidence of

direct payments being favoured by local authorities where coverage is shrinking and

where alternative methods of provision are relatively expensive. As direct payments

are broadening across the client groups it is crucial that investment in support

mechanisms is maintained. This is particularly important due to the fact that direct

payments are increasingly being used by service users with ‘less ability’. Individual

level data can beneficially explore this relationship further.

Finally, we have further shown that some influences are common across the user

groups while other have varying effects. Overall this indicates that the policy of direct

payments at the local level is shaped by broader policy drivers across user groups and

to some extend local authorities. The clear-cut variation between local authorities in

terms of direct payments utilisation, together with our findings on local patterns,

emphasises the importance for performance monitoring and further research taking

such factors into account, acknowledging the extent to which these influence the

policy choices of local authorities.

Page 20: Increasing the personalisation of social care in England ...€¦ · 1 Paper for Health Economists' Study Group (HESG) Meeting Glasgow Caledonian University 23rd-25th of June 2014

20

References Department of Health, 2000. Community Care (Direct Payments) Amendment

Regulations,

Ellis, K., 2007. Direct Payments and Social Work Practice: The Significance of

“Street-Level Bureaucracy” in Determining Eligibility. British Journal of Social

Work, 37(3), pp.405–422.

Ellis, K., 2011. “Street-level Bureaucracy” Revisited: The Changing Face of Frontline

Discretion in Adult Social Care in England. Social Policy & Administration,

45(3), pp.221–244.

Fernandez, J.-L. et al., 2007. Direct payments in England: factors linked to variations

in local provision. Journal of Social Policy, 36(01), pp.97–121.

Fernandez, J.-L. et al., 2013. Implications of setting eligibility criteria for adult social

care services in England at the moderate needs level,

Fernández, J.-L., Snell, T. & Wistow, G., 2013. Changes in the patterns of social care

provision in England: 2005/6 to 2012/13, PSSRU Discussion Paper 2867

Forder, J. et al., 2012. Evaluation of the personal health budget pilot

programme.PSSRU discucssion paper 2840_2

Foster, M. et al., 2006. Personalised social care for adults with disabilities: a

problematic concept for frontline practice. Health & social care in the

community, 14(2), pp.125–35.

Glasby, J. & Littlechild, R., 2009. Direct payments and personal budgets: Putting

personalisation into practice, Bristol: Policy Press.

Glendinning, C., Halliwell, S., Jacobs, S., Rummery, K., & Tyrer, J. (2000). Bridging

the gap: using direct payments to purchase integrated care. Health & Social Care

in the Community, 8(3), 192–200.

Glendinning, C. et al., 2008. Evaluation of the Individual Budgets Pilot Programme:

Final Report. Available at: http://kar.kent.ac.uk/13322/ [Accessed May 20,

2014].

Glendinning, C., 2012. Home care in England: markets in the context of under-

funding. Health & social care in the community, 20(3), pp.292–9.

Glendinning, C., 2008. Increasing Choice and Control for Older and Disabled People:

A Critical Review of New Developments in England. Social Policy &

Administration, 42(5), pp.451–469.

Health and Social Care Information Centre, 2012. Community Care Statistics 2010-

11: Social Services Activity Report, England,

Hillcoat-Nalletamby, S. & Ogg, J., Moving beyond “ageing in place”: older people’s

dislikes about their home and neighbourhood environments as a motive for

wishing to move. Ageing and Society.

Holloway, S., Neville, S. & Watts, R., 2011. Longitudinal Study of Cash Payments

for. Adult Social Care in Essex, London.

Leece, D. & Leece, J., 2006. Direct Payments: Creating a Two-Tiered System in

Social Care? British Journal of Social Work, 36(8), pp.1379–1393.

Manning, W.G. & Mullahy, J., 2001. Estimating log models: to transform or not to

transform? Journal of health economics, 20(4), pp.461–94.

Milner, J. & O’Byrne, P., 1998. Assessment in Social Work, Hampshire: Palgrave.

Priestley, M. et al., 2006. Direct Payments and Disabled People in the UK: Supply,

Demand and Devolution. British Journal of Social Work, 37(7), pp.1189–1204.

Wittenberg, R. et al., 2011. Projections of demand for and costs of social care for

older people in England, 2010 to 2030, under current and alternative funding

systems.