Salisbury 2018.

Methods	Study design: cluster‐randomised controlled trial; 2 arms: '3D' intervention; usual care. Clusters at practice level. Allocations were made in blocks of 2 per area, with an intervention and a control practice allocated simultaneously Definition of multi‐morbidity: at least 3 types of chronic conditions from a list of 11 compiled by grouping the 17 chronic conditions included in the UK Quality and Outcomes Framework (QOF) pay‐for‐performance programme into types of conditions with similar management considerations Study duration: March 2014 to September 2017; 15‐month follow‐up period
Participants	Description: patients with multi‐morbidity (797 + 749) from 33 practices: 16 intervention practices, 17 usual care practices Geographical location: Bristol and Greater Manchester in England, and Ayrshire in Scotland Setting: GP surgeries providing National Health Service (NHS) primary medical care. Patients' usual general practice and usual primary care providers Age, mean (SD): intervention group 71.0 (11.6), control group 70.7 (11.4) Number of long‐term conditions (SD): 3.2 (0.5) Gender: intervention group female 51%; control group female 50% Ethnicity: intervention group: white 99%; other or unknown 1%; control group: white 99%; other or unknown 1% Language: no English language was an exclusion criterion; no further details provided Socioeconomic status: intervention group fully retired from work 71%; control group fully retired from work 69% IMD for practices: intervention group: English 15.6 (9.6), n = 11, Scottish 24.2 (20.0), n = 5; control group: English 15.8 (12.2), n = 12, Scottish 26.4 (18.3), n = 5 Place of residence: not reported Education: not reported Frailty: not reported Mobility/functional ability: not explicitly reported; however, some aspects captured in ‘health‐related quality of life’ measure Receipt of carer support: not reported Communication vulnerability: some likely to have been excluded by eligibility criteria; however, 1% of both groups had a learning disability and 4% of both groups had dementia
Interventions	Intervention name: the ‘3D’ intervention Intervention type: patient‐focused, practitioner‐focused, and related to organisational change; a comprehensive, holistic intervention, utilising multi‐disciplinary practitioner training and organisational change with the aim to maximise continuity, co‐ordination, and efficiency of care Aims and rationale: a patient‐centred care model, encapsulating strategies recommended in recent international guidelines (ref), replacing disease‐focused review of each health condition with one 6‐monthly comprehensive multi‐disciplinary review. The 3D intervention has the aim of improving continuity, co‐ordination, and efficiency of care for patients with multi‐morbidity. Using the 3D review, study authors seek to address the burden of treatment for patients with multi‐morbidity. Study authors report that if each condition is considered in isolation, patients can be prescribed numerous drugs and lifestyle changes, and are expected to attend frequent healthcare appointments Materials: an electronic template integrated within the EMIS electronic medical records system reinforces the patient‐centred approach and is interactive, with prompts for clinicians that change depending on the patient’s combination of chronic conditions. In addition, an appointment letter asked the patient to think about the health problems that bother them most, and patients were provided with a 3D wallet card, naming their responsible doctor Procedures: each 3D review consists of 2 appointments (a nurse appointment and a named responsible physician appointment) and a records‐based medication review by a pharmacist. “The nurse focuses on identifying the health problems most important to the patient; asking about pain, function, and quality of life; screening for depression and dementia; and then addressing the disease‐specific care the patient requires. Findings are printed as a patient‐held agenda to inform the subsequent consultation with the doctor. The pharmacist uses the patient’s electronic medical records to review medication, and makes recommendations about simplifying and optimising treatment. The physician considers the nurse and pharmacist reviews, discusses treatment adherence, and agrees on a collaborative health plan with the patient. The patient is given a printed copy of the plan, which specifies how the patient and clinicians will address the agreed goals over the next 6 months through routine consultations" (Procedures, pp 43‐44) Elements related specifically to patient involvement in decision‐making: training sessions with clinicians included how to identify patients’ priorities and agree on a health plan with patients that was ‘SMART’ (Specific; Measurable; Achievable; Realistic; Time‐bound). The 3D review seeks to identify the health problems of most importance to the patient, along with additional problems from the clinician’s point of view. All problems are shared with the patient after the first consultation in an ‘agenda’ document, to encourage collaboration. During a subsequent GP consultation, a holistic, collaborative plan is made to address the agenda, guided by the patient’s top 4 priorities for action. This health plan is documented and printed as the result of an agreed, negotiated discussion Mode of delivery: assumed patients sent appointment letter by post. Location of patient appointments not recorded but assumed face‐to‐face. Medication reviews records‐based and not appearing to involve patients directly Who delivered the intervention? Each practice identified a local champion to support implementation. At practice level, changes were made to flag patients with multi‐morbidity in electronic medical records systems; promote continuity of care; provide a 3D wallet card for patients, naming the responsible doctor; encourage patients to ask for longer appointments when needed; train clinicians and receptionists; replace separate disease‐focused reviews with one 6‐monthly whole‐person 3D review; and implement an interactive 3D electronic data entry template. Each patient was allocated a usual responsible GP and nurse, who would conduct the 3D reviews. All practice clinical staff involved in delivering the intervention received 2 half‐days of training. Administrative staff were trained in a separate meeting Where was the intervention provided? The intervention was delivered at the patient's usual GP practice When and how often or how much of the intervention was provided? The intention was for patients to receive 2 3D reviews within a 12‐month period, with reviews carried out every 6 months Was the intervention tailored? Disease‐specific questions and prompts (presented to clinicians via the electronic template) varied depending on the patient's combination of conditions Was the intervention modified or adapted? The intervention was not modified nor adapted
Outcomes	Health‐related quality of life Tool: patient self‐report using the EQ‐5D‐5L measure (Herdman 2011); higher scores = higher perceived quality of life Review category: health status Timing of measurement: baseline, at 9 months (data not extracted), and at 15 months (Results have been recalculated for analysis using data as presented in the trial report) Treatment burden Tool: patient self‐report using the Multi‐morbidity Treatment Burden Questionnaire (MTBQ) (Duncan 2018); higher score = higher perceived burden Review category: treatment burden Timing of measurement: baseline, at 9 months (data not extracted), and at 15 months (Results have been recalculated for analysis using data as presented in the trial report) Medication adherence Tool: patient self‐report using the Morisky Medication Adherence Scale (Morisky 2008); higher score = greater adherence Review category: patient enablement and engagement Timing of measurement: baseline and at 15 months (Results have been recalculated for analysis using data as presented in the trial report) Patient‐centred care: proportion who discussed the problems most important to them Tool: patient self‐report using a single question and a 5‐point scale, adapted from the General Practice Patient Survey (Ipsos MORI 2015); higher score = higher perception of discussing priorities Review category: patient evaluation of care/the intervention Timing of measurement: baseline, at 9 months (data not extracted), and at 15 months (Results have been recalculated for analysis using data as presented in the trial report) Carer experience Tool: carer self‐report using a 6‐item carer experience questionnaire (Al‐Janabi 2011); score 0 to 100; higher scores = better experience with the health care provided Review category: patient and carer outcomes Timing of measurement: baseline and at 15 months (Results have been recalculated for analysis using data as presented in the trial report) Number of consultations in primary care: nurse Tool: routine review of medical records by the research team Review category: resource use and cost Timing of measurement: baseline and at 15 months (Results have been recalculated for analysis using data as presented in the trial report) Number of indicators of high‐risk prescribing Tool: review of routine electronic health records using an approach developed for a previous trial (Dreischulte 2016) Review category: quality of care Timing of measurement: at 15 months (Results have been recalculated for analysis using data as presented in the trial report)
Control or usual care	Usual care. In UK, mainly consists of: nurse‐led disease‐specific chronic disease reviews focused on meeting the outcomes of the UK Quality and Outcomes Framework pay‐for‐performance scheme
Funding source and potential conflicts of interest	Funding: National Institute for Health Research, Health Services and Delivery Research programme, project number 12/130/15 Potential conflicts of interest: One study author is a member of the NIHR Health Services and Delivery Research Board Another study author chairs the Guideline Development Group of the National Institute for Health and Clinical Excellence (NICE) Multimorbidity Clinical Guideline NG56 and was a member of the NIHR Health Services and Delivery Research researcher‐led panel All other study authors declared no competing interests
Notes	Other outcomes, considered of less relevance to the review: Burden of illness: self‐rated health Burden of illness: how much illness affects the individual's life Burden of illness: anxiety Burden of illness: depression Polypharmacy: number of different drugs prescribed in last 3 months Patient‐centred care: doctor Patient‐centred care: nurse Patient‐centred care: proportion experiencing care as joined up Patient‐centred care: proportion having written a care plan, health plan, or treatment plan Satisfaction with care Continuity of care: Continuity of Care Index Discontinuity of care: Visit Entropy measure Number of consultations in primary care: physician Number of consultations and secondary care: hospital admissions Number of consultations and secondary care: outpatient attendance Additional cost‐effectiveness data Quality of disease management: proportion of applicable QOF chronic disease targets met Qualitative process assessment (reported elsewhere)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sequence generation appears sufficient to ensure random assignment of participants "The randomisation system was run from the Bristol Randomised Trials Collaboration by the trial statistician, who was masked to practice identifiers. Allocations were done in blocks of two in each area, with an intervention and a control practice allocated simultaneously"
Allocation concealment (selection bias)	Low risk	Randomisation carried out after practice recruitment "Patients were assessed for eligibility and invited to participate before practice allocation, and were not informed of their practice’s allocation until they had given consent and completed baseline measures. ... The randomisation system was run from the Bristol Randomised Trials Collaboration by the trial statistician, who was masked to practice identifiers. Allocations were done in blocks of two in each area, with an intervention and a control practice allocated simultaneously so that concealment of allo­ cation was maintained"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not done. Some potential for this to influence outcomes but not clear if this occurred "Because of the nature of the intervention, practices and participants were aware of their treatment allocation"
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Some blinding/automation but not for all outcomes "Outcome data were self­reported or based on automated extraction of data from the electronic medical records, except for details of hospital use, which were collected manually by researchers who were aware of practice allocation. Analysis was done by the trial statistician (DG), who was masked to allocation, except for details of health­care use, for which masking could not be maintained"
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Intention‐to‐treat analysis; attrition not reported; lack of data for some items and some imbalance between groups but not enough to be classed as high risk; method of imputation reported (for the primary outcome) "We used multiple imputation by chain equations including baseline, 9­month, 15­month, and EQ­5D­5L data as available, intervention group, stratification and minimisation variables, and other covariates that were informative of missingness"
Selective reporting (reporting bias)	Low risk	Protocol available. All pre‐specified outcomes reported either in the paper or in the full report
Other bias	Low risk	Selective recruitment of cluster participants: Cluster participants were recruited before group assignment, and the same participants were followed up over time; therefore low risk of selective recruitment of cluster participants; no other obvious sources of bias "Patients were assessed for eligibility and invited to participate before practice allocation"

3D: intervention that was both patient‐ and practitioner‐focused and addressed organisational change.

CBT: cognitive‐behaviour therapy.

CDSMS: Chronic Disease Self‐Management Support programme.

EHR: electronic health record.

EQ‐5D‐5L: EuroQoL Group Quality of Life Questionnaire based on a five‐level scale.

GP: general practitioner.

IMD: XXX.

ITT: intention‐to‐treat.

MI: motivational interview.

MTBQ: Multi‐morbidity Treatment Burden Questionnaire.

NHS: National Health Service.

NICE: National Institute for Health and Care Excellence.

QOF: Quality and Outcomes Framework.

SD: standard deviation.

SE: standard error.

SMART: Specific; Measurable; Achievable; Realistic; Time‐bound plan,