A cross-sectional study of Swiss ambulatory care services use by multimorbid patients in primary care in the light of the Andersen model.

Background: Multimorbidity is frequently encountered in primary care and is associated with increasing use of healthcare services. The Andersen Behavioral Model of Health Services Use is a multilevel framework classifying societal, contextual, and individual characteristics about the use of healthcare services into three categories: 1. predisposing factors, 2. enabling factors, and 3. need factors. The present study aimed to explore multimorbid patients' use of ambulatory healthcare in terms of homecare and other allied health services, visits to GPs, and number of specialists involved. A secondary aim was to apply Andersen's model to explore factors associated with this use.

Method: In a cross-sectional study, 100 Swiss GPs enrolled up to 10 multimorbid patients each. After descriptive analyses, we tested the associations of each determinant and outcome variable of healthcare use, according to the Andersen model: predisposing factors (patient's demographics), enabling factors (health literacy (HLS-EU-Q6), deprivation (DipCare)), and need factors (patient's quality of life (EQ-5D-3L), treatment burden (TBQ), severity index (CIRS), number of chronic conditions, and of medications). Logistic regressions (dichotomous variables) and negative binomial regressions (count variables) were calculated to identify predictors of multimorbid patients' healthcare use.

Results: Analyses included 843 multimorbid patients; mean age 73.0 (SD 12.0), 28-98 years old; 48.3% men; 15.1% (127/843) used homecare. Social deprivation (OR 0.75, 95%CI 0.62-0.89) and absence of an informal caregiver (OR 0.50, 95%CI 0.28-0.88) were related to less homecare services use. The use of other allied health services (34.9% (294/843)) was associated with experiencing pain (OR 2.49, 95%CI 1.59-3.90). The number of contacts with a GP (median 11 (IQR 7-16)) was, among other factors, related to the absence of an informal caregiver (IRR 0.90, 95%CI 0.83-0.98). The number of specialists involved (mean 1.9 (SD 1.4)) was linked to the treatment burden (IRR 1.06, 95%CI 1.02-1.10).

Conclusion: Multimorbid patients in primary care reported high use of ambulatory healthcare services variably associated with the Andersen model's factors: healthcare use was associated with objective medical needs but also with contextual or individual predisposing or enabling factors. These findings emphasize the importance of adapting care coordination to individual patient profiles.