Cheng 2012.

Methods	Cluster‐RCT. Participants clustered by community nurse service Setting: Thirteen community nurse services based at six different hospitals, covering a wide region of Hong Kong Funding: Not reported.  Completed as a PhD thesis Recruitment (patients): Recruited by a liaison nurse in each centre Recruitment (healthcare providers): Department heads from the 13 community nurse services recruited the nurses for the study. Inclusion criteria: at least three years nursing experience, specialist training in community nursing, and expected to work at the service for the duration of the study Training/support: Training was provided to all nurses participants in the study, and involved both classroom learning and community‐based case examples.  Knowledge and application of the training was formally evaluated through paper‐based and home‐based assignments.  All nurses worked under the supervision of a senior ranked nurse (e.g. an Advanced Practice Nurse, Manager or Nursing Officer) Inclusion criteria (patients): Participants needed to have had a chronic illness and be newly referred to community nursing; aged 18 or over; cognitively intact (able to express complaints, and discomfort). Exclusion criteria (patients): living and receiving care in an institution; diagnosis of acute confusion, acute psychiatric illness, dementia, terminal illness such as end stage renal failure; referral for a one‐off procedure such as blood‐taking Consideration of people with comorbidities: More the 70% of the enrolled patients had more than one active medical problem.  Mean number of medical diagnoses was 2.4 (SD 1.3)
Participants	Patients: 96 participants (37.5% male). Mean age 73 year (SD 11); 100% Chinese Principle health problems: hypertension (46.9%); diabetes mellitus (22.9%); heart failure (20.8%); cancer (17.7%); osteoarthritis (11.5%); chronic lung disease (10.4%) Treatment currently receiving: Community nursing Description of healthcare providers: Thirteen community nurses (registered nurses and enrolled nurses, enrolled in a ratio of 4:1)
Interventions	Study aim: To test the hypotheses that patients with chronic health conditions receiving community nursing structured around mutual goal setting would have higher rates of goal achievement, better functional outcomes, better perceived health status, higher self‐efficacy, and lower health service utilisation in comparison to a control group of patient receiving community nursing without mutual goal setting Intervention: (n = 53) Routine community nursing plus seven sessions of mutual goal setting over eight weeks.  Goals were collaboratively developed with each patient, starting with a pre‐determined list of common goals, with the option of adding new goals if required. GAS scales were developed for all goals, with an expert panel of nurses involved in the development of any new scales. A goal setting record was documented and signed by both the nurse and patient, and reviewed in follow up meetings Control: (n = 43) Routine community nursing without mutual goal setting. Care goals were documented in the Community Based Nursing Service information system, but were not established through a process of mutual goal setting with each patient (i.e. these goals were set by nurses, not discussed with patients, and just used to document an expected outcome arising as part of the care plan) Delivery: Mutual goal setting was delivered over eight‐week. The duration of delivery of the routine nursing care was not reported Fidelity: All nurse participants worked under the supervision of a senior ranked nurse. Delivery of the intervention was monitored by observation and audit of the study records Consumer involvement outside of the intervention: None reported
Outcomes	Timing of outcomes: Four, 12, and 24 weeks after baseline data collection. Goal achievement (percentage of goals achieved) Chronic Disease Self‐efficacy Scale Disability Index of the Health Assessment Questionnaire Perceived health status, measured using the World Organization of National Colleges, Academics, and Academic Associations of General Practices/Family Physicians Charts Satisfaction Scale in Community Nursing Number of emergency department visits Days of hospitalisation Mortality Hospital readmissions
Goal setting characteristics	Comparison of interest: Collaborative goal setting plus strategies to enhance goal pursuit versus usual care with no structured or required approach to goal setting Patient involvement in goal setting: Goals collaboratively set by the patient and healthcare professional Family involvement in goal setting: None reported Name of goal setting approach: GAS within the context of King's theory of goal attainment (King 1981) Development of a plan for goal pursuit: A plan for goal pursuit was collaboratively developed by both the nurse and patient Written copy of goals provided to patients: Not reported Level of goal difficulty: The setting of realistic goals was emphasised Goal areas of focus: Mixed, including activity limitations, participation restrictions, body structure and function, health knowledge, health behaviour, psychological state
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation sequence was generated 'using computerized software (System Randomizer)' (p. 87)
Allocation concealment (selection bias)	High risk	Concealment of group allocation during patient recruitment was not possible due to the use of cluster randomisation. Recruitment of patients was undertaken by a liaison nurse in each centre who would have known whether the centre was in the experimental or control group, therefore which group the patients were going into.
Blinding of participants and personnel (performance bias) All outcomes	High risk	The intervention required active involvement of the patient and healthcare professionals, so blinding not possible
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	'The raters had no knowledge as to which study group the patients belonged' (p. 141). However, a number of the outcomes were self‐reported by participants who were not blinded to group allocation
Incomplete outcome data (attrition bias) All outcomes	High risk	Outcome data were not collected for 28% (27/96) of the sample population. For 18% of the study population this was due to death or hospitalisation before the 24‐week assessment point. 'Findings of this study were analyzed by only those who completed the study at the 24‐week follow‐up' (p. 157)
Selective reporting (reporting bias)	Unclear risk	Unable to find a protocol for this study published prior to it being conducted, so unable to compare the outcomes reported with those planned to be measured at the outset
Other bias	Low risk	Risk of cross‐group contamination low due to use of cluster‐RCT methods. No evidence of other sources of bias