Study design and participants
The FIMA study was a randomized, controlled intervention study involving a comparison between a physician led interprofessional medication assessment and usual care in public home care settings. The complete study design of the FIMA study has been published previously . The Research Ethics Committee of Northern Savo Hospital District and Kuopio University Hospital approved the FIMA study protocol on February 3, 2015. The FIMA study was registered with Clinical Trials.gov on March 20, 2015 (identifier: NCT02398812). Reporting follows the CONSORT 2010 statement.
We screened and recruited patients receiving regular home care services in the study areas. The inclusion criteria were age ≥ 65 years and registration to public home care services, and at least one of the following: ≥ 6 medicines in use, dizziness, orthostatic hypotension or a recent fall. We excluded patients whose medication was not managed by the home care and patients undergoing active cancer therapy.
A total of 512 patients were recruited (Figure 1). The characteristics of the participants have been described previously . Written informed consent was obtained from all individual patients included in the study or their closest proxy if the patient had cognitive impairments. After baseline measurements, patients were randomized to receive the intervention or care-as-usual. Intervention and usual care groups were treated similarly except for the interprofessional medication assessment.
Medication use and patients’ diagnoses were verified according to a structured protocol [11, 12]. Comorbidity was defined according to a modified Charlson Comorbidity Index (CCI) . Performance in daily activities, the patient’s physical and cognitive performance, depressive symptoms and quality of life were assessed. Sociodemographic variables were also collected.
The structured medication assessment included a review of medication, an assessment of the clinical information, and a meeting of an interprofessional team consisting of a pharmacist, physician and registered nurse working regularly in home care; they conducted the medication assessment within two weeks after the baseline measurements. Patients’ updated and verified medication lists, baseline measurements, and electric medical records including their medical histories were available during the assessment.
Before the team meeting, the pharmacist reviewed the patients’ medication lists using four databases: SFINX® (currently INXBASE®) for drug-drug-interactions, PHARAO® (currently RISKBASE®) that complements SFINX® with regard to 11 clinically relevant adverse effects, RENBASE® for renal risks  and the Database of Medication for the Elderly (Meds75+) . The physician gathered information from medical records and on the patients’ current clinical status.
The interprofessional team meeting discussed the patient’s current health status and functioning and reviewed the patient’s medications. The physician made clinical decisions and wrote conclusions into the patient’s medical records at the end of the team meeting. The nurse updated the medication lists and informed patient about the changes; if necessary, the patient could participate in the interprofessional team meeting. The average time for the interprofessional team meeting was 20 minutes, with the structured review done by the pharmacist lasting 27 minutes.
All pharmacists had a qualification in comprehensive medication review or current continuing professional development in clinical pharmacy. All interprofessional team members undertook a one-day training course or a personal introduction concerning the FIMA protocol.
Patients randomised to usual care did not receive an interprofessional medication assessment. Information on their medication use was collected in a similar manner as in the intervention group but their baseline medication lists were reviewed by a pharmacist only after the six-month measurements had been conducted.
Katz index of Activities of Daily Living (ADL)  and the Lawton and Brody scale of Instrumental Activities of Daily Living (IADL) scale  were used to evaluate each patient’s performance. The maximum score in ADL is six; in IADL it is eight, with lower scores indicating increased requirement for assistance in daily activities. The Timed Up and Go (TUG) test was used to assess mobility, lower extremity strength and balance. The time taken to complete the TUG test correlates with the level of functional mobility . The Mini-Mental State Examination (MMSE) was used for screening cognitive functions; MMSE scores ≤24 indicate impaired cognitive function . The Geriatric Depression Scale (GDS-15) was used for assessing depressive symptoms. Sum scores ≥6 are suggestive of depression . The preference-based, five-dimension instrument provided by EuroQol (EQ-5D-3L)(1)  was used for measuring health-related quality of life. These measurements were carried out at baseline and repeated at the six-month follow-up.
Data were analysed according to randomisation group irrespective of whether the patients received the intervention as planned (the intention to treat principle). Baseline characteristics of the sample were summarized using proportions, percentages, and means with standard deviation (SD).
We used linear mixed models (LMM) with a random subject effect to detect differences in ADL, IADL, TUG, MMSE, GDS-15, and EQ-5D-3L between the usual care and intervention groups. Treatment (FIMA vs. usual care), time (baseline vs. 6-month follow-up), and sex served as factors, and age and CCI (excluding dementia) at baseline served as covariates. Based on our preliminary analyses (shown as request), the final models also included a treatment-time-sex interaction. IBM® SPSS® Statistics Version 25 served as the statistical platform.