Wu 2006.

Study characteristics
Methods	Aim of study: to investigate the effects of compliance and periodic telephone counselling by a pharmacist on mortality in patients receiving polypharmacy Study design: RCT (unit of allocation: individual) Number of arms/groups: 2
Participants	Description: both patient/consumer and carer Geographic location: China Setting: specialist medical clinic at a hospital Inclusion criteria: non‐compliant (pharmacist assessed medical clinic records), ≥ 5 drugs on ≥ 2 consecutive visits to the clinic Exclusion criteria: non‐Cantonese dialects or a different language, had conditions that prevented effective communication (deaf, mute, dementia, psychological disorders), living in nursing homes with supervised treatment Number of participants randomised: 442 (219 and 223) Number of participants included in analysis: 442 (219 and 223) (but 43 died by follow‐up) Age: 71.2 ± 9.4 vs 70.5 ± 11.1 Gender: female 51% vs 52% Ethnicity: not specified ‐ all spoke Cantonese Number of medications: drugs for chronic illnesses: 6.0 ± 1.3 vs 5.9 ± 1.2 Frailty/Functional impairment: not specified Cognitive impairment: dementia excluded Comorbidities: not specified
Interventions	Group 1 ‐ Telephone counselling by a pharmacist: patients received a 10‐ to 15‐minute telephone call from pharmacist at midpoint between clinic visits (6 to 8 calls over 2 years). Pharmacist asked about patient's treatment regimens, clarified any misconceptions, explained side effects, reminded of next clinic appointment, reinforced importance of compliance Group 2 ‐ Usual care: no telephone intervention Co‐intervention: all patients received 10‐ to 15‐minute educational talk by pharmacist during screening. Pharmacist determined compliance using structured questionnaire Provider: pharmacist Where: telephone to home When and how often: 6 to 8 telephone calls for 2 years Intervention personalised: yes
Outcomes	Timing of outcome assessment: screening or enrolment and 2 years Medication adherence (subjective) : patient asked if he or she had missed any doses; changed regimens in terms of dose, frequency, and timing; or had drugs left over. Compliant 80% to 120%. This information was checked against dispensing information Adverse clinical health outcomes (objective): all‐cause mortality Adverse clinical health outcomes (objective): hospitalisations: rate of admission to hospital, number of emergency department visits, and hospital stay 2 years before and after screening
Notes	Trial registration: SRCTN48076318 Consumer involvement: not specified Funding source: Hong Kong Government Health Care and Promotion Fund and MSD international grant Dropout: 25 and 38 died
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Pharmacist blinded to randomisation codes, which were computer‐generated by statistician and sealed in envelopes labelled with consecutive numbers
Allocation concealment (selection bias)	Low risk	Envelopes opened by clinic nurse in an ascending manner, and patients allocated
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding was not possible because the intervention was complex and caregivers were involved
Blinding of outcome assessment (detection bias) All outcomes	High risk	Blinding not possible; could have had blinded research review of outcomes
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	ITT; no loss to follow‐up, but 60 defaulters before randomisation, 31 of whom died
Selective reporting (reporting bias)	Low risk	Baseline reports as compliant/non‐compliant, follow‐up reports who remains compliant/non‐compliant (number can be computed); otherwise as per methods
Other bias	Unclear risk	Sample size 1067 to account for non‐compliance and achieve significant reduction in mortality ‐ not reached