Chrischilles 2014.

Study characteristics
Methods	Aim of study: to examine the impact of a personal health record (PHR) on medication use safety among older adults Study design: RCT (single‐centre open‐label parallel group study with unequal randomisation (3:1); unit of allocation: individual) Number of arms/groups: 2
Participants	Description: patient/consumer Geographic location: USA Setting: patient's home (online) Inclusion criteria: age 65+, used a computer in past month to visit websites or to send/receive email, responded to questionnaire Number of participants randomised: 1163 Number of participants included in analysis: 1075 (802 vs 273) Age: mean ± SD 72.5 ± 6.0 vs 72.0 ± 6.3 Gender: female: 461 (57.5%) vs 150 (54.9%) Ethnicity: non‐Hispanic white: 782 (99%) vs 267 (98.2%) Number of medications: mean ± SD: prescription: 4.1 ± 3.2 vs 4.2 ± 3.2, OTC: 4.1 ± 2.8 vs 4.3 ± 3.1 Frailty/Functional impairment: physical health (SF‐12): 45.9 ± 10.6 vs 46.1 ± 10.3 Cognitive impairment: memory problems: 80 (10%) vs 31 (11.4%) Comorbidities: medical conditions (from list of 19): 3.6 ± 2.3 vs 3.6 ± 2.2
Interventions	Group 1 ‐ Personal Health Records (PHRs): participants sent an invitation to use study PHR for a period of 1 year and a quick‐start guide. Users could enter, view, and print their current and past medicines, allergies, health conditions, and health event tracking over time. PHR also had user‐friendly medication safety messages based on the Assessing Care of Vulnerable Elders project (ALCOVE‐3) medication use quality indicators. PHR displayed a message when a user entered a medication with an associated ALCOVE‐3 safety concern (drug‐drug interactions e.g. warfarin; dosage concerns e.g. acetaminophen; important lab monitoring e.g. loop diuretics; risk awareness e.g. NSAIDs, bleeding; drugs that should be avoided e.g. barbiturates). Three levels of increasing detail ‐ brief alert, summary level, and detailed explanation. Participants who did not log in were sent a reminder letter 3 to 4 weeks after initial invitation Group 2‐ Usual care: no access to study PHR Co‐intervention: N/A Provider: online Where: online (with baseline and follow‐up questionnaires mailed) When and how often: continuous for 1 year at patient discretion Intervention personalised: individual medication‐specific messages
Outcomes	Timing of outcome assessment: baseline and 6 months Medication adherence (subjective): modified Morisky self‐reported adherence. Answers never, rarely, sometimes, often, always (instead of original yes/no). Mailed questionnaire Adverse clinical health outcomes (subjective): experienced medication side effects in past 3 months (yes/no ‐ reported as percentage of participants experiencing side effects) Other (subjective): medication management problems: mean (SD) number of medication management problems. List of 8 problems including questions on multiple prescribers, multiple pharmacies, mail‐order prescriptions, confusion whether medication was taken, taking medication without knowing indication, problems affording medications, feeling that medications are not working, feeling that medications are not doing what they were intended to do
Notes	Trial registration: NCT02012712 Consumer involvement: PHR was developed and refined using participatory design and focus group sessions with older adults, as well as evaluation in a usability laboratory Funding source: Agency for Healthcare Research and Quality grant and National Institutes of Health grant Dropout: 23 before intervention, 65 lost to follow‐up Fidelity: 61.2% attempted to log on to PHR; 55.2% performed some activity with PHR. More than 40% entered at least 1 medication into PHR
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomised in a 3:1 ratio using computerised random numbers. Groups comparable
Allocation concealment (selection bias)	Low risk	Notification of study group assignment was sent by mail to all trial participants by an investigator with no clinical involvement in the trial
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not blinded ‐ participants knew allocation
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Mailed questionnaires and online results; no outcome assessors
Incomplete outcome data (attrition bias) All outcomes	Low risk	1075 of 1163 included in analysis; ITT
Selective reporting (reporting bias)	Low risk	Reported as per methods
Other bias	High risk	61.2% attempted to log on to PHR; 55.2% performed some activity with PHR. More than 40% entered at least 1 medication. ‐ so poor fidelity of intervention. Reimbursed for completing baseline and follow‐up questionnaires