One-Year Adverse Outcomes Among US Adults With Post-COVID-19 Condition vs Those Without COVID-19 in a Large Commercial Insurance Database.

Importance: Many individuals experience ongoing symptoms following the onset of COVID-19, characterized as postacute sequelae of SARS-CoV-2 or post-COVID-19 condition (PCC). Less is known about the long-term outcomes for these individuals. Objective: To quantify 1-year outcomes among individuals meeting a PCC definition compared with a control group of individuals without COVID-19. Design, Setting, and Participants: This case-control study with a propensity score-matched control group included members of commercial health plans and used national insurance claims data enhanced with laboratory results and mortality data from the Social Security Administration's Death Master File and Datavant Flatiron data. The study sample consisted of adults meeting a claims-based definition for PCC with a 2:1 matched control cohort of individuals with no evidence of COVID-19 during the time period of April 1, 2020, to July 31, 2021. Exposures: Individuals experiencing postacute sequelae of SARS-CoV-2 using a Centers for Disease Control and Prevention-based definition. Main Outcomes and Measures: Adverse outcomes, including cardiovascular and respiratory outcomes and mortality, for individuals with PCC and controls assessed over a 12-month period. Results: The study population included 13 435 individuals with PCC and 26 870 individuals with no evidence of COVID-19 (mean [SD] age, 51 [15.1] years; 58.4% female). During follow-up, the PCC cohort experienced increased health care utilization for a wide range of adverse outcomes: cardiac arrhythmias (relative risk [RR], 2.35; 95% CI, 2.26-2.45), pulmonary embolism (RR, 3.64; 95% CI, 3.23-3.92), ischemic stroke (RR, 2.17; 95% CI, 1.98-2.52), coronary artery disease (RR, 1.78; 95% CI, 1.70-1.88), heart failure (RR, 1.97; 95% CI, 1.84-2.10), chronic obstructive pulmonary disease (RR, 1.94; 95% CI, 1.88-2.00), and asthma (RR, 1.95; 95% CI, 1.86-2.03). The PCC cohort also experienced increased mortality, as 2.8% of individuals with PCC vs 1.2% of controls died, implying an excess death rate of 16.4 per 1000 individuals. Conclusions and Relevance: This case-control study leveraged a large commercial insurance database and found increased rates of adverse outcomes over a 1-year period for a PCC cohort surviving the acute phase of illness. The results indicate a need for continued monitoring for at-risk individuals, particularly in the area of cardiovascular and pulmonary management.

Medication non-adherence patterns and profiles for patients with incident myocardial infarction: Observations from a large multi-morbid US population.

BACKGROUND: Consistent adherence levels to multiple long-term medications for patients with cardiovascular conditions are typically advocated in the range of 50% or higher, although very likely to be much lower in some populations. We investigated this issue in a large cohort covering a broad age and geographical spectrum, with a wide range of socio-economic disability status. METHODS: The patients were drawn from three different health plans with a varied mix of socio-economic/disability levels. Adherence patterns were examined on a monthly basis for up to 12 months past the index date for myocardial infarction (MI) using longitudinal analyses of group-based trajectory modelling. Each of the non-adherent patterns was profiled from comorbid history, demographic and health plan factors using main effect logistic regression modelling. Four medication classes were examined for MI: betablockers, statin, ACE inhibitors and anti-platelets. RESULTS: The participant population for the MI/non-MI cohorts was 1,987,605 (MI cohort: mean age 62 years, 45.9% female; non-MI cohort: mean age 45 years, 55.3% females). Cohorts characterized by medication non-adherence dominated the majority of MI population with values ranging from 74% to 82%. There were four types of consistent non-adherence patterns as a function of time for each medication class: fast decline, slow decline, occasional users and early gap followed by increased adherence. The characteristics of non-adherence profiles eligible for improvement included patients with a prior history of hypertension, diabetes mellitus and stroke as co-morbidities, and Medicare plan. CONCLUSIONS: We found consistent patterns of intermediate non-adherence for each of four drug classes for MI cohorts in the order of 56% who are eligible for interventions aimed at improving cardiovascular medication adherence levels. These insights may help improve cardiovascular medication adherence using large medication non-adherence improvement programs.

Improving Stroke Risk Prediction in the General Population: A Comparative Assessment of Common Clinical Rules, a New Multimorbid Index, and Machine-Learning-Based Algorithms.

BACKGROUND: There are few large studies examining and predicting the diversified cardiovascular/noncardiovascular comorbidity relationships with stroke. We investigated stroke risks in a very large prospective cohort of patients with multimorbidity, using two common clinical rules, a clinical multimorbid index and a machine-learning (ML) approach, accounting for the complex relationships among variables, including the dynamic nature of changing risk factors. METHODS: We studied a prospective U.S. cohort of 3,435,224 patients from medical databases in a 2-year investigation. Stroke outcomes were examined in relationship to diverse multimorbid conditions, demographic variables, and other inputs, with ML accounting for the dynamic nature of changing multimorbidity risk factors, two clinical risk scores, and a clinical multimorbid index. RESULTS: Common clinical risk scores had moderate and comparable c indices with stroke outcomes in the training and external validation samples (validation-CHADS2: c index 0.812, 95% confidence interval [CI] 0.808-0.815; CHA2DS2-VASc: c index 0.809, 95% CI 0.805-0.812). A clinical multimorbid index had higher discriminant validity values for both the training/external validation samples (validation: c index 0.850, 95% CI 0.847-0.853). The ML-based algorithms yielded the highest discriminant validity values for the gradient boosting/neural network logistic regression formulations with no significant differences among the ML approaches (validation for logistic regression: c index 0.866, 95% CI 0.856-0.876). Calibration of the ML-based formulation was satisfactory across a wide range of predicted probabilities. Decision curve analysis demonstrated that clinical utility for the ML-based formulation was better than that for the two current clinical rules and the newly developed multimorbid tool. Also, ML models and clinical stroke risk scores were more clinically useful than the "treat all" strategy. CONCLUSION: Complex relationships of various comorbidities uncovered using a ML approach for diverse (and dynamic) multimorbidity changes have major consequences for stroke risk prediction. This approach may facilitate automated approaches for dynamic risk stratification in the significant presence of multimorbidity, helping in the decision-making process for risk assessment and integrated/holistic management.