Objective
To determine whether longitudinal health data accounts for end‐organ injury or death in the setting of chronic hypertension.
Design
Cohort of 64 799 deliveries to 61 854 women.
Setting
US claims data for the preiod 2008–2019.
Population
Women with a delivery hospitalisation and chronic hypertension.
Methods
Risk for a composite of acute end‐organ injury or death during the delivery hospitalisation and 30 days postpartum was analysed. Adjusted logistic regression models were derived with discrimination for each model estimated by the C‐statistic. Poisson regression was used to estimate adjusted risk ratios. Starting with models using data from pregnancy, further adjustment was performed accounting for healthcare use in the year prior to pregnancy, including hospitalisations, emergency department encounters, prescription medications and pre‐pregnancy diagnoses.
Main outcome measures
Acute end‐organ injury or death.
Results
The composite outcome occurred among 5.7% of 64 799 deliveries. For patients with commercial insurance, filling non‐hypertensive medications from ≥11 different classes, compared with none (adjusted risk ratio, aRR 4.07, 95% CI 2.86–5.79), three or more hospitalisations before pregnancy, compared with none (aRR 4.75, 95% CI 3.46–6.52), and chronic kidney disease diagnosed in the year before pregnancy (aRR 2.35, 95% CI 1.88, 2.94) were associated with increased risk. For pregnancies covered by commercial insurance, the C‐statistic increased from 0.615 (95% CI 0.599–0.630) in the model with pregnancy data only to 0.796 (95% CI 0.783–0.808) for the model additionally including healthcare use in the year before pregnancy. Findings with Medicaid were similar.
Conclusions
Prepregnancy care use predicted adverse maternal outcomes. These data may be important in risk stratification.