Cost Effectiveness of Biomarker Tests for Irritable Bowel Syndrome With Diarrhea: A Framework for Payers

Abstract

Background & aims: Diagnosis of diarrhea-predominant irritable bowel syndrome (IBS-D) relies on the Rome IV symptom-based criteria, which are imperfect for separating functional vs organic disease. Biomarker tests for IBS-D might be added to symptom data to allow clinicians to make more accurate and precise diagnoses in a cost-effective manner. We tested the economic consequences of using a range of hypothetical IBS-D biomarkers, and explored at what cost and level of accuracy a biomarker becomes cost effective. We produced a framework for payers to evaluate the return on an investment of implementing IBS-D biomarkers of varying accuracy and cost.

Methods: We used decision analysis software to evaluate a hypothetical cohort of patients who met Rome IV criteria for IBS-D. We conducted cost-utility and budget impact analyses of 2 competing approaches: usual care or an IBS biomarker-based diagnostic approach. Patients in the usual care group received empiric IBS treatment; non-responders received additional diagnostic tests for organic disease. In the group evaluated with a biomarker test, those with a positive result received IBS treatment before additional diagnostic analyses, whereas patients with a negative result underwent upfront diagnostic testing. Outcomes were incremental cost per quality-adjusted life year gained (third-party payer perspective) and incremental per-member per-month cost.

Results: In the base-case analysis, using a willingness-to-pay threshold of $100,000/quality-adjusted life year, we found that biomarkers are not cost effective when the biomarker test costs more than $846, even if the test is 100% accurate in detecting IBS-D. In probabilistic analysis using 1,000 simulations, most trials (75% or more) show that the biomarker-based diagnostic approach is cost effective above the following accuracy thresholds: a $100 biomarker test with 51% accuracy, a $200 test with 57% accuracy, a $300 test with 63% accuracy, a $400 test with 69% accuracy, a $500 test with 76% accuracy, a $600 test with 82% accuracy, a $700 test with 89% accuracy, and a $800 test with 94% accuracy.

Conclusions: In decision analysis of a hypothetical cohort of patients who met Rome IV criteria for IBS-D, we identified cost and accuracy thresholds that can guide investigators and payers as they develop, validate, price, and/or reimburse IBS-D biomarker tests for use in everyday clinical practice.

Keywords: Budget Impact; Diagnostic; Modeling; QALY.

FIGURE 1

Truncated decision model. Individuals in the usual care arm first received empiric IBS-D therapy and needed to fail two treatment courses before undergoing additional diagnostic testing with endoscopy, celiac serologies, and CRP. Those in the IBS-D biomarker-based approach first had a biomarker test. Individuals with a positive test received IBS-D therapy and diagnostic testing was only performed after three failed treatment courses. On the other hand, those with a negative biomarker test had immediate diagnostic testing, which determined whether the patient was treated for organic disease (celiac disease, IBD, or microscopic colitis) or IBS-D. □=decision node; ○=probability node; ◁= terminal node.

FIGURE 2

Nomogram depicting the preferred strategy when the IBS-D biomarker costs $100 and at a WTP threshold of $100,000/QALY gained. For example, a biomarker that is 80% sensitive and 40% specific is not cost effective versus usual care. However, when increasing the specificity of the same biomarker to 50%, its use becomes cost effective.

FIGURE 3

At varying biomarker costs and accuracies, we performed probabilistic sensitivity analyses using 1,000 trials and determined the proportion of trials within the $100,000/QALY WTP threshold. For example, if a $300 biomarker is 60% accurate, then 54.8% of trials would fall within the budget. Conversely, when increasing the accuracy to 75%, 99.6% of trials found that it is cost effective versus usual care. Note: tested scenarios included those where the biomarker’s sensitivity, specificity, and thereby accuracy, were equal.