Accounting for Blood Pressure Seasonality Alters Evaluation of Practice-Level Blood Pressure Control Intervention

Background: Despite the large body of literature evaluating interventions to improve hypertension management, few studies have addressed seasonal variation in blood pressure (BP) control. This underreported phenomenon has implications for interpreting study findings and informing clinical care. We share a methodology that accounts for BP seasonality, presented through a case study - HealthyHearts NYC, an intervention aimed at increasing adherence to the Million Hearts BP control evidence-based guidelines in primary care practices. Methods: We used a randomized stepped-wedge design (n = 257 practices). Each intervention included 13 visits from practice facilitators trained in improving practice-level BP control over 12 months. Two models were used to assess the intervention effect - one that did not account for seasonality (model 1) and one that did (model 2). Model 2 was a re-specification of model 1 to include our proposed two fixed-effects terms to address BP seasonality. Results: Model 1 showed a significant negative association between the intervention and BP control (IRR = 0.98, 95% CI = 0.96-0.99, P ≤ 0.05). In contrast, Model 2, which did address seasonality, showed no intervention effect on BP control (IRR = 0.99, 95% CI = 0.97-1.01, P = 0.19). Conclusions: These findings reveal that analyses that do not account for BP seasonality may not present an accurate picture of intervention effects. In our case study, accounting for BP seasonality turned a negative association into a null association. We recommend that when evaluating BP control, studies compare outcome measures across similar seasons and that the measurement period last long enough to account for seasonal effects.