Multimorbidity within households and health and social care utilisation and cost

Problem

Multimorbidity prevalence is rising and people with multimorbidity often require significant interaction with health and care systems. Treatment guidelines emphasise reducing treatment burden and improving coordination of care across services. Yet recent large-scale studies of care for people with multimorbidity have not shown immediate success and so more is needed to understand how to improve their care.

As the daily responsibility for managing and coordinating their health conditions commonly falls primarily on the individuals themselves and on their informal carers (who are often in the same household), we set out to test whether the household health context matters for the use and cost of care for people with multimorbidity.

Approach

Difficulty in identifying households within electronic health records (EHRs) likely contributes to the lack of research in this area. To address this, we used a local sample of EHRs linked to detailed household composition data from local authority records and then replicated the analysis in a national study where co-residence was inferred from anonymised address data.

Linked data from health providers and local government in Barking and Dagenham provided a retrospective cohort of people aged 50+ in two-person households between April 2016 and March 2018. Two-part regression models were applied to estimate annualised use and cost of hospital, primary, community, mental health and social care by multimorbidity status of individuals and co-residents, adjusted for age, sex and deprivation. Applicability at the national level was tested using the Clinical Practice Research Datalink.

Findings

In our sample of two-person households, over 45% of multimorbid people were co-resident with another multimorbid person. They were 1·14 (95% CI 1·00, 1·30) times as likely to have any community care activity and 1·24 (95% CI 0·99,1·54) times as likely to have any mental health care activity than those co-resident with a healthy person. They had more primary care visits (8·5 (95% CI 8·2,8·8) vs 7·9 (95% CI 7·7,8·2)) and higher primary care costs. Outpatient care and elective admissions did not differ between these groups. Findings in the national data were similar.

Consequences

The study provides early evidence that some forms of care utilisation among people with multimorbidity varies by their household health context. The findings raise questions about how to deliver health and social care that acknowledges the household context for people with multimorbidity. For example, this could include scheduling community care to households or developing health care initiatives to households based on the principles of the group care approach.