Malet‐Larrea 2016.

Study characteristics
Methods	Cluster‐RT (pharmacies were the cluster unit of randomisation)
Participants	31; 17 intervention and 14 control, this was also the final sample that was analysed Setting: community pharmacists Diagnostic criteria: participants were ≥ 65, used ≥ 5 medications for ≥ 6 months, with the ability to complete the EuroQol 5D questionnaire
Interventions	Organisational IIntervention group: pharmacists allocated to the intervention group provided the medication with follow‐up service according to national guidelines. The medication review with follow‐up service started with a comprehensive interview undertaken in a private area of the pharmacy. The pharmacist collected relevant information about the participant’s health problems, medicines used, clinical and biological parameters (gathered through medical records provided by the participant or measured in the pharmacy), medication use, lifestyle habits and concerns about diseases and medications. Pharmacists also assessed the level of control of health problems by using information referred by participants’ and/or clinical and biological parameters, depending on the type of health problem (i.e. pain versus hyperlipidaemia) and classified every health problem as controlled, uncontrolled or unknown. After performing a comprehensive medication review, the pharmacist identified negative clinical outcomes related to medicines and drug‐related problems. Subsequently, an action plan was agreed upon by the participant and the physician if required. This medication review with follow‐up service was focused on both participants’ outcomes and medication use process and required a commitment to follow‐up. The usual care consisted of dispensing medicines prescribed by physicians and advice on minor ailments.
Outcomes	Medication‐related hospital admission was the primary outcome of this sub analysis. Hospital admissions were recorded in participants’ visits to the pharmacies and the medication related ones were identified through the expert panel after the fieldwork. Kappa values ranging from 0.61 to 1 were considered as an acceptable incidence rate ratio to measure the agreement among experts. The cost of hospital admissions estimated by diagnosis‐related group was a secondary outcome and the diagnosis‐related groups were recorded after the fieldwork.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Pharmacies were randomised to the intervention or control group by an independent researcher.
Allocation concealment (selection bias)	Low risk	An independent researcher performed randomisation using a computer‐generated list of random numbers.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Neither the participants nor pharmacists were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	There was little attrition and comparable rates for intervention and control group.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	For the sub analysis, the expert panel was blind as to which group the participants belonged so whether a hospital admission was medication‐related was not affected.
Selective reporting (reporting bias)	Low risk	No evidence of selective reporting
Protection against contamination bias	Unclear risk	Although pharmacies were randomised to control and intervention groups, informal contact between pharmacists may have led to contamination.
Other bias	Unclear risk	There is no evidence of other bias.