Association between severe maternal morbidity and high health service use in the first postpartum year: population-based study in British Columbia, Canada

Background: Severe maternal morbidity (SMM) includes life-threatening complications occurring during pregnancy, at delivery, or within 42 days thereafter, affecting 1.5% - 2.0% of births in Canada. Health service utilization patterns in the year following SMM are unknown. In this study, we compared high health service use from 43 days until 365 days postpartum in SMM-affected and unaffected pregnancies in British Columbia (BC), Canada. Objectives: 1) Assess associations between SMM and composite high health service use between days 43 and 365 post-delivery.2) Assess associations between SMM and use of each individual healthcare visit subtype. 3) Assess whether there is a difference in the association between SMM and high health service use based on presence or absence of (i) a 6-week postpartum visit; (ii) pre-existing chronic comorbidities; (iii) preterm birth; (iv) substance use; (v) multiple gestation; (vi) previous high health service use in the 2 years prior to delivery.Methods: This retrospective cohort study used de-identified, linked administrative datasets within Population Data BC including the BC Perinatal Database Registry, the Discharge Abstract Database (DAD), the Medical Services Plan (MSP), the National Ambulatory Care Reporting System (NACRS), the BC Consolidation file, the Chronic Disease Registry, and the BC Vital Statistics Deaths dataset. Our cohort included all hospital deliveries in BC residents from 2013 to 2020 captured from within the BC perinatal database. We included records with complete identifiers for individuals 18 or older, excluding maternal deaths during pregnancy, at delivery or within 42 days postpartum. SMM was a composite of diagnostic and procedure codes, defined using the Canadian perinatal surveillance system as occurring from obstetric hospitalizations at 20 weeks’ gestation until delivery and up to 42 days postpartum. High health service use was defined as being the top 5% in the cohort for any non-obstetric health service utilization (i.e., a composite of emergency department (ED) visits, hospitalizations, visits to a primary care provider (PCP) and specialist visits identified using the DAD, NACRS, and MSP). We then measured the association between SMM and high health service use using log-binomial regression with generalized estimating equations to account for clustering by individual with >1 pregnancy during the study period. We adjusted for maternal age, parity, income, urban vs rural/remote residence, attendance at 1st trimester antenatal visit, pre-existing chronic conditions within 2 years prior to delivery, tobacco use, and substance use during pregnancy. Analyses were repeated for specific types of health service use (ED, hospital, PCP, and specialist) and with each SMM subtype as the exposure of interest (sepsis, hemorrhagic, hypertensive, non-hypertensive cardiovascular conditions, surgical-related morbidity, intensive care unit admission, assisted ventilation, and other SMM). Stratified analyses were conducted according to the absence or presence of the following pregnancy-related factors: a 6-week postpartum visit, preterm birth (<37 weeks’ gestation), multiple gestation. We also stratified analysis according to indicators of general health captured within two years prior to delivery; namely, previous high health service use, chronic morbidity, tobacco use, and substance use. We conducted a sensitivity analysis by repeating analyses after excluding those who died or moved out of BC in the first postpartum year (n=2930). To assess the potential impact of Covid-19 on our results, we restricted births to those occurring in the pre-pandemic era (April 1st, 2013, to February 28th, 2019). Additionally, we assessed the 90th and 99th percentiles as alternative thresholds to define high health service use.Results: The cohort included 261,287 deliveries (5575 (2.1%) with SMM). SMM-affected individuals were twice as likely to have high health service use (> 15 visits) between 43 and 365 days postpartum; 511/5575 (9.2%) vs. 10890/255712 (4.3%); adjusted relative risk [aRR] = 1.96, 95% CI 1.78-2.17). This pattern was true for all health service types but most prominent for specialist visits (> 3 visits); 489/5575 (8.8%) with high specialist use with SMM vs. 9162/255712 (3.6%) with high specialist use without SMM; aRR = 2.21, 95% CI 1.99-2.45. Of the SMM subtypes, non-hypertensive cardiovascular SMM was most strongly linked with high overall health service use (33/154, 21.4% vs. 10890/255712, 4.3%; aRR= 5.18, 95% CI 3.28-8.16). There was marked heterogeneity in the association of SMM and high health service use for those with term versus without preterm birth (aRR= 2.02, 95% CI 1.80-2.28 vs. aRR = 1.46, 95% CI 1.22-1.74), for those without versus with prior chronic conditions (aRR= 2.34, 95% CI 2.03-2.70 versus aRR = 1.76, 95% CI 1.54-2.01), for those without versus with a visit to a specialist in the 2 years prior to delivery (aRR= 2.37, 95% CI 2.01-2.79 vs. aRR = 1.73, 95% CI 1.53-1.96), and for those without versus with high health service use in the 2 years prior to delivery (aRR = 2.07, 95% CI 1.83-2.33 vs. aRR = 1.34, 95% CI 1.11-1.62). There were no appreciable differences in our results after removing postpartum deaths and migrations out of BC from the cohort, or when restricting analyses to the pre-pandemic era. Using the 90th percentile to define high health service use yielded slightly weaker associations than the main analysis (95th percentile), while the 99th percentile produced wider CI due to fewer individuals meeting the threshold.Conclusion: Our study revealed high use of health services between 43 and 365 days after delivery in SMM compared with unaffected pregnancies. In this study we have quantified gaps and healthcare needs for women following SMM which can be leveraged to adapt postpartum health service provision

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.