The comparative effectiveness, safety, value, and adherence of newer P2Y12 inhibitors versus clopidogrel in the context of heterogeneity
MetadataShow full item record
Background P2Y12 inhibitors are a class of medications indicated with aspirin as part of dual antiplatelet therapy (DAPT) for patients with acute coronary syndrome after undergoing percutaneous coronary intervention (PCI). Taking a P2Y12 inhibitor along with aspirin after PCI reduces the incidence of cardiovascular death, nonfatal MI, and stroke in the following year. Clopidogrel, the first oral P2Y12 inhibitor, was approved in 1997 and was used widely as part of DAPT therapy. However, there are some issues with clopidogrel such as low bioavailability and slow onset of platelet inhibition compared to newer agents, heterogeneity in patient response, and drug resistance. These concerns were addressed with two newer oral agents, prasugrel and ticagrelor, which were approved in 2009 and 2011, respectively. This study aims to examine the heterogeneity in treatment effects of these antiplatelet agents and assess how healthcare providers and patients change their behaviors as a result. In Aim 1, I examined whether physicians identify heterogeneity of treatment effects (HTE) with P2Y12 inhibitors and change their prescribing patterns as a result. This is extended into Aim 2, which explored whether physicians’ adaptions in prescribing patterns was a cost effective strategy. Aim 3 assessed patients’ adherence to P2Y12 inhibitors and determined if any factors were associated with heterogenous impacts along the adherence distribution. Methods In Aim 1, an instrumental variable approach with person-centered treatment effects was used to assess patient-level comparative effectiveness and safety outcomes from January 2010 to December 2017. These outcomes were used to study whether physician adapted their prescribing patterns over the study period to match the most optimal P2Y12 inhibitor with the patient. In Aim 2, results from Aim 1 were extended to develop a hybrid lifetime Markov model to assess whether physicians’ adaptive prescribing with P2Y12 inhibitors was a cost effective strategy compared to universal treatment with clopidogrel or the newer P2Y12 inhibitors. Inputs for the 1-year short term model were informed from Aim 1 results, while inputs for the lifetime model were informed from the literature. In Aim 3, conditional and unconditional quantile regression models were used along with more traditional logistic regression models to examine medication adherence of the three P2Y12 inhibitor, and to determine which covariates had heterogenous impacts along the adherence distribution. Patients who received drug eluting stents were measured for 185 days. Results In Aim 1, 52,823 patients were included for analysis. Patients on ticagrelor and prasugrel had a significantly lower probability of major adverse cardiovascular events [-3.97 percentage points (95% CI, -6.97 to -0.26)] and significantly lower probability of major bleeding events [-2.93 percentage points (95% CI, -4.83 to -0.70)] compared to patients on clopidogrel. Physicians were able to better align patients who would benefit on clopidogrel from 17.39% in 2010 to 26.40% by 2015, but patient outcomes were not significantly different than when everyone received ticagrelor or prasugrel. In Aim 2, physicians’ adaptive prescribing resulted in 11.63 life-years (LYs), 9.92 quality-adjusted life-years (QALYs), and $72,403 total costs; universal clopidogrel resulted in 11.59 LYs, 9.84 QALYs, and $72,670 total costs; and universal prasugrel or ticagrelor resulted in 11.67 LYs, 10.00 QALYs, and $73,325 total costs. Universal treatment with prasugrel or ticagrelor was a dominant strategy compared to universal treatment with clopidogrel and a cost-effective strategy compared to targeted prescribing with an incremental cost-effectiveness ratio of $11,144/QALY. In Aim 3, adherence to P2Y12 inhibitors were generally high (88.74%, 90.52%, 88.95%, and 89.60% for branded clopidogrel, generic clopidogrel, prasugrel, and ticagrelor, respectively). The unconditional quantile regression approach showed prasugrel and ticagrelor had significantly lower adherence compared to branded clopidogrel, especially around the 30th percentile. Other factors that impacted adherence the most were having comorbid depression and living in the southern region. Conclusions The research from this dissertation provided a deeper understanding of P2Y12 inhibitors and how heterogeneity of treatment effects impacted physician prescribing and in turn the cost-effectiveness of targeted prescribing with these antiplatelet agents. This research also revealed which factors impacted adherence the most along the entire adherence distribution so that future strategies to improve adherence could be better informed.