Understanding hospital activity and outcomes for people with multimorbidity using electronic health records

Background As the prevalence of multimorbidity grows, provision of effective healthcare is more challenging. Both multimorbidity and complexity in delivery of healthcare may be associated with worse outcomes. Methods We studied consecutive, unique emergency non-surgical hospitalisations for patients over 50 years old to three hospitals in Scotland, UK between 2016 and 2024 using linked primary care and hospital records to define multimorbidity (2 + long-term conditions), and timestamped hospital electronic health record (EHR) contacts with care providers to study patterns and intensity of inpatient care. The primary outcome was emergency hospital readmission within 30 days of discharge, analysed using multivariable logistic regression. Results Across 98,242 consecutive admissions, 84% of the study population had multimorbidity, 50% had 4 + long-term conditions, and 37% had both physical and mental health conditions. Both higher condition count and total number of care provider contacts were independently associated with the primary outcome in fully adjusted models (adjusted odds ratio [aOR] 1.61, 95% CI 1.51 to 1.72 for 4 + conditions compared to no multimorbidity, p < 0.001; aOR 1.10, 95% CI 1.03 to 1.17 for > 12 care provider contacts compared to 1–4, p = 0.01). While multimorbidity was associated with longer hospital stays with more contacts, the distribution of contacts and rehabilitation activity did not differ by multimorbidity or emergency readmission status. Conclusions Higher count multimorbidity was associated with an increased risk of readmission, but we observed uniformity in care despite differential outcomes across multimorbidity groups. This may suggest that EHR data-driven approaches could inform person-centred care and improve hospital resource allocation.