Bond 2007.

Study characteristics
Methods	Aim of study: to test the hypothesis that a comprehensive MEDMAN service would increase the proportion of patients receiving treatment according to the National Service Framework in England and Wales; would improve overall patient health status; and would be cost‐effective Study design: RCT (pharmacy/GP; unit of allocation: individual) Number of arms/groups: 2
Participants	Description: patient/consumer Geographic location: UK Setting: community pharmacy (+ primary care (GP)) Inclusion criteria: patients registered with GP > 17 years and with CHD (previous MI, angina, CABG, and/or angioplasty); pharmacies (only pharmacies with private consultation area were eligible to participate) Exclusion criteria: illiterate/innumerate, history of alcohol/drug misuse, terminal/serious illness, severe mental illness, unable to provide informed consent or otherwise unsuitable for the trial as determined by GP Number of participants randomised: 1493 (I: 980, C: 513) Number of participants included in analysis: questionnaires analysed: 712 vs 373, clinical records analysed: 868 vs 466 Age: mean ± SD intervention 68.7 ± 9.2 vs control 68.8 ± 9.1 Gender: F: 307 (32.6%) vs 147 (29.4%) Ethnicity: not specified Number of medications: prescribed medications median (IQR) of 738, intervention: 7 (5 to 10) Frailty/Functional impairment: not specified Cognitive impairment: not specified Comorbidities: not specified
Interventions	Group 1 ‐ Community Pharmacy Medicines Management (MEDMAN): patients received a study registration card and a letter asking them to visit their nominated pharmacy to initiate service. Initial consultation informed by extracted medical data supplied by researchers. Further consultations provided according to pharmacist‐determined patient need. Consultations included assessments of the following: therapy, medication compliance, lifestyle (e.g. smoking, exercise, diet), and social support (e.g. difficulties collecting prescriptions and opening bottles). Recommendations were recorded on a referral form, which was sent to the GP, who returned annotated copies to pharmacists Group 2 ‐ Usual care from GP and community pharmacy Co‐intervention: N/A Provider: pharmacist (community) Where: community pharmacy When and how often: initial consultation, then as pharmacist‐determined need Intervention personalised: yes ‐ assessment of therapy, compliance, lifestyle, social ‐ and further consultations as needed
Outcomes	Timing of outcome assessment: baseline and 12 months Medication adherence (subjective): 12 statements about medicine‐taking were summated to derive self‐reported compliance score (range 12 to 60). 12‐Item scale extended scope of MARS questionnaire, introduced a time dimension, and rephrased some questions to make them more patient friendly Knowledge about medicines (subjective): patients were asked whether they "knew more about their medicines compared with a year ago" on a 5‐point Likert scale. Dichotomous; those who said agree/strongly agree Satisfaction with intervention (subjective): responses to 15 positive and negative statements regarding their most recent pharmacy visit. Overall score 15 to 75 (higher = better) Health‐related quality of life (subjective): SF‐36 and EuroQoL‐5D Condition‐specific outcomes (objective): patients reaching CHD targets. Total score for patients reaching 8 targets (aspirin, lipid, BP, smoking, alcohol, physical activity, diet, and BMI) Cost‐effectiveness (objective): health economics analysis. Total NHS‐related study cost: NHS resource use based on information extracted from GP‐held records at baseline and at follow‐up. NHS costs included costs of intervention and other treatment (e.g. medicines, hospital, other health consultations)
Notes	Trial registration: N/A Consumer involvement: not specified Funding source: Department of Health for England and Wales, managed by National Pharmaceutical Association, Royal Pharmaceutical Society of Great Britain, Company Chemist Association, and Co‐operative Pharmacy Technical Panel, led by PSNC Dropout: before intervention: 3 died, 49 withdrew (total 52, 39 vs 13); post intervention: 38 vs 9 withdrew, 20 vs 19 died Unpublished data included: full trial report provided by trial authors
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients randomised 2:1 (intervention:usual care) independently of research team using a password‐protected computer programme in permuted blocks stratified by practice
Allocation concealment (selection bias)	Low risk	Patients consented before randomisation; randomisation done independent of research team
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not possible to blind participants nor staff to intervention. Pharmacies not told which control patients had nominated their pharmacy
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Audit clerks and researchers conducting statistical analyses were blinded to patient randomisation. Self‐reported data were collected by postal questionnaire
Incomplete outcome data (attrition bias) All outcomes	Low risk	81% and 79% of questionnaires analysed. Intention‐to‐treat, but patients with missing data excluded. Potential selection bias resulting from loss to follow‐up or missing data was tested and adjusted for, using the Heckman selection correction. When evidence of selection bias was found, unbiased effect of the intervention was reported. 98 and 99% of clinical record forms analysed
Selective reporting (reporting bias)	Unclear risk	Krska paper; not as per protocol
Other bias	Unclear risk	Did not reach required sample size. Sample size calculation: 1920 (1280 vs 640)