Browse the corpus

Cardiovascular disease remains the leading cause of death globally and in the United States.1 2 Major modifiable risk factors include hypertension, smoking, poor diet, diabetes, and socioeconomic deprivation.3 Among these, high systolic blood pressure is the most substantial contributor to global cardiovascular burden.4 Improving risk factor control in underserved populations continues to be a critical public health priority. In this context, Inoue and colleagues reanalysed data from the Oregon Health Insurance Experiment to examine whether Medicaid coverage affects cardiovascular risk.5 Their original analysis reported significant reductions in systolic blood pressure and glycated haemoglobin (HbA1c) among individuals predicted to benefit most from insurance coverage. However, a post-publication correction identified a coding error. The updated results revised the estimated systolic blood pressure reduction from −4.96 mmHg to −2.93 mmHg (95% confidence interval (CI) −5.82 to −0.32), and the difference in HbA1c is no longer statistically significant.5

it most from insurance coverage. However, a post-publication correction identified a coding error. The updated results revised the estimated systolic blood pressure reduction from −4.96 mmHg to −2.93 mmHg (95% confidence interval (CI) −5.82 to −0.32), and the difference in HbA1c is no longer statistically significant.5 Although the updated systolic blood pressure reduction remains statistically significant, its clinical relevance is less certain. The study population was relatively young and had moderate baseline cardiovascular risk. Previous studies, including the Global Burden of Disease project and pooled cohort analyses, indicate that levels of systolic blood pressure above 110 to 115 mmHg are associated with increased long term mortality and cardiovascular events.4 6 However, these associations reflect cumulative exposure over many years. A burden of proof study shows that ischemic heart disease risk begins to increase at systolic blood pressure levels as low as 120 mmHg, with a 39% greater risk relative to 100 mmHg (95% uncertainty interval 1.34 to 1.44), rising steeply to more than fourfold at 165 mmHg.7 However, these estimates are based on aggregated population level data from randomised trials and observational cohorts, and primarily reflect long term risk accumulation.7 They emphasise that public health strategies to reduce chronic exposure to elevated systolic blood pressure are key, rather than implying that modest short term reductions would provide clinical benefit in groups at moderate risk.7 International guidelines suggest that individuals with systolic blood pressure between 130 and 139 mmHg and elevated cardiovascular risk may benefit from treatment, particularly when additional risk factors are present.8 Still, whether a short term reduction of 2 to 3 mm Hg in a moderate risk subgroup would yield meaningful health benefit remains unclear.

individuals with systolic blood pressure between 130 and 139 mmHg and elevated cardiovascular risk may benefit from treatment, particularly when additional risk factors are present.8 Still, whether a short term reduction of 2 to 3 mm Hg in a moderate risk subgroup would yield meaningful health benefit remains unclear. Evidence from large individual participant level meta-analyses supports the potential value of even modest systolic blood pressure reductions when implemented at scale. A meta-analysis of 344 716 participants from 48 randomised clinical trials found that each 5 mmHg reduction in systolic blood pressure was associated with approximately a 10% lower risk of major cardiovascular events (hazard ratio 0.91, 95% CI 0.89–0.94 in people with no cardiovascular disease; 0.89, 0.86–0.92 in people with prior disease).9 Another meta-analysis of 358 707 participants from 51 trials showed benefits across a wide range of baseline blood pressure and age groups, including individuals aged 80 years or older.10 These findings reinforce Geoffrey Rose’s principle that small shifts in population wide risk factors can lead to substantial aggregate health gains.11 The use of a causal forest model by Inoue and colleagues is a strength of the study. This machine learning approach allowed the authors to explore heterogeneity in treatment effects across subgroups, a growing area of interest in health policy and precision medicine. The study also builds on a randomised controlled trial design, enhancing internal validity and minimising confounding.

ngth of the study. This machine learning approach allowed the authors to explore heterogeneity in treatment effects across subgroups, a growing area of interest in health policy and precision medicine. The study also builds on a randomised controlled trial design, enhancing internal validity and minimising confounding. Machine learning models are increasingly used to improve prediction and subgroup identification in cardiovascular and diabetes care. Recent work has shown that such models can outperform traditional risk scores in complex patient populations.12 However, translating these methods into real-world policy making remains challenging. Their application requires high quality, individual level data and robust external validation, which are not routinely available in Medicaid administrative data. Several limitations of the current study warrant attention. Firstly, the findings are based on a single state experiment with relatively short follow-up, limiting generalisability to broader populations. Secondly, while the model identifies individuals most likely to benefit, implementing this approach would require access to detailed clinical and behavioural information that is often unavailable in routine data sources. Thirdly, the mechanisms behind the observed systolic blood pressure reduction remain unclear. Finally, while expanded coverage may reduce financial barriers and improve access to care, the lack of significant change in glycated haemoglobin suggests that broader cardiometabolic improvements were not observed within the study period.

mechanisms behind the observed systolic blood pressure reduction remain unclear. Finally, while expanded coverage may reduce financial barriers and improve access to care, the lack of significant change in glycated haemoglobin suggests that broader cardiometabolic improvements were not observed within the study period. The results also reflect a common challenge in evaluating health policy interventions. Modest but statistically significant findings must be interpreted with caution. Medicaid can provide an essential foundation for preventive care, especially in underserved populations. However, insurance coverage alone may not be sufficient to drive clinically meaningful outcomes without complementary interventions focused on diagnosis, treatment adherence, and risk reduction. Evidence from real-world data supports this view. Analyses of Medicaid expansion have shown improved access to care and management of chronic conditions. Still, the observed benefits often vary by state, population, and outcome. Context matters, and targeted implementation strategies are likely needed to realise the full potential of coverage expansion. Further research should aim to identify which populations benefit most from Medicaid coverage and through which pathways. Understanding the behavioural, clinical, and social mechanisms that mediate these effects will be crucial for designing effective and equitable interventions. Longitudinal studies with more detailed measurement of healthcare use, medication adherence, and lifestyle factors may help to clarify the link between coverage and outcomes.

the behavioural, clinical, and social mechanisms that mediate these effects will be crucial for designing effective and equitable interventions. Longitudinal studies with more detailed measurement of healthcare use, medication adherence, and lifestyle factors may help to clarify the link between coverage and outcomes. In conclusion, the updated findings from Inoue and colleagues suggest that Medicaid coverage may lead to modest reductions in systolic blood pressure among specific subgroups. However, the broader implications for clinical practice and policy are limited by uncertainty regarding clinical significance, mechanisms of action, and applicability beyond the study setting. These findings highlight both the potential and the complexity of using insurance coverage as a tool to improve cardiovascular health. They also reinforce the need for careful, transparent evaluation of policy interventions to ensure that they are both effective and equitable. Competing interests: The BMJ has judged that there are no disqualifying financial ties to commercial companies. CY and SJT reported a research grant from the Canadian Institutes of Health Research (grant number 177747). No other competing interests declared.