Raanan Shamir, MD

Cost Effectiveness of Population Screening

15.9.2010

Screening-WHO Definition

An investigation whereby:

a simple test is applied to

an asymptomatic population without consulting a health care

system in order to identify a disease.

Mass screening

Celiac disease fulfils all requirements for a mass screening program: It is common Specific screening tests are

available Definite treatment (gluten-free

diet) is available If unrecognized, morbidity and

complications may ensue.

However… Most available data about the

epidemiology of CD is derived from studies in symptomatic patients

GFD is not easy to maintain and may constitute a cumbersome lifestyle

There is a paucity of data on the cost–effectiveness (CE) of CD screening

On GFD IgA and IgG TTG decreased significantly over 12 months

About 25% admitted sub optimal compliance on social occasions explaining the continuing (although low) positivity in 8/28 subjects after 18 months

Prevalence of CD is high in Italian schoolchildren.

2/3 of cases were asymptomatic. Acceptance of the program was

good, as was dietary compliance. Given the high prevalence and

possible complications of untreated CD, mass screening deserves careful consideration.

Tommasini A, et al. Arch Dis Child 2004

 

Med Desc Making 2006;26:1-12

Methods

We developed a state-transition Markov model, from a societal perspective, to evaluate whether it is cost effective to screen the entire population at the age of 18 years (range 6-30 years), for CD.

Since the model uses values that can vary between studies and countries, the influence of establishing a certain value on the model was examined using the sensitivity analysis (examining the model within the range of preset values).

Markov Model for CD Screening

Unknown CD

No CD

Known CD no GFD

Known CD Strict

diet

Dead

Methods (Cont.)

Effectiveness was measured by life expectancy in each strategy (TTG, EMA, IgA and AGA IgA combinations).

We calculated the incremental average cost-effectiveness (CE) ratio for each strategy (additional expected cost divided by additional expected effectiveness) compared with the next least expensive strategy.

Costs

We considered only direct health care costs, representing the average payments for each coded procedure based on the 2004 Medicare Fee Schedule (http://www.hgsa.com/professionals/feedb-2004.shtml)

The cost of biopsy includes additional average costs of endoscopy and its complications (bleeding, perforations, death)

The costs of evaluation of symptoms include the following: Physician’s visits, laboratory tests (two physicians visits, complete blood count, chemistry panel, CD serology), and endoscopy that is mandatory for the diagnosis of CD

Cost values in $

Results

• All screening strategies resulted in life year saving and reductions in mortality related to CD

• Base-case analysis at an annual discount rate of 3% revealed a US$ 49,491 per life-year gained for screening compared to no screening using the EMA strategy

ICER as a function of prevalence

ICER as a function of SMR in CD

Conclusions (1)

Mass screening would be CE in populations with a high prevalence of CD, over a wide range of ages at screening, assuming that the mortality rate is higher in undiagnosed CD, and that implementing and adhering to a GFD in cases diagnosed by screening reduces the mortality rate

Conclusions (2)

From a CE perspective, EMA is the best serological marker for mass screening for CD

Due to the uncertainties regarding the validity of our assumptions, screening for CD would be justified only if these assumptions are proved to be correct

To determine the CE of mass screening within the young adult population for CD

To document the changes in quality of life produced by CD screening

To define the parameters constituting the highest impact on the CE

T. Hershcovici1 M. Leshno, E. Goldin, E. Israeli & R. Shamir. Aliment Pharmacol Ther 2010;31:901-10

Mass screening, the 2010 Update

METHODS

Design: State transition Markov model

Perspective: Third-party payer Interventions: CD screening compared

to a no-screening strategy Outcome Measures:

Quality Adjusted Life Years (QALYs) Incremental Cost-Effectiveness Rate (ICER)

Positive Serology

IBS

IDA

Morbidity

Mortality

QALYs

SMR

Latent Silent Overt

%over time

Screening

GFD QALYs

Strategy

Cost

(US$) QALYsICER

(US$/QALYs)

No

Screening 24.94 26.9031  

Screening 158.64 26.9057* 48.960

* The screening strategy resulted in a gain of 0.0027 QALYs.

Results

Parameters with Significant Influence on ICER

Time delay to CD diagnosis

Utility of CD patients adhering to GFD

Prevalence of CD

Screening would be CE if the time delay to diagnosis is longer than 6 years, and utility of

GFD adherence is greater than 0.978.

Two-Way Sensitivity Analysis

Screening is not CE for a time delay less than 3.5 years independent from CD prevalence*

Screening is not CE for a time delay less than 5.2 years independent from the utility of adherence to GFD*

*ICER=50,000US$/QALY

Our model suggests that mass screening for CD of the young-adult general population is associated with improved QALYs and is a cost effectiveness strategy However:

Shortening of the time delay to diagnosis by heightened awareness of

health-care professionals may be a valid alternative to screening

Conclusions

Base-case values and ranges used in sensitivity analysis

Parameters that do not Influence ICER Significantly

Adherence to GFD

Probability to suffer of IDA or IBS-like symptoms

The utility of IDA or IBS-like symptoms

Cost of endoscopic biopsy and of serologic testing

Standardized Mortality Rate

Screening for Celiac Disease-Justified?

CD is the most common food-based disease

Clinically based early diagnosis is difficult and is frequently delayed**

There are diagnostic serologic tests that are both specific and sensitive and can be applied

to a large population**

Undiagnosed patients remain at risk for complications

A gluten-free diet (GFD) relieves symptoms and decreases morbidity and mortality

Perform CD screening in high-risk conditions

associated with increased prevalence of

disease

Mass-screening for CD remains controversial

CD Screening-Recommendations

Gastroenterology 2005;128:S104

Markov Model – No-Screening Arm

Markov Model–Screening Arm

MortalitySMR

No Screening

IBS

IDA

Morbidity

QALYs Celiac Disease

GFD

Latent Silent Overt

Percentage over time

Base–Case Analysis:Estimation of Parameters

Prevalence of CD in the general population: 1%

Standardized Mortality Rate (SMR):Untreated CD patients: 1.6

CD patients on GFD: 1.1

Average time delay to CD diagnosis from onset of symptoms: 6 years

Base–Case Analysis:Estimation of Parameters

Compliance to GFD:Symptomatic CD patients: 82% at 10 years

Asymptomatic CD patients: 60% at 10 years

Utility of GFD: 0.98

Base–Case Analysis:Estimation of Parameters

•Sensitivity of IgA anti-tTG : 95%

•Specificity IgA anti-tTG :98%

•Sensitivity of IgG anti-tTG* :98.7%

•Specificity IgG anti-tTG* :98.6%

*in IgA deficient patient

Base–Case Analysis:Estimation of Parameters

Annual probability to develop symptoms in asymptomatic CD patient:2.8%

Prevalence of symptoms in CD patients:IBS-like symptoms: 30%

Iron-deficiency anemia: 50%

Utility of symptoms: IBS-like symptoms: 0.73

Iron-deficiency anemia: 0.76

Sensitivity Analysis on Prevalence of CD

Screening remained cost-effective for a CD prevalence range between 0.875-2%.

Sensitivity Analysis on Time Delay to CD diagnosis

For a time delay of less than 5.9 years, screening for CD is not cost-effective.

Sensitivity Analysis on Utility of Adherence to GFD

For utility of adherence to GFD between 0.95-0.978, screening for CD is not cost-effective .

For an estimated utility less than 0.95, the no-screening strategy becomes dominant.

Sensitivity Analysis on SMR

A SMR between 1.2-2.7 for symptomatic CD patients had no major effect on ICER.

•Developed a case-finding strategy based on testing for TTG IgA antibodies in subjects showing

predefined signs and symptoms or belonging to at-risk groups .

•69 primary care doctors and 60 primary care paediatricians agreed to participate .

•1041 adults and 447 children were selected for TTG-Ab testing during the year of the study

(2001).

Berti I, et al. Case-finding startegy, 2006

Berti I, et al. Case-finding startegy, 2006

•31 subjects (2.08%, 19 adults and 12 children) were ultimately diagnosed as CD, while no cases of coeliac

disease had been diagnosed by the participating doctors in the previous year, 29 subjects were diagnosed as

coeliacs in the year after the completion of the study (2002) .

•The prevalence of confirmed CD in the population under study increased from 1:1506 to 1:1073 in adults and

from 1:827 to 1:687 in children from year 2000 to 2001. When cases diagnosed in 2002 are included, the

prevalence is 1:832 and 1:602, respectively .

Berti I, et al. Case-finding startegy, 2006Conclusions

•Case-finding is a feasible and successful strategy for detecting undiagnosed CD patients and has

the important added value of increasing the awareness of the disease among primary care

physicians• It represents a cost-effective alternative to

population screening for reducing the burden of undiagnosed CD