Short Term Impact and Markov Modeling Analysis of the Cost-Effectiveness of the Meta Salud Diabetes Cluster Randomized Controlled Trial for Cardiovascular Risk Reduction Among Diabetics in the Mexican National Primary Care Health System

Abstract

TITLE: Short Term Impact and Markov Modeling Analysis of the Cost-Effectiveness of the Meta Salud Diabetes Cluster Randomized Controlled Trial for Cardiovascular Risk Reduction Among Diabetics in the Mexican National Primary Care Health System. BACKGROUND: The cost-effectiveness of the Meta Salud Diabetes (MSD) trial is not well established. This was a clinically integrated lifestyle intervention designed to reduce modifiable cardiovascular (CVD) risk factors and promote diabetes control via self-help groups within the structure of a support group (GAM: Grupo de Ayuda Mutua) in government-run (community) Health Centers (Centros de Salud) in Sonora, Mexico. The MSD trial has previously been shown to demonstrate a positive trend in reducing CVD risk in adults with type 2 diabetes in a low-resource setting. OBJECTIVE: To conduct a cost-effectiveness analysis of the use of a self-management diabetes program (Meta Salud Diabetes) delivered in a culturally appropriate context (Sonora Mexico) from the payer perspective. DESIGN: Cohort level Markov based modeling. DATA SOURCES: Clinical and cost data from the MSD trial (a parallel, two-arm, cluster-randomized, behavioral clinical trial across 22 clinics in Sonora, Mexico, between August 2016 and October 2018) and data from published literature. TARGET POPULATION: Mexican adults aged 45 years or older with known type 2 diabetes and receiving care in government-run (community) Health Centers (Centros de Salud) in Sonora, Mexico. TIME HORIZON: 10 years. PERSPECTIVE: Payer. METHODS: A Markov based decision model was built through TreeAge Pro to analyze the cost-effectiveness of MSD compared to GAM. In each treatment strategy, two patient groups were modeled: patients with controlled diabetes (defined as A1c ≤7.0g/dl) and patients with uncontrolled diabetes (A1c>7.0g/dl). Patients could transition through various health states determined by transition probabilities. The incremental cost per glycemic improvement (HbA1c improvement of 1% and above) per patient at the end of 13 weeks was also determined. Uncertainty evaluation via probabilistic sensitivity analyses were conducted. A range of willingness-to-pay (WTP) thresholds (US$10,500 to US$31,500 per QALY) was used in the analyses. Glycated hemoglobin levels (HbA1c) were used as proxy for diabetes control. OUTCOME MEASURES: Direct medical costs, quality-adjusted life-years (QALYs), net monetary benefit, incremental cost-effectiveness ratios (2023 U.S. dollars per quality-adjusted life-year [QALY] gained) of the intervention compared to usual care expressed as discounted costs and benefits from the payer’s perspective. RESULTS OF BASE-CASE ANALYSIS: In the base case, MSD was highly cost-effective. The intervention added 0.01 QALYs relative to the GAM only strategy at an incremental cost of $18 per additional QALY. At $10,500 willingness-to-pay per QALY, the MSD strategy was mostly cost-effective in 62% of the ICER iterations suggesting a moderate level of confidence in this strategy as an optimal pathway. In the short term (13 weeks) there was an incremental cost of $17.79 per person to cause an incremental improvement in HbA1c in about 7% of the population. A scenario-based deterministic approach of varying various health personnel related costs yields a lower ICER range of $250.68 when training is delivered by 2 Community Health Workers and a higher end range ICER of $961.19 when training is delivered by 2 physicians to cause at least a 1% reduction in HbA1c in 13 weeks. RESULTS OF SENSITIVITY ANALYSIS: Probabilistic sensitivity analysis consistently showed that the MSD intervention was an optimal strategy compared to GAM only (usual care). Model inputs with greatest impact on the ICER were cost of MSD, proportion of patients with controlled diabetes entering the cycle, discount rate, age of the cohort. LIMITATIONS: Individual level patient characteristics were not included. Indirect costs were not captured. CONCLUSION: Aggressive non-pharmacological control of diabetes using the Meta Salud Diabetes curriculum in addition to usual care offered at Grupos de Ayuda Mutua (GAM) centers in this low resource setting in Mexico for the control of cardiovascular risk factors in diabetes was cost-effective. PRIMARY FUNDING SOURCE: None