• Modeling costs and opportunistic infections for Maryland Medicaid HIV/AIDS patients: Effect of patient non-adherence to antiretroviral drugs

      Huang, Xing-Yue; Mullins, C. Daniel (2001)
      HIV/AIDS is the most important infectious disease. The significant benefits of new antiretrovirals to HIV/AIDS patients have been documented but these benefits depend on strict patient adherence. Although various published articles have focused on patient non-adherence, most of them have failed to assess the economic and clinical impacts of non-adherence. Maryland Medicaid claims data were used to evaluate risk factors, economic impact and impact of non-adherence on opportunistic infections (OI). The study population included patients who were continuously enrolled in Maryland Medicaid program from the index date to July 30, 1997. The mean non-adherence rate was 20%. Caucasian, male and/or patients who resided in Baltimore County had a significantly lower non-adherence rate compared to their counterparts. In addition, patients treated with protease inhibitor (PI) had a significantly lower non-adherence rate than their counterparts. Race, gender, and resident of Baltimore County were not significantly associated with a ≥95% adherence rate. Age was positively and PI treatment was negatively associated with a ≥95% adherence rate. Results consistently indicate that patients with a ≥95% adherence rate had lower monthly HIV/AIDS-related costs than their counterparts. In addition, Caucasian patients and/or patients with Medicaid and Medicare coverage had a lower monthly cost than their counterparts. However drug dependents had a significantly higher monthly cost than their counterparts. Patients who died, patients with OI and/or patients with severe mental illness had a significantly higher monthly cost than their counterparts. The impact of a ≥95% adherence rate on OI was insignificant. However, the results should be interpreted with caution. Limited sample size and low prevalence of OIs raised the issue of lack of statistical power, which may impede detection of the impact. In conclusion, this study confirms that patient social and clinical factors are significantly associated with non-adherence in a Medicaid HIV/AIDS population. Most importantly, a ≥95% patient adherence rate is associated with a lower monthly cost. The impact of a ≥95% adherence rate on risks of having OI is insignificant which may be due to lack of statistical power. Future studies should assess the clinical impact on more prevalent OIs in a larger patient population.