|
Testing for celiac sprue in irritable bowel syndrome with predominant diarrhea: a cost-effectiveness analysis |
Spiegel B M R, DeRosa V P, Gralnek I M, Wang V, Dulai G S |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology Two strategies for the management of patients with who met the Rome II criteria for diarrhoea-predominant irritable bowel syndrome (IBS-D) were compared. The first was the direct initiation of empirical IBS treatment. The second was initial screening for coeliac sprue (CS). In the second strategy, tests that proved to be positive were followed by endoscopic biopsy.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised a hypothetical cohort of patients with Rome II-positive IBS-D symptoms who were negative in alternative organic diagnoses. Patients who had normal serum chemistry panel, complete blood count, stool examination for ova and parasites, thyroid stimulating hormone level, erythrocyte sedimentation rate and structural evaluation of the colon, were included in the study. Patients who had suffered from gastrointestinal bleeding, dysphagia, and weight loss, or had alarm signs like iron-deficiency anaemia and occult blood-positive stools, were excluded from the study.
Setting The setting was not explicitly stated, but it was likely to have been secondary care. The economic study was carried out in the USA.
Dates to which data relate The effectiveness data were derived from studies published between 1967 and 2003. The cost data appear to have been derived from official sources published in 2003, although it was reported as 2002 in the table.
Source of effectiveness data The effectiveness data were derived from a review and synthesis of completed studies, augmented by authors' assumptions.
Modelling A decision analytic model, a Markov model, was designed using decision analysis software (DATA 4.0 TreeAge Software) in order to evaluate the cost-effectiveness of the two strategies. A Bayesian approach was used to integrate CS prevalence and different characteristics of CS diagnostic tests. The base-case time horizon of the model was 10 years and comprised 120 cycles, each of which were one month in duration. After receiving treatment it was assumed that patients entered the Markov cycle after symptomatic improvement under therapy, and could either remain in remission or develop recurrent symptoms. The model was based on various assumptions about the prevalence of CS in IBS, CS diagnostic tests, gluten-free diet in sprue, IBS therapy and provider behaviour. The assumptions were thoroughly analysed in the paper and were too numerous to report in this abstract.
Outcomes assessed in the review The authors conducted a systematic review of the medical literature. The input parameters used in the model were:
the prevalence of underlying CS in Rome-positive IBS patients;
the sensitivity and specificity of a serologic test for CS;
the probability that CS is latent or potential;
the probability that CS is associated with immunoglobulin (Ig) A deficiency;
the probability of compliance with a gluten-free diet;
the probability of improvement of CS symptoms under a gluten-free diet;
the probability of initial symptomatic improvement with generic IBS therapy;
the probability of symptom recurrence following initial symptomatic improvement with generic IBS therapy;
the proportion of physicians who test for CS following failure with IBS therapy;
the mean diagnostic delay time between failed IBS therapy and testing for CS;
the frequency of physician visits for ongoing IBS symptoms; and
the frequency of physician visits for resolved IBS symptoms.
Study designs and other criteria for inclusion in the review The authors only included publications in English. They did not report the study designs included in their review.
Sources searched to identify primary studies MEDLINE was searched for primary studies. In addition, the authors also carried out an abstract review with the use of a CD-ROM (Digestive Disease Week Abstracts-on-Disc; AGA).
Criteria used to ensure the validity of primary studies Methods used to judge relevance and validity, and for extracting data Number of primary studies included Overall, 12 primary studies provided effectiveness evidence.
Methods of combining primary studies Values biased against CS were selected from the literature.
Investigation of differences between primary studies The authors investigated the pros and cons of each primary study.
Results of the review The base-case estimates (range in the literature) were as follows.
The prevalence of underlying CS in hypothetical cohort with Rome-positive IBS symptoms was 3.4% (range: 0 - 11.4).
The sensitivity of the serologic test for CS was 85% (range: 85 - 98) and the specificity was 94% (range: 94 - 100).
The probability that CS is latent or potential was 35% (no range).
The probability that CS is associated with IgA deficiency was 5% (range: 2 - 10).
The probability of compliance with a gluten-free diet was 50% (range: 50 - 80).
The probability that a gluten-free diet improves CS bowel symptoms was 70% (range: 70 - 80).
The probability of initial symptomatic improvement with generic IBS therapy was 75% (range: 35 - 75).
The probability of recurrent symptoms following initial symptomatic improvement with generic IBS therapy was 50% (no range).
The mean diagnostic delay tie between failed IBS therapy and testing for CS was 6 months (12 months).
The proportion of physicians who test for CS following failure with IBS therapy was 25% (no range).
The frequency of physician visits for ongoing IBS symptoms was every 3 months (no range).
The frequency of physician visits for resolved IBS symptoms was once yearly (no range).
With the exception of the last three results, which were based on authors' assumptions, all the results were derived from the literature.
Methods used to derive estimates of effectiveness The authors made assumptions to derive some estimates of effectiveness.
Estimates of effectiveness and key assumptions Some parameters of the model were derived using authors' assumptions, owing to the lack of precise data in the literature. These parameters were the proportion of physicians who test for CS following failure with IBS therapy, the frequency of physician visits for ongoing IBS symptoms, and the frequency of physician visits for resolved IBS symptoms.
Further assumptions were made. These concerned the prevalence of CS in IBS, the prevalence of CS in IBS, CS diagnostic tests in IBS, gluten-free diet in sprue, and IBS therapy. All assumptions were explicitly justified.
Measure of benefits used in the economic analysis The measure of benefit used was the proportion of patients with symptomatic improvement at the end of the time horizon (symptomatic improvement).
Direct costs The direct hospital costs were evaluated. These included the costs of the initial general medicine office visit, the follow-up medicine office visit, the generic screening test for CS, diagnostic upper endoscopy for a positive CS test, inpatient admission of ulcer peroration from upper endoscopy, Medicare DRG for bowel perforation, the emergency room fee, initial surgical consultation, the surgeon's fee, the anaesthesiologist's fee, the surgeon's follow-up visit, and the cost per month of generic IBS therapy. The costs of diagnostic upper endoscopy for a positive CS test covered the endoscopist's consultation fee and procedure fee, facility fee and biopsy interpretation fee. Average wholesale prices were used for the drugs. All the costs were derived from official sources published in 2003. The unit costs were reported, but the quantities of resources used were not analysed separately. The costs were discounted at a rate of 3%.
Statistical analysis of costs The costs were treated deterministically.
Indirect Costs The indirect costs were not included in the analysis.
Sensitivity analysis A multivariable sensitivity analysis ("tornado analysis") was undertaken to assess the robustness of the results when the input parameters were varied and to determine the most influential parameters. Using each of these influential parameters, a one-way sensitivity analysis was performed to determine the threshold estimates at which CS testing dominated IBS therapy. In addition, the authors conducted a probabilistic Monte Carlo simulation based on the assumption that all variables were triangular in distribution. One thousand trials were estimated and the authors reported the median value and the 2.5 and 97.5 percentile values. The authors also performed a threshold analysis assuming three different amounts (thresholds) that third-party payers were willing to pay for additional symptomatic improvement. The ranges used in the sensitivity analyses were derived from the literature.
Estimated benefits used in the economic analysis In the empiric IBS therapy group, 50.9% of the patient cohort achieved symptomatic improvement at the end of the 10-year period. The equivalent proportion in the CS testing group was 51.6%.
Cost results The average cost per patient was reported for each strategy. The average cost per patient treated was $4,023 in the IBS therapy group and $4,100 in the CS test group.
Synthesis of costs and benefits An incremental analysis was performed. The CS testing strategy incurred an incremental cost of $11,000 per additional symptomatic improvement.
The multivariable sensitivity analysis demonstrated that the results were sensitive to the prevalence of underlying CS, the specificity of the diagnostic test for CS, the probability that a gluten-free diet improves symptom prevalence and the cost of IBS therapy.
The one-way sensitivity analysis showed that CS testing dominated IBS therapy if the prevalence of underlying CS in patients with Rome-positive IBS exceeded 8%, the specificity of the CS test exceeded 98%, the probability that a gluten-free diet improves symptoms of CS exceeded 95%, or if the cost per month of IBS therapy exceeded $135.
The results of the probabilistic Monte Carlo simulations demonstrated that the median incremental cost-effectiveness ratio was $12,983 per additional symptomatic improvement (the 2.5 and 97.5 percentiles were -$32,520 and $41.031, respectively).
The percentage of trials beneath the $20,000, $50,000 and $100.000 willingness-to-pay thresholds were 78.0%, 89.1% and 94.7%, respectively.
Authors' conclusions Compared with therapy for irritable bowel syndrome (IBS), coeliac sprue (CS) screening was cost-effective for the majority of patients with diarrhoea-predominant IBS.
CRD COMMENTARY - Selection of comparators The authors explicitly justified their choice of the comparators. You should decide if they represent a widely used technology in your own setting.
Validity of estimate of measure of effectiveness A systematic review was undertaken to derive the effectiveness estimates. The conduct of the review was not described in detail, so it was unclear if it was conducted satisfactorily. The authors selected estimates of effectiveness from the primary studies that biased against CS, and made assumptions to derive other estimates of effectiveness. The choice of assumptions was explicitly justified and the estimates were investigated in a sensitivity analysis. As the time horizon of the model was 10 years, the effectiveness estimates were appropriately discounted.
Validity of estimate of measure of benefit The measure of benefit used in the economic analysis was the proportion of patients with symptomatic improvement over a 10-year period. The choice was explicitly justified.
Validity of estimate of costs The cost analysis was performed from the perspective of the third-party payer. As such, all the relevant categories of costs were included in the analysis. All unit costs and ranges of costs (for use in the sensitivity analysis) were reported. However, the quantities of resources used were not reported, thus impeding the reproducibility of the study in other settings. The unit costs were taken from published official sources and the price year was reported. There was disparate reporting of the publication year for cost data. Although the costs were treated deterministically, an appropriate sensitivity analysis was undertaken to test the robustness of the estimates used. The costs were discounted appropriately.
Other issues The authors did not compare their findings with those from other studies so it is not known how far their results agree with other published results. They did, however, explicitly address the issue of the generalisability of the results to other settings. The authors do not appear to have presented their results selectively. The study enrolled patients who met the Rome-II criteria for IBS-D and this was reflected in the authors' conclusions.
The authors reported several limitations to their study. First, their base-case estimates did not reflect all populations, and they were based on primary studies of different quality and study designs that referred to different populations. Second, there is evidence that the prevalence of CS in IBS-D might have been overestimated. Third, the cost-effectiveness of the CS test might have been overestimated because of inaccurate data in the literature concerning CS test characteristics. Fourth, the authors' analysis did not address the issue of health-related quality of life of the patients in the two groups, owing to the lack of validated utility measures. Finally, they did not include indirect societal costs in their analysis, owing to the lack of data, but reported that the inclusion of such costs was likely to bias the results in favour of the CS testing strategy.
Implications of the study The authors did not make explicit recommendations for changes in policy or practice. However, they explicitly recommended research to characterise, with precision, the performance of screening CS tests in sub-groups of patients with IBS symptoms. They specifically suggested that a prospective trial should be undertaken to evaluate the costs and effects according to IBS sub-groups. Future research should also try do derive utility measures for health states in IBS and CS so that a cost-utility analysis may be conducted. The authors also suggested the evaluation of possible comparators, for example, the use of a gluten-free diet and "test and treat" for CS without endoscopic confirmation.
Bibliographic details Spiegel B M R, DeRosa V P, Gralnek I M, Wang V, Dulai G S. Testing for celiac sprue in irritable bowel syndrome with predominant diarrhea: a cost-effectiveness analysis. Gastroenterology 2004; 126(7): 1721-1732 Indexing Status Subject indexing assigned by NLM MeSH Celiac Disease /diagnosis /economics /epidemiology /prevention & Cost-Benefit Analysis; Diarrhea /diagnosis /economics; Diet; Glutens; Humans; Irritable Bowel Syndrome /diagnosis /economics /epidemiology /prevention & Mass Screening /economics; Prevalence; Sensitivity and Specificity; control; control AccessionNumber 22004000851 Date bibliographic record published 31/12/2005 Date abstract record published 31/12/2005 |
|
|
|