Interventions to change the behaviour of health professionals and the organisation of care to promote weight reduction in overweight and obese adults

Gerd Flodgren; Katherine Deane; Heather O Dickinson; Sara Kirk; Hugh Alberti; Fiona R Beyer; James G Brown; Tarra L Penney; Carolyn D Summerbell; Martin P Eccles

doi:10.1002/14651858.CD000984.pub2

. Author manuscript; available in PMC: 2014 Nov 18.

Published in final edited form as: Cochrane Database Syst Rev. 2010 Mar 17;(3):CD000984. doi: 10.1002/14651858.CD000984.pub2

Interventions to change the behaviour of health professionals and the organisation of care to promote weight reduction in overweight and obese adults

Gerd Flodgren ², Katherine Deane ³, Heather O Dickinson ¹, Sara Kirk ⁴, Hugh Alberti ⁵, Fiona R Beyer ⁶, James G Brown ⁷, Tarra L Penney ⁸, Carolyn D Summerbell ⁹, Martin P Eccles ¹

PMCID: PMC4235843 EMSID: EMS57069 PMID: 20238311

Abstract

Background

The prevalence of obesity is increasing globally and will, if left unchecked, have major implications for both population health and costs to health services.

Objectives

To assess the effectiveness of strategies to change the behaviour of health professionals and the organisation of care to promote weight reduction in overweight and obese people.

Search methods

We updated the search for primary studies in the following databases, which were all interrogated from the previous (version 2) search date to May 2009: The Cochrane Central Register of Controlled Trials (which at this time incorporated all EPOC Specialised Register material) (The Cochrane Library 2009, Issue 1), MEDLINE (Ovid), EMBASE (Ovid), CINAHL (EBSCO), and PsycINFO (Ovid). We identified further potentially relevant studies from the reference lists of included studies.

Selection criteria

Randomised controlled trials (RCTs) that compared routine provision of care with interventions aimed either at changing the behaviour of healthcare professionals or the organisation of care to promote weight reduction in overweight or obese adults.

Data collection and analysis

Two reviewers independently extracted data and assessed study quality.

Main results

We included six RCTs, involving more than 246 health professionals and 1324 overweight or obese patients. Four of the trials targeted professionals and two targeted the organisation of care. Most of the studies had methodological or reporting weaknesses indicating a risk of bias.

Meta-analysis of three trials that evaluated educational interventions aimed at GPs suggested that, compared to standard care, such interventions could reduce the average weight of patients after a year (by 1.2 kg, 95% CI −0.4 to 2.8 kg); however, there was moderate unexplained heterogeneity between their results (I² = 41%). One trial found that reminders could change doctors’ practice, resulting in a significant reduction in weight among men (by 11.2 kg, 95% CI 1.7 to 20.7 kg) but not among women (who reduced weight by 1.3 kg, 95% CI −4.1 to 6.7 kg). One trial found that patients may lose more weight after a year if the care was provided by a dietitian (by 5.6 kg, 95% CI 4.8 to 6.4 kg) or by a doctor-dietitian team (by 6 kg, 95% CI 5 to 7 kg), as compared with standard care. One trial found no significant difference between standard care and either mail or phone interventions in reducing patients’ weight.

Authors’ conclusions

Most of the included trials had methodological or reporting weaknesses and were heterogeneous in terms of participants, interventions, outcomes, and settings, so we cannot draw any firm conclusions about the effectiveness of the interventions. All of the evaluated interventions would need further investigation before it was possible to recommend them as effective strategies.

Medical Subject Headings (MeSH): Body Weight, Controlled Clinical Trials as Topic, Delivery of Health Care [organization & administration; standards], Obesity [psychology; *therapy], Overweight [psychology; therapy], Patient Education as Topic, Professional Practice [organization & administration; *standards], Randomized Controlled Trials as Topic, Weight Loss

MeSH check words: Adult, Female, Humans, Male

BACKGROUND

Description of the condition

The prevalence of obesity is increasing, both in the developed and developing world and if left unchecked, it will have huge implications for population health and for health services expenditure in coming decades (WHO 1998; WHO 2004). Benefits of weight loss for obese people have been shown in short-term studies, as measured by reduction in cardiovascular risk factors (for example lipids, insulin, and blood pressure) and improvement in psychological status (Garrow 2000).

Description of the intervention

Information on the effectiveness of interventions to promote weight loss in patients is available. Although there are gaps in the evidence, a number of potentially effective weight loss interventions have been identified: diet, exercise, and behavioural strategies for adults, in combination where possible; the use of maintenance strategies such as continued therapist contact; selected use of pharmaceutical interventions in conjunction with strategies to change lifestyle; and surgery for selected morbidly obese patients (Colquitt 2005; Curioni 2006; NICE 2006; Padwal 2003; Shaw 2005; Shaw 2006; Thomas 2007).

The extent to which health professionals deliver such interventions within routine healthcare is uncertain. In the past, health professionals’ application of effective patient weight loss strategies may have been limited because of an abundance of research of variable quality with no consistent or clear conclusions, other than an apparent pessimism about the long-term effectiveness of treatments overall. Even with the availability of reviews stating the effectiveness of patient interventions (Douketis 1999; EHCB 1997; Glenny 1997; NHLBI 1998; O’Meara 2000), health professionals may be inconsistent in their application of such guidelines in routine care, often citing barriers such as lack of time, lack of access to the guidelines, or lack of confidence in the guidelines’ conclusions and their relevance to their clinical practice (Cabana 1999). Other potential barriers to effective management of obesity may include lack of access to appropriate support services and a lack of motivation to work with this patient group due to negative perceptions of overweight and obese people or of the efficacy of treatments (Frank 1993; HEA 1995; Price 1987; Puhl and Brownell 2001; Puhl and Heuer 2009; Summerbell 1998).

Interventions aimed at improving the way healthcare professionals work to reduce the weight of people who are obese or overweight can be divided into those targeted at the health professionals themselves and those targeted at the organisation of care.

Why it is important to do this review

In the UK the proportion of obese people is increasing rapidly and, if this trend continues, the UK could be a predominantly obese society by 2050 (Foresight 2007). It is essential to develop and implement effective strategies to prevent and treat obesity at the level of the individual, family, and healthcare provider, as well as in the environment (Foresight 2007). The purpose of this review is to evaluate the effectiveness of interventions at the provider level, including both interventions targeted directly at health professionals and interventions targeted at the organisation of care.

OBJECTIVES

To assess the effectiveness of strategies to change either or both the behaviour of health professionals and the organisation of care to promote weight reduction in overweight and obese people.

METHODS

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs).

Types of participants

Fully qualified health professionals, working with overweight or obese adults.

Due to variability in the classification of overweight and obesity in primary studies, we included all trials enrolling adults described as overweight or obese. We used definitions based on body mass index (BMI - in kilogram/metre²): overweight was defined as a BMI over 25 but less than 30 and obesity as a BMI of 30 or over (EHCB 1997; NHLBI 1998).

We included studies of patients if a reduction in weight was specified as an objective of the intervention and outcome weight data were provided for the overweight or obese subpopulations within these patient groups. Thus, all patients in an included study had to be overweight or obese, or results from the overweight or obese subpopulation had to be provided separately.

Settings of studies

Healthcare organisations, defined as organisations that had health care as their primary objective. All patients in an included study had to be recruited in the context of a healthcare setting .

Types of interventions

We included any intervention that aimed to help a health professional implement an intervention targeting weight reduction in overweight or obese people. These interventions can be divided into two main categories according to the EPOC taxonomy (see Appendix 1) as one of the following.

i. Interventions targeting health professionals

Interventions aimed at improving the effectiveness of health professionals working to reduce the weight of overweight or obese people. This category includes strategies such as providing professionals with information or training on appropriate practice.

ii. Interventions targeting the organisation of care

Interventions aimed at changing the organisation of care directed at reducing the weight of overweight or obese people. This category includes interventions that were predominantly about changes in organisational systems, such as the introduction of multi-disciplinary teams, changes in skill mix, or in the setting of service delivery.

Comparators

We included only studies that had standard care as the comparator arm of the study.

We planned the following comparisons:

interventions targeting health professionals versus standard care;
interventions targeting the organisation of care versus standard organisation of care.

The standard care comparator groups had to meet either of these two criteria:

study participants receiving routine weight management service(s) in the context of their normal healthcare provision and setting, or:
study participants being informed of the availability of routine weight management service(s) in the context of their normal healthcare provision and setting.

Excluded studies

We excluded the following types of studies.

Studies that varied the clinical content or intensity of care, or both, of the intervention aimed at reducing weight, without a normal care control group. Therefore we excluded studies comparing the effectiveness of different durations of follow up, intervention, or frequency of consultation with obese or overweight people.
Studies that reported neither patients’ weight nor body mass index.
Studies that reported only knowledge or attitudes of health professionals or patient satisfaction, with no objective measure of professional performance or patient outcomes.

Types of outcome measures

We included any objective measure of provider performance consistent with EPOC guidelines (EPOC 2002) or patient outcomes. We also planned to report any available cost data.

Main outcomes

Patient’s body weight.

Other outcomes

Patient outcomes

body mass index (BMI); satisfaction with provider practice or healthcare provision; psychological outcomes (self-esteem, stress, depression, dietary restraint); morbidity (measures of disease status and sick leave); measures of body fat; effects on risk factors (differences in cholesterol levels, blood pressure); patient behaviour (attendance levels at weight management or physical exercise programmes); the number of withdrawals from treatment.

Health professional outcomes

measures of health practitioners’ behaviour, knowledge, attitudes, or satisfaction.

Search methods for identification of studies

For the first version of this review:

We utilised expertise used to develop search strategies for EPOC (EPOC 2009a) and the Effective Health Care Bulletin on obesity (EHCB 1997) to develop a search strategy for this review.

We searched the following databases : MEDLINE Ovid CD-ROM (1966 - 1/1998), PsycLit Silverplatter CD-ROM (1974 - 12/1997), EMBASE (Ovid via Bids) (1979 - 12/1997), Cinahl ARC Service (WinSPIRS online) (1982 - 11/1997), SIGLE Blaiseline (1980 - 11/1997), Sociofile ARC service (WinSPIRS online) (1974 - 10/1997), Dissertation Abstracts Dialog Corporation Dialog service (1861 - 1/1998), Conference Papers Index Dialog Corporation Dialog service (1973 - 1/1998), Resource Database in Continuing Medical Education (searched 6/1997). Copies of the full search strategies are available on request from the first author.

We also searched the following Cochrane Review Group specialised registers using ‘overweight’ and ‘obesity’ as the basis for key terms (EPOC 2009a): EPOC (5/1997), Cochrane Depression, Anxiety and Neurosis Group (8/1997), Cochrane Diabetes Group Register (8/1997). We also searched The Cochrane Controlled Trials Register (CCTR) (9/1997).

We undertook the following searches of key journals according to Cochrane criteria: International Journal of Obesity (1977 - 12/1997), European Journal of Clinical Nutrition (1988 - 12/1997), Journal of Human Nutrition and Dietetics (1988 - 12/1991), Human Nutrition: Clinical Nutrition (1982 - 12/1987), Human Nutrition: Applied Nutrition (1982 - 12/1987), Health Psychology (1993 - 12/1997), Obesity Research (1993 - 1994). We contacted experts in this field through the Association for the Study of Obesity (ASO), the British Dietetic Association (BDA), and the Journal of the American Dietetic Association (JADA) and asked them to notify us of potentially relevant papers. EPOC colleagues undertaking a review of preventive care identified further potentially relevant studies from their searches and the reference lists of included studies.
For the second version of this review:

For this update, we searched the following databases using the original search strategies: Medline Ovid CD-ROM Database (1/1997 - 4/2000), EMBASE Ovid via BIDS web (1/1998 - 2/2000), Cinahl Ovid CD-ROM Database (11/1997 - 2/2000), PsycLit Ovid Online (1/1997 - 5/2000), Sigle Blaiseline (1980 - 4/2000). We also searched the EPOC register and pending databases (4/2000). We did not search additional databases that were searched for the original production of the review again due to their low yield of relevant studies and the cost of accessing these databases.
For the third version of this review:

We updated the search for primary studies in the following databases, which were all interrogated from the previous (version 2) search date to May 2009:

The Cochrane Central Register of Controlled Trials (CENTRAL) (which at this time incorporated all EPOC Specialised Register material) (The Cochrane Library 2009, Issue 1), MEDLINE (Ovid), EMBASE (Ovid), CINAHL (EBSCO), PsycINFO (Ovid). We did not search SIGLE at this time as it had ceased being updated.

Search strategies for primary studies incorporate the methodological component of the EPOC search strategy combined with selected index terms and free text terms. We translated the MEDLINE search strategy into the other databases using the appropriate controlled vocabulary as applicable. We identified further potentially relevant studies from the reference lists of included studies.

We have included full search strategies for all databases in Appendix 2, Appendix 3, Appendix 4, and Appendix 5.

Data collection and analysis

Selection of studies

We searched for randomised controlled trials (RCTs) that compared routine provision of care with interventions aimed either at changing the behaviour of health professionals or the organisation of care to promote weight reduction in overweight and obese adults. We downloaded all titles and abstracts retrieved by electronic searching to the reference management database Endnote and removed duplicates. Two review authors (from KD, ME, GF, SK, TP, HA, JB) then examined the remaining references independently. We excluded those studies which clearly did not meet the inclusion criteria and obtained copies of the full text of remaining references. Two review authors (from KD, ME, GF, SK, TP, HA, JB) assessed the eligibility of these papers independently. Two authors (from KD, ME, GF, SK) resolved disagreements by discussion. We documented reasons for exclusion.

Data extraction and management

Two reviewers (from KD, ME, SK, TP, HA, JB, HD, FB) independently extracted data on study design, patient characteristics, interventions, outcomes, and risk of bias to a form specially designed for the review (Appendix 6). We noted the length of follow up for outcome measurement because short-term studies may be misleading, given that patients do not always maintain their initial weight losses (EHCB 1997).

Assessment of risk of bias in included studies

We used The Cochrane Collaboration’s tool for assessing risk of bias (Higgins 2008) on six standard criteria: adequate sequence generation, concealment of allocation, blinded or objective assessment of primary outcome(s), adequately addressed incomplete outcome data, free from selective reporting, and free of other risk of bias. We used three additional criteria specified by EPOC (EPOC 2009b): similar baseline characteristics, reliable primary outcome measures, and adequate protection against contamination.

Measures of treatment effect

Where possible, we extracted the mean weight of participants in each arm at the end of the study and the standard deviation of this mean. We calculated the difference between the intervention and control arms in the final mean weight of participants, and its standard error, to summarise the effect of treatment. Some studies presented the overall treatment effect and its standard error, but not the final weight in each arm and, in these cases, we used these reported treatment effects directly.

If final mean weights in each arm and their SDs (or the difference in final mean weight and its SE) were not reported, we extracted the mean change in weight between baseline and the end of the study in each arm, and its standard deviation; hence we calculated the difference between the intervention and control arms in the mean change in weight of participants, and its standard error, to summarise the effect of treatment.

For one study (Rogers 1982), final weight was not reported but the average amount that patients were overweight was reported by treatment arm at the end of the study, so we used this as the primary outcome.

If results were presented at more than one time point, we used the results for the longest duration of follow up in our primary meta-analysis.

Unit of analysis issues

We noted whether studies randomised patients or healthcare providers (either GPs or GP practices). If analysis did not allow for clustering of patients within healthcare providers, we recorded a unit of analysis error, as such analysis tends to over-estimate the precision of the effect of treatment ( Goldstein 2003).

Dealing with missing data

If primary outcome data were missing, or only imputed data were reported, we contacted trial authors to request data on the outcomes among participants who were assessed.

Assessment of heterogeneity

We assessed heterogeneity between studies by visual inspection of forest plots, by estimation of the percentage of heterogeneity between trials which could not be ascribed to sampling variation (Higgins 2003), by a formal statistical test of the significance of the heterogeneity (Deeks 2001) and, when possible, by sub-group analyses (see below). If there was evidence of substantial heterogeneity, we investigated and reported the possible reasons for this.

Assessment of reporting biases

We had planned to examine funnel plots corresponding to meta-analysis of the primary outcome in order to assess the potential for small study effects such as publication bias. However, as we only found six included studies, three of which had data suitable for meta-analysis, we did not examine funnel plots.

Data synthesis

We have summarised the outcome data extracted from papers in Analysis 1.1.

The mean differences between the patient’s weight (or change in weight) in the intervention and standard care arms at the end of the each trial are presented in separate forest plots for educational interventions (Cohen 1991; Martin 2006; Moore 2003), reminders (Rogers 1982) and organisational interventions (Pritchard 1999; Sherwood 2006). For both educational interventions and organisational interventions, we used the generic inverse variance facility of Review Mananager 5 (RevMan 5) (RevMan 2008), because the trials of Moore 2003 and Pritchard 1999 reported the final weight (or change in weight) in the intervention arms relative to the standard care arm, rather than the final weight (or change in weight) in both intervention and standard care arms.

For the three trials that considered educational interventions, we pooled results in a meta-analysis using the mean difference method (Deeks 2008). We used a random-effects model with inverse variance weighting (DerSimonian 1986).

Subgroup analysis and investigation of heterogeneity

We performed sub-group analyses and grouped trials based on whether they reported final values of patients’ weight or the change from baseline. We considered factors such as the type of intervention, whether it was evidence-based, and the length of follow up in interpretation of any heterogeneity.

We had also planned to sub-group analyses by overweight and obesity, as they may hold different implications for health and treatment, but this was not possible because the included papers did not distinguish between such patients.

Sensitivity analysis

We examined results after one year’s follow up (or as close as possible to one year) in a sensitivity analysis.

RESULTS

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies.

Results of the search

First version of review: from initial searches of electronic databases and requests for help, EH screened 7193 abstracts. Two authors (EH and AMG) independently assessed and cross-checked any studies that appeared relevant (244) . Two authors (EH and CS or SK) independently assessed 107 full text copies of papers; 95 studies failed to meet the inclusion criteria. We included 12 studies in the review.

Second version of the review: from the searches of electronic databases, EH screened 1612 abstracts. This elicited 33 abstracts for dual screening (potentially relevant studies/number of hits): MEDLINE 4/390, EMBASE 1/422, PsycLit 0/141, Sigle 0/71, CINAHL 0/521, EPOC register 13/21, EPOC pending database 15/46. Two authors (EH and CS) independently assessed and cross-checked these abstracts . In addition, we identified 18 studies through other means (direct notification to the authors, searches for other Cochrane reviews). From these, we retrieved 43 full paper copies which two authors (EH and CS) independently assessed for inclusion criteria.. Of these 43, we agreed to include seven in the review update. However, one of these papers (Haring 1976) was an additional report of an included paper (Rogers 1982), therefore we included only six new studies in the second version of this review. Tne authors resolved all discrepancies easily by discussion. Third version of review: our searches found 12,097 references. We excluded 11,698 of these from a reading of their title and abstract (where available) as they did not meet our inclusion criteria. We retrieved 399 references in full and two authors (from KD, ME, SK, TP, HA, JB) independently assessed these for inclusion. Of these papers, 367 clearly did not fulfil the inclusion criteria. Of the remaining 32 studies, 18 were included in the previous version of the review. Since some of the inclusion and exclusion criteria had changed for the current update (principally study design, routine comparator, and requirement that study be conducted in a healthcare setting), a number of studies previously included were now excluded (Atkinson 1977; Balch 1976; Counterweight Prog 2004; Ferstl 1975; Hagen 1974; Hakala 1994; Jeffery 1979; Jeffery 1982; Levitz 1974, Lindstrom 1976; McDonald 1984; Meyers 1996; Ogden 1997; Perri 1987; Richman 1996; Simkin-Silverman 1997). We excluded eight further studies after detailed inspection; we have reported the reasons for exclusion of these 26 studies in the Characteristics of excluded studies table. We identified six papers that met the inclusion criteria of this review, three of which (Cohen 1991; Pritchard 1999; Rogers 1982) were included in the previous version of the review.

Included studies

See: Characteristics of included studies and Table 1.

Table 1. Summary of methods and participants of included studies.

Study ID	Methods	Participants	Patients
Cohen 1991	Design: CRCT Unit of allocation: Provider Unit of analysis error: Yes Power calculation: No	Providers: 18 family physicians at one family health center	Patients: 30 overweight or obese individuals (BMI > 27.3 or 27.8 for males and females respectively) Age: I: 59.3 and C: 59.7 years Gender: 73% females Ethnicity:unclear Diabetes :unclear IHD: all patients were hypertensive
Martin 2006	Design: CRCT Unit of allocation: Provider Unit of analysis error: No Power calculation: Yes	Providers:8 physicians from two clinics	Patients: 144 overweight or obese patients (BMI> 25) Age:I:40.7 (12.6), C:43.0 (11.4) years Gender: 100% female Ethnicity: 100% african-american Diabetes and IHD:unclear
Moore 2003	Design: CRCT Unit of allocation: Practice Unit of analysis error: No Power calculation: Yes	Providers: 245 healthcare staff from 44 practices	Patients: 843 obese patients (BMI>30) Age:I:48.8 (10.9), C:48.8 (12.2) years Gender: I: 75% and C: 73% female Ethnicity: unclear Diabetes and IHD:unclear
Pritchard 1999	Design: RCT Unit of allocation: Patients Unit of analysis error: No Power calculation: Yes	Providers: one dietitian and unclear number of GPs at one practice	Patients: 270 overweight or obese patients (BMI≥25) Age:Unclear Gender: 72% females Ethnicity: unclear Diabetes: unclear IHD:Unclear
Rogers 1982	Design: RCT Unit of allocation: Unclear Unit of analysis error: Unclear Power calculation: No	Providers:Unclear number of physicians at cardiac, pulmonary and renal clinics	Patients: 147 obese patients (whose weight exceeded 20% of their ideal weight) Age:unclear Gender: 77% females Ethnicity: unclear Diabetes: 33.3% IHD:Unclear
Sherwood 2006	Design: RCT Unit of allocation: Patient Unit of analysis error:No Power calculation: Yes	Providers: Unclear number of trained nutritionists and/or exercise specialists	Patients: 1801 overweight or obese patients (23/77) Age:50.7 yrs Gender: 72% females Ethnicity: 91% Caucasian Diabetes 5.5% IHD:Unclear (27.3% on IHD related medication)

Study	Reason for exclusion
Ashley 2001	Lack of standard care arm. Participants not recruited in the context of a healthcare setting
Atkinson 1977	No standard care arm. The intervention was not led solely by qualified healthcare professionals
Balch 1976	No standard care arm. Recruitment of overweight participants did not occur in a healthcare setting
Ben Noun 1988	The comparator was not normal care.
Boltri 2007	Not all of the patients were overweight or obese (Table 1). It did not report weight or weight loss at the end of sudy (i.e. no objective outcome measure was reported)
Counterweight Prog 2004	No outcome data for the control arm available at present. (Personal communication with the authors.)
De Mello 2004	Study used participants who were children. The participants were not recruited in the context of a healthcare setting, nor was the intervention led by a qualified healthcare professional
Donnelly 2007	No standard care arm. Unclear if subjects were recruited in the context of a healthcare setting. There was also concern that the “experienced health educators” used in this study were not qualified health professionals
Dunstan 2006	No standard care arm. Not all participants were recruited in the context of a healthcare setting. Overweight and obese participants were not supervised by health professionals in the gym (but by YMCA staff)
Ferstl 1975	No standard care arm, not all participants were recruited in the context of a healthcare setting, nor did the intervention take place in healthcare setting; the whole intervention was devised for this study and delivered at a non-profit institute for therapy research
Finnish DPS Group 1999	The study participants were not recruited in the context of a healthcare setting. “The study subjects were recruited through various methods, e. g. from epidemiological surveys and by opportunistic population screenings with special emphasis on the high-risk groups such as obese subjects and first-degree relatives of Type II diabetic patients. Subjects were also recruited through advertising in local newspapers.” (Page 794/ col 2 / para 5 of Eriksson 1999)
Hagen 1974	Recruitment of overweight participants did not occur in a healthcare setting. Not all participants were over 18
Hakala 1994	RCT organisation: inpatient vs outpatient. Too much variation in content between the two groups.
Jeffery 1979	No standard care arm. It only compared the frequency of therapist contact. The intervention has been designed for the study at Stanford University and it’s unclear if any sort of care program was in place for overweight undergraduates at the University
Jeffery 1982	No standard care arm. Participants were not recruited in the context of a healthcare setting. The intervention was not led by qualified healthcare professionals
Kromann 1985	This study appears to be a CBA with a convenience sample of patients. Also the standard care arm physicians received additional training and is described by the authors as not representative of other GP practices
Levitz 1974	Participants were not recruited in the context of a healthcare setting. Weight loss intervention led by non-health professionals
Lindstrom 1976	Patients not recruited in the context of a healthcare setting. The intervention was not led by a qualified healthcare professional
McDonald 1984	No objective patient outcome data reported i.e. it did not report patient weight or weight loss at the end of the study
Meyers 1996	No standard care arm, the face to face group is not “standard care” and also the intervention was designed solely for the study. The participants were not recruited in the context of a healthcare setting
Ogden 1997	No objective outcome measures i.e. it did not report weight or weight loss at the end of study
Perri 1987	Patients not recruited in context of a healthcare setting. No standard care arm
Richman 1996	Not an RCT (controlled before and after study).
Simkin-Silverman 1997	No objective outcome measures i.e. it did not report weight or weight change at the end of study
Vinicor 1987	Not all of the participants were overweight or obese (Table 1, pg 350)
Willaing 2004	Not all of the participants were obese or overweight.

Outcome or subgroup title		Effect size
1 Results	Other data	No numeric data
1.1 Professional interventions vs. standard care	Other data	No numeric data
1.2 Organisational interventions vs. standard care	Other data	No numeric data

Outcome or subgroup title	No. of studies	Statistical method	Effect size
1 Weight (kg) at longest follow up	3	Mean Difference (Random, 95% CI)	−1.24 [−2.84, 0.37]
1.1 Change in weight between baseline and end of study	2	Mean Difference (Random, 95% CI)	−1.77 [−2.80, −0.74]
1.2 Weight at end of study	1	Mean Difference (Random, 95% CI)	1.3 [−1.86, 4.46]
2 Weight (kg) at 1 year follow up (or closest timepoint available)	3	Mean Difference (Random, 95% CI)	−1.29 [−2.77, 0.20]
2.1 Change in weight between baseline and end of study	2	Mean Difference (Random, 95% CI)	−1.77 [−2.80, −0.74]
2.2 Weight at end of study	1	Mean Difference (Random, 95% CI)	1.0 [−1.96, 3.96]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Weight (kg) at longest follow up	1	108	Mean Difference (IV, Fixed, 95% CI)	−3.75 [−8.46, 0.96]
1.1 Amount overweight at end of study (Men)	1	20	Mean Difference (IV, Fixed, 95% CI)	−11.2 [−20.66, −1.74]
1.2 Amount overweight at end of study (Women)	1	88	Mean Difference (IV, Fixed, 95% CI)	−1.30 [−6.73, 4.13]

Outcome or subgroup title	No. of studies	Statistical method	Effect size
1 Weight (kg) at longest follow up	2	Mean Difference (Random, 95% CI)	Totals not selected
1.1 Doctor/dietitian vs. standard care	1	Mean Difference (Random, 95% CI)	Not estimable
1.2 Dietician vs. standard care	1	Mean Difference (Random, 95% CI)	Not estimable
1.3 Mail intervention vs. standard care	1	Mean Difference (Random, 95% CI)	Not estimable
1.4 Telephone intervention vs. standard care	1	Mean Difference (Random, 95% CI)	Not estimable
2 Weight (kg) at 1 yr follow up (or closest timepoint available)	2	Mean Difference (Random, 95% CI)	Totals not selected
2.1 Doctor/dietitian vs. standard care	1	Mean Difference (Random, 95% CI)	Not estimable
2.2 Dietician vs. standard care	1	Mean Difference (Random, 95% CI)	Not estimable
2.3 Mail intervention vs. standard care	1	Mean Difference (Random, 95% CI)	Not estimable
2.4 Telephone intervention vs. standard care	1	Mean Difference (Random, 95% CI)	Not estimable

Study	Comparisons	Main process effect	Main patient outcome
Professional interventions vs. standard care
Cohen 1991	I: One educational teaching session C: Usual care.	Not done, except for recording how many patients the 18 GPs I: N=10; C: N=8) recruited each	See Table 2. Weight change from baseline (SD) (kg) 0-6 months I: (n=15) −1.8 (3.4) C: (n=15) 0.56 (2.5) I-C = −2.36 (favours I) 6-12 months I: (n=15) 0.94 (3.3) C: (n=15) 0.73 (2.2) I-C = 0.21 (favours C) 0-12 months I: (n=15) −0.88 (4.0) C: (n=15) 1.3 (3.0) I-C = −2.18 (favours I)
Martin 2006	I: GP targeted intervention: remote educational teaching session and interventions tailored to the character of the overweight and obese patients by a multidisciplinary team delivered over 6 months C: Standard care	No post intervention assessment of GP practice.	See Table 2 (ITT analysis) and page 1417/Col. 1/end of para 2 Weight change from baseline (SD) (kg) 0-6 months I: (n=69) −1.44 (3.30) C: (n=69) 0.25 (3.30) I-C: −1.69 (favours I)
Moore 2003	I: Three 90 minutes educational sessions over 4 weeks targeted at GPs and their teams C: Usual care	Values are numbers responding “yes” at 12 months Evidence that weight discussed in consultation. (n = 650) I: 186 C: 129 Odds ratio 2.0 (95%CI:1.3 to 3.2) P=0.003 Weight recorded. (n=650) I: 197 C: 137 Odds ratio 2.0 (95% CI:1.3 to 3.3) P = 0.004 Target weight recorded (n=643) I: 46 C: 9 Odds ratio 13.6 (95% CI: 4.2 to 44. 3) P <0.001 Dietary targets recorded (n=648) I: 48 C: 14 Odds ratio 4.5 (95% CI:1.2 to 16.7) P=0.02 Exercise targets recorded (n=648) I: 46 C: 25 Odds ratio 1.9 (95%:0.7 to 5.0) P=0.2	See Figure 2 and Table 2. Difference in weight, I- C (SE) 3 months I: n=331 C: n=333 I-C: 0.6(1.38) 12 months I: n=279 C: n=286 I-C: 1.0(1.51) (Favours C) 18 months I: n=256 C: n=275 I-C: 1.3(1.61) (Favours C) Difference in BMI (kg/m²): I- C (SE) 3 months I-C:-0.2 (0.52) (Favours I) 12 months 0 (0.52) 18 months I-C: 0.1 (0.55) (Favours C)
Rogers 1982	I:Computerised reminders C: Usual care	Number of diets given or reviewed: Reminders vs control: Year 1: 2 (4.8%) Year 2: 4 (9.1%) Done both years: 7 (13.5%) Not done: 24 (27.5%) p = 0.007 ’for all obese patients combined for sex’ (but not clear which of the above figures this is for). (No SD/SEs so not possible to calculate CIs.)	See Table 5, page 71. Follow-up: 147 patients classified as obese, 23 dropped out, but data collected for: Mean kg. overweight : 10-15 months Men: I: (n = 15) 20.4 (7.5) C: (n = 11) 25.7 (12.6) I-C= −5.3, SE=4.35 (favours I) Women: I: (n = 42) 23.4 (13.3) C: (n = 46) 24.8 (11.5) I-C= −1.4, SE=2.76 (favours I) 22-24 months Men: I: (n = 11) 15.8 (6.6) C: (n = 9) 27.0 (13.2) (favours I) Women: I: (n = 42) 23.6 (14.7) C: (n = 46) 24.9 (10.8) (favours I)
Organisational interventions vs. standard care
Pritchard 1999	I1: Doctor/ dietitian I2: Dietician C: Usual care	None measured	Page 314: sections on “Weight outcomes”, “Blood pressure outcomes”; n from Table 2 I1: n= 92 I2: n=88 C: n = 90 Weight change relative to control 12 months I1-C: −6.7 (0.42) I2-C: −5.6 (0.39) I1 - I2: −1.1 (0.92) Change in blood pressure relative to control (mmHg) 12 months I1-C: −12 (1.56) I2-C: −7 (1.56) I1 - I2: −5 (1.56) Total cost per group: I1-C: (n=93) $8240.30 I2-C: (n=89) $5715.06 C: (n=91) $2103.53 Additional cost per kg lost: I1-C: (n=93) $9.76 I2-C: (n=89) $7.30
Sherwood 2006	I1: Mail delivered intervention I2: Phone delivered intervention C: Usual care	None measured	See Table 2 and Table 5. Weight change (kg), mean (SD): 18 months I1: (n=600) −2.27 (5.9) I2: (n=601) −2.35 (5.9) C: (n=600) −1.91 (5.9) 24 months I1: (n=600) −0.73 (5.4) I2: (n=601) −0.93 (5.4) C: (n=600) −0.59 (5.4) I1-C: −0.14 I2-C: −0.34 (Favours I1, I2) Total costs/participant: I1: (n=600) $50.45 I2: (n=601) $127.39 C: (n=600) $42.18 Cost/weight loss of 1kg: I1: (n=600) $72.08 I2: (n=601) $132.70 C: (n=600) $71.50

	1.1 Reviews scope: Any intervention that aims to improve the way health professionals work to reduce the weight of overweight or obese people. That is the effect(s) of a behavioural/ educational, financial, organisational or regulatory intervention (s) is evaluated	RELEVANT SUPPORTING TEXT AND LOCATION (page/column/paragraph)
YES	The effect of intervention(s) that aims to improve the way health professionals work to reduce the weight of overweight or obese people is evaluated. NB the population must be overweight or obese OR the overweight or obese population’s results are segregated for at least one of our significant outcomes (weight loss or objective measure of health professional’s behaviour change)
NO
UNCLEAR	The intervention does not appear to be clearly described. Discuss the paper with KD before beginning data extraction

	1.2 Randomised controlled trial (RCT)	RELEVANT SUPPORTING TEXT AND LOCATION (page/column/paragraph)
YES	Statement of random allocation of health professionals, patients, episodes of care, locations of care, etc given by authors
NO	No statement of random allocation of health professionals, patients, episodes of care, locations of care, etc
UNCLEAR	Discuss the paper with KD before beginning data extraction

	1.3.1 Paper reports objective measurement of provider performance/behaviour or patient outcome(s)	RELEVANT SUPPORTING TEXT AND LOCATION (page/column/paragraph)
YES	E.g. Primary outcome : Patient weight loss, OR Secondary Patient outcomes: psychological outcomes (depression, dietary restraint); morbidity (measures of disease status, sick leave); fat or BMI measures; effects on risk factors (differences in cholesterol levels, blood pressure); patient behaviour (attendance levels at weight management or physical exercise programmes); and number of withdrawals from treatment OR Secondary Health professional outcomes: measures of health practitioners behaviour, knowledge.
NO	E.g. self-report data, measures of attitudes or beliefs or perceptions or satisfaction. Studies reporting only knowledge or attitudes of health professionals or patient satisfaction with no objective measure of professional performance or patient outcomes are to be excluded
UNCLEAR	Discuss the paper with KD before beginning data extraction

	1.3.2 Relevant and interpretable data presented or obtainable (e.g. by reading points off a graph)
YES	Data is presented or obtainable
NO	Relevant data is not presented and is clearly unobtainable
UNCLEAR	Discuss the paper with KD before beginning data extraction

Methods	Design: CRCT Unit of allocation: Provider: Family practice residents were randomly assigned to either an experimental or a control group (Pg 25/ Col 1/ Para 1) Unit of analysis: Patient: (Table 2/ Pg 27). But no attempt was made to account for the clustering effect: Analysis of covariance was used to compare weight change and blood pressure change between the experimental and control groups and to compare blood pressure change between the weight losers and weight gainers, adjusting for initial values. The unpaired T test (two-tailed) was used to compare the experimental and control groups and the weight gainers and losers with respect to baseline age, weight, BMI, mean arterial pressure, number of medications, and number of visits to the physician. The Mann-Whitney U test was used to compare the experimental and control groups and the weight losers and gainers with regard to change in the number of medications. The Spearman rank correlation coefficient was used to assess the correlation between change in number of medications and change in blood pressure. (Pg 26 /Col 1/ para 7) Power calculation: No power calculation
Participants	The number randomised into the trial: *Episodes of care:* Number of visits to family practitioner: 9.7 (SD 3.0) intervention group; 5.2 (SD 2.4) control group, (Table 2/pg 27) *Patients:* 30 (Pg 26/ Col 2/ Para 1) (31 patients originally randomised but data for one patient who was excluded due to another health problem is not presented) *Providers:* 18 (Pg 26/ Col 2/ Para 1) Int: 10; Control: 8 *Practices:* 1 (The Lawrenceville Family Health Center, Pg 25/ Col 2/ Para2) *Hospitals:* N/A *Communities or regions:* N/A Characteristics of participating healthcare providers: Profession: Physicians: 18 family practice physicians (Pg 26/ Col 2/ Para1) Level of training: In post-graduate training: Residents (Pg 25/ Col 2/ Para 2) Age of health professional: Unclear Years since graduation or in practice: Unclear Proportion of eligible providers (or allocation units) who participated in the evaluation: Unclear Characteristics of the participating patients: Clinical problem(s) of participating patients: Overweight (BMI ≥ 25 but ≤ 30): Unclear: To be included participants had to be obese defined by a body mass of 27.8 or more in males and 27.3 in females. (Pg 25 /Col 2/ Para 3). Obese (BMI ≥ 30): Unclear: To be included participants had to be obese defined by a body mass of 27.8 or more in males and 27.3 in females. (Pg 25/ Col 2/ Para 3). Mean BMI 34.2 (int) and 34.0 (controls) (Table 1/ pg 26) But no distinction between overweight and obese populations provided. *Diabetes:* Unclear: *Ischemic heart disease:* All patients were hypertensive (systolic blood pressure > 139 or diastolic blood pressure > 89) (Pg 25/ col 2/ para 3) Other characteristics of participating patients: Age: Mean: 59.3 years (intervention group) 59.7 years (control group) (Table 1/pg 26) Baseline Weight /BMI Intervention: 91.8 kg /34.2 Control: 91.7 kg /34.0 *Gender:* 22 females, 8 males, equally distributed between the two groups (Pg 26/ Col 2/ Para 1) Ethnicity: Unclear *Other:* All hypertensive. Diagnosis of hypertension based on an average systolic BP of 140mm Hg or more on 2 or more readings or an average diastolic BP of 90mm Hg on 2 or more readings recorded in the FHC record. (Pg 25/ Col 2/ Para 3) Mean arterial pressure 105.6 (intervention group), 105.9 (control group) (Table 1/pg 26) Setting: Reimbursement system: Unclear Setting of care: General practice or community-based: The Lawrenceville Family Health Center, (Pg 25/ Col 2/ Para 2) Academic status of the setting of care: University (teaching) hospital: The University of Pittsburgh, St Margaret Memorial Hospital (Pg 25/ Col 1/ Para 3). The Lawrenceville Family Health Centre is the model family practice unit for the family practice residents at St Margaret Memorial Hospital. (Pg 25/ Col 2/ Para 2) Country: USA: Pittsburgh, PA (Pg 25/ Col 1/ Para 3)
Interventions	Professional intervention: Intervention group: At a residents physicians’ meeting all residents were informed of the broad principles of the trial; details that would influence the status of experimental or control groups were excluded. Physicians assigned to the experimental group were taught about the importance of weight reduction in managing hypertension and were provided with information about the effects of specific foods on body weight. The teaching session was conducted by a behavioural psychologist who has special interest and expertise in weight reduction. During the teaching session the physicians were questioned about their knowledge of the caloric content of foods and were given practical strategies for changing the dietary habits of their patients. The goal of the dietary advice was to reduce the caloric content of the diet without radically changing the patient’s life style. Methods of encouraging patients, such as reinforcement, were also discussed. The residents were given an instruction sheet that included low-calorie alternatives to high calorie foods. Other key strategies included seeing patients monthly and reviewing the previous day’s food intake with the patient. (Pg 25/ Col 2/ Para 4 - Pg 26/ Col 1/ Para 1) Control group: At a residents physicians meeting all residents were informed of the broad principles of the trial; details that would influence the status of experimental or control groups were excluded. The physicians in the control group received no special instructions or materials (Pg 25/ Col 2/ Para 4) Timing of intervention: *Proximity to clinical decision-making:* Remote educational sessions: A single training session was provided at the start of the trial (Pg 25/ Col 2/ Para 4 - Pg 26/ Col 1/ Para 1) *Frequency/number of intervention events:* A single training session was provided at the start of the trial (Pg 25/ Col 2/ Para 4 - Pg 26/ Col 1/ Para 1) *Duration of intervention:* A single training session of unknown duration was provided at the start of the trial (Pg 25/ Col 2/ Para 4 - Pg 26/ Col 1/ Para 1) Healthcare professional recipient: *Intervention group:* A single training session was provided to the experimental group physicians at the start of the trial (Pg 25/ Col 2/ Para 4 - Pg 26/ Col 1/ Para 1) *Control group:* The physicians in the control group received no special instructions or materials (Pg 25/ Col 2/ Para 4) Intervention deliverer: *Intervention group:* The teaching session was conducted by a behavioural psychologist who has special interest and expertise in weight reduction (Pg 25/ Col 2/ Para 5) *Control group:* N/A Types of targeted behaviour of the health professionals: To use the practical strategies taught to change their patient’s dietary habits (Pg 25/ Col 2/ Para 4 - Pg 26/ Col 1/ Para 1) Development of the intervention: Consultation with professional recipients: Unclear Evidence base of intervention: Unclear Consumer involvement: Not specified Barriers to change: Not done Source of funding for study: This study was conducted as part of Dr Cohen’s fellowship at St. Margaret Memorial Hospital (Pg 25/ Col 1/ Para 3) but the source of the funds wasn’t stated Ethical approval: Not reported
Outcomes	Outcomes measured: Weight change Blood pressure change (Change in mean arterial pressure change in mm Hg) Change in the number of medications Number of visits to the physician (Tables 2 - 3/ Pg 27) Length of time outcomes measured after initiation of the intervention: All outcomes measured at baseline, 6 months and 12 months (Table 1/ Pg 27) Ceiling effect: *Identified by investigator:* Unclear *Identified by reviewer:* No, potential for weight loss in population clear and demonstrated. (Table 1/ Pg 27). However, it was not clear to what extent physicians were already doing the intervention behaviours Losses to follow up: Number randomised: *Intervention group:* 15 (Pg 26/ Col 2/ Para 1) *Control group:* 15 (Pg 26/ Col 2/ Para 1) Number completing follow up: *Intervention group:* 15. Over the entire 12-month period of study there were no dropouts from the experimental group (Pg 26/ Col 2/ Para 2) *Control group:* 15. Over the entire 12-month period of study there were no drop-outs from the control group (Pg 26/ Col 2/ Para 2) Reasons for loss to follow up: *Intervention group:* N/A *Control group:* N/A Economic variables: Costs of the intervention: Not reported Changes in direct healthcare costs as a result of the intervention: Not reported Changes in non-healthcare costs as a result of the intervention: Not reported Costs associated with the intervention linked with provider or patient outcomes in an economic evaluation: Not reported
Notes	Unit of analysis error: Results were analysed without allowing for clustering of patients within physicians (page 26/Col1/Bottom para)
*Risk of bias*
Item	Authors’ judgement	Description
Adequate sequence generation?	Unclear	The residents were stratified by residency year and randomly assigned to either control or experimental groups (Pg 25/ Col 2/ Para 4)
Allocation concealment?	Unclear	The residents were stratified by residency year and randomly assigned to either control or experimental groups (Pg 25/ Col 2/ Para 4)
Blinding? Primary outcome	No	Weight change: At each visit the patients weight was recorded and any weight change noted (Pg 26/ Para 1/ Col 4)
Blinding? Secondary outcomes	Unclear	Mean arterial pressure change: At baseline the blood pressure was measured by a nurse who had been trained in accordance with recommendations of the American Heart Association (Pg 26/ Col 1/ Para 2) But no information was presented on later measurements. Number of visits: No information presented
Blinding? Secondary outcomes	No	Change in number of antihypertensive medications: Management of the patient’s hypertension medication was left to the resident (Pg 26/ Para 1/ Col 4)
Incomplete outcome data addressed? Primary outcome	Yes	Weight change: Over the entire 12 month period of the study there were no drop-outs from either experimental or control groups (Pg 26/ Col 2/ Para 2)
Incomplete outcome data addressed? Secondary outcomes	Yes	Mean arterial pressure change / Number of visits / Change in number of antihypertensive medications: Over the entire 12 month period of the study there were no drop-outs from either experimental or control groups (Pg 26/ Col 2/ Para 2)
Free of selective reporting?	Yes	The authors state they intend to measure changes in weight, arterial blood pressure, number of antihypertensive agents prescribed, number of visits, variables which are all presented in the paper (Table 1 and Table 3/ Pg 27)
Baseline characteristics similar?	Yes	Table 1/pg26
Reliable primary outcome measures? Average weight change	Yes	Weight change: At baseline the patient’s weight was measured by a nurse who had been trained in accordance with recommendations of the American Heart Association (Pg 26/ Col 1/ Para 2). At 6 and 12 months the patient’s weight were noted by the same trained nurse (Pg 26/ Col 1/ Para 6)
Protection against contamination?	Yes	The physicians in the control group received no special instructions or materials. Physicians in the experimental group were asked not to share information from the educational sessions or special materials with control physicians. (Pg 25/ Col 2/ Para 4) There was no evidence of contamination between the experimental and control groups during the 12 months of the study. Chart audit revealed no use of the educational materials by control residents and interviews with them disclosed no awareness of information from the teaching session. (Pg 26/ Col 2/ Para 1)

Methods	Design: CRCT Unit of allocation: Provider: Clinician: Individualised (stratified per clinic, balanced, nested design) (Pg 1413/ Col 2/ Para 3) Unit of analysis: The analysis of the primary response variable, weight change at 6 months, was effected with a mixed linear model that included treatment group (two levels) and clinic (four physician practices were recruited from each of two clinics) as fixed effects in a factorial arrangement. An additional random effect was introduced to account for sampling variability among physician practices and to provide the appropriate test statistic for the treatment effect, due to the nesting of subjects within practice. (Pg 1415/ Col 2/ Para 2) Power calculation: The sample size of 20 participants per physician was chosen based on a power analysis indicating that 16 subjects per physician (128 patients total) would give 80% power to detect a difference of 23% in success proportion under a one-tailed hypothesis. The final target sample size of 20 subjects per physician was judged adequate to allow for attrition and other sources of exclusion. The power analysis was conducted using a binomial model for a proportion of success of 5 lb or 2.27 kg within a physician practice in achieving weight loss. (Pg 1414/ Col 1/ Para 2)
Participants	The number randomised into the trial: *Episodes of care:* Tailored intervention group received six monthly active treatment visits during which their physician delivered the intervention. Each visit lasted ~15 minutes. (Pg 1414/ Col 2/ Para 3) Unclear for standard care participants - received no special instructions and were seen, as needed, for regular medical care. (Pg 1415/ Col 1/ Para 2) *Patients:* 144 (Fig 1 / Pg 1416) *Providers:* 8 (Pg 1413 / Col 2/ Para 3) *Practices:* 2 (Pg 1413 / Col 2/ Para 3) *Hospitals:* N/A *Communities or regions:* N/A Characteristics of participating healthcare providers: Profession: 8 physicians: 4 clinicians in each group, from 2 clinics (Pg 1413/ Col 2/ Para 3) Level of training: Fully trained (presumed rather than stated) (Pg 1413/ Col 2/ Para 3) Age of health professional: Unclear Years since graduation or in practice: Unclear Proportion of eligible providers (or allocation units) who participated in the evaluation: Unclear Characteristics of the participating patients: Clinical problem(s) of participating patients: pg refs needed Overweight (BMI ≥ 25 but ≤ 30): Not stated, all patients BMI ≥ 25 (Pg 1413/ Col 2/ Para 2) Obese (BMI ≤ 30): Not stated, all patients BMI ≥ 25 (Pg 1413/ Col 2/ Para 2) *Diabetes:* Unclear *Ischemic heart disease:* Unclear Other characteristics of participating patients: *Age:* Intervention group: Mean 40.69 years, SD 12.59 (n = 73). Control group: Mean 42.97 years, SD 11.38 (n = 71). (Table 1/ Pg 1415) Range: Inclusion criteria 18 to 65 years (Pg 1413/ Col 2/ Para 2) Baseline Weight kg(SD) Intervention (n = 71): 103.0 (17.95) Control (n = 73): 100.86 (20.8) *Gender:* 100% female (Pg 1413/ Col 2/ Para 2) *Ethnicity:* 100% African American (Pg 1413/ Col 2/ Para 2) *Other:* All low income. All with no serious or uncontrolled medical condition (Pg 1413/ Col 2/ Para 2) Setting: Reimbursement system: Fee for service: each physician received a $35.00 reimbursement for each office visit, which was the amount reimbursable under state Medicaid rules for similar office visits.” (Pg 1414/ Col 1/ Para 3) Setting of care: General practice-based (Pg 1413/ Col 2/ Para 2) Academic status of the setting of care: Non-teaching or university affiliated (Pg 1413/ Col 2/ Para 2) Unclear Country: USA, Baton Rouge, LA (Pg 1413/ Col 2/ Para 2)
Interventions	Professional intervention: Intervention group: Physicians from both groups initially received 2 hours of instruction on general obesity treatment, as outlined by the National Heart, Lung, and Blood Institute clinical practice guideline on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults (36). The four physicians providing tailored interventions then received an additional 7 hours of training, which addressed the assessment of stage of change, motivational interviewing, and techniques for the behavioral treatment of obesity. This training also included instruction on appropriate dietary recommendations, such as ways to reduce dietary fat intake, appropriate fruit and vegetable intake, how to read food labels, and how to modify recipes. Tailored Intervention Patients in the tailored intervention group received six monthly active treatment visits during which their physician delivered the intervention. Each visit lasted 15 minutes. Physicians received protocols for each monthly visit, and participants received both oral recommendations from their physician and handouts summarising the focus of each visit. The treatment materials delivered by the physician were individually prepared and tailored to each patient by a multidisciplinary research team consisting of the physician, a health psychologist, a registered dietitian, and an exercise physiologist. Physicians provided feedback and input to the multidisciplinary team, although the actual materials were written and prepared by the other research team members. Physicians had the option of either delivering the treatment using a prepared script or delivering the intervention with the assistance of an outline of main points to be covered. The content of the tailored interventions was obtained from the information provided by participants during the baseline assessment visit. Based on current eating practices and preferences, a dietitian provided recommendations to assist each participant in making healthier food choices and provided meal preparation tips. The exercise physiologist provided tailored physical activity recommendations based on the participant’s current activity levels, activity preferences, and any barriers to activity reported (e.g., medical conditions, lack of social support, unsafe neighbourhoods). A health psychologist developed tailored behavioral change recommendations based on Social Cognitive Theory, the Transtheoretical Model, and behavioral principles that targeted constructs such as self-efficacy, motivational readiness to change, social support, pros/cons of behavior change, self-reinforcement, realistic goal setting, stimulus control, and contingency management. The recommendations written by each expert were incorporated into the tailored intervention materials that were presented by the physician to the patient. In addition, the recommendations were tailored to the cultural and socioeconomic status backgrounds of the participants by taking cultural preferences into account when formulating dietary and exercise plans, providing educational materials prepared specifically for African Americans, and giving low-cost alternatives when making diet and physical activity recommendations. Topics of the monthly meetings included introductory information on weight loss, ways to decrease dietary fat, ways to increase physical activity, dealing with barriers to weight loss, healthy alternatives when eating out and shopping, and ways to staymotivated during weight loss efforts. (Pg 1414/ Col 2/ Para 2 - Pg 1415/ Col 1/ Para 1) Control group: “All physicians, regardless of treatment condition, initially received 2 hours of instruction on general obesity management, as outlined by the National Heart, Lung and Blood Institute clinical practice guideline on the Identification, Evaluation and Treatment of Overweight and Obesity in Adults.” (Pg 1414/ Col 2/ Para 2). “Standard care physicians were instructed to provide their usual obesity management conducted during a typical office visit. Standard care participants received no special instructions and were seen, as needed, for regular medical care. Information provided by standard care participants during the initial assessment was not used during any subsequent office visit.” (Pg 1415/ Col 1/ Para 2) Timing of intervention: Proximity to clinical decision-making: Remote educational sessions (Pg 1414/ Col 2/ Para 2) Frequency/number of intervention events: Unclear (Pg 1414/ Col 2/ Para 2) Duration of intervention: 7 hours of additional training (Pg 1414/ Col 2/ Para 3) Healthcare professional recipient: Intervention group: Unclear - simply states all physicians received 2 hours of training and those providing tailored intervention received an additional 7 hours of training. (Pg 1414/ Col 2/ Para 2) Control group: Unclear - simply states all physicians received 2 hours of training (Pg 1414/ Col 2/ Para 2) Intervention deliverer: Intervention group: Unclear who delivered the training session. (Pg 1414/ Col 2/ Para 2) The treatment materials delivered by the physician were individually prepared and tailored to each patient by a multidisciplinary research team consisting of the physician, a health psychologist, a registered dietitian, and an exercise physiologist. (Pg 1414/ Col 2/ Para 3) Control group: Unclear (Pg 1414/ Col 2/ Para 2) Types of targeted behaviour of the health professionals: Tailored intervention to patients during six, monthly active treatment visits. (Pg 1414/ Col 2/ Para 3-4) Development of the intervention: Consultation with professional recipients: Unclear Evidence-base of intervention: Unclear -Although the two hours of training provided to all the physicians was evidence-based: 2 hours of instruction on general obesity treatment, as outlined by NHLBI 1998. (Pg 1414/ Col 2/ Para 2) The additional 7 hours training for the intervention group was not referenced. Consumer involvement: Unclear Barriers to change: Not done Source of funding for study: The study was supported by The National Institute of Diabetes and Digestive and Kidney Diseases (Grant R01 DK57476) and co-sponsored by the Centers for Disease Control and Prevention, the Centre for Chronic Disease Prevention and Health Promotion, Division of Nutrition and Physical Activity, and by the Office of Research on Women’s Health (Pg 1419/ Col 1/ Para 3) Ethical approval: Unclear if the boards that approved the study were ethics boards: The study was approved by the institutional review boards of the Pennington Biomedical Research Centre, the Louisiana State University Health Sciences Centre, and the Baton Rouge Medical Centre (Baton Rouge, LA). (Pg 1414/ Col 1/ Para 3)
Outcomes	Outcomes measured: Weight change BMI Length of time outcomes measured after initiation of the intervention: Baseline and 6 months (Pg 1414/ Col 1/ Para 4) Ceiling effect: Identified by investigator: Unclear Identified by reviewer: Unclear Losses to follow up: Number randomised: Intervention group: 71 (Fig 1/ Pg 1416) Control group: 73 (Fig 1/ Pg 1416) Number completing follow up: Intervention group: 48 (6 months) (Fig 1/ Pg 1416) Control group: 58 (6 months) (Fig 1/ Pg 1416) Reasons for loss to follow up: Intervention group: 8 lost to follow up (1 died, 7 lost contact). 4 missed 6-month appointment. 3 no longer met medical inclusion criteria. (Fig 1/ Pg 1416) Control group: 19 lost to follow up (5 scheduling conflicts, 14 lost contact). 1 missed 6-month appointment. 3 no longer met medical inclusion criteria. (Fig 1/ Pg 1416) Economic variables: Costs of the intervention: Not reported Changes in direct healthcare costs as a result of the intervention: Not reported Changes in non-healthcare costs as a result of the intervention: Not reported Costs associated with the intervention linked with provider or patient outcomes in an economic evaluation: Not reported
Notes	No unit of analysis error. See Page 1415/Col2/“Statistcial analysis” section
*Risk of bias*
Item	Authors’ judgement	Description
Adequate sequence generation?	Yes	(Pg 1414/ Col 1/ Para 1)
Allocation concealment?	Unclear	Not reported
Blinding? Primary outcome	Unclear	Patient weight change. Blinding not referred to.
Blinding? Secondary outcomes	Unclear	BMI
Blinding? Secondary outcomes	Unclear	N/A
Incomplete outcome data addressed? Primary outcome	Yes	Patient weight change: Good explanation of missing data and an intent-to-treat (ITT)1 analysis using baseline values carried forward for dropouts was completed. (Pg 1417/ Col 1/ Para 2). Overall 20% attrition rate; more drop outs from the intervention group (23 of 71 in the intervention group and 15 of 73 in the control group) . Also “Participants who dropped out of the study differed from study completers in that they tended to be younger, 35.4 (11.6) vs. 43.3 (11.6) years (P <= 0.01). In addition, the dropouts among the intervention group tended to have smaller waist circumferences (P < 0.01) and to be younger (P < 0.05) than the standard care dropouts. (Pg 1416/ Col 1/ Para 1)
Incomplete outcome data addressed? Secondary outcomes	Unclear	Not reported
Free of selective reporting?	No	Participants’ weight and BMI were obtained at baseline and 6 months (Pg 141 / Col 1/ Para 4) but only weight change was reported in Table 1/ Pg 1416
Baseline characteristics similar?	Yes	Patients in the intervention group are lighter, younger and have smaller waist circumferences but authors state that the differences are not statistical significant (Table 1/ Pg 1415 and Pg 1415 / Col 2/ Para 3)
Reliable primary outcome measures? Average weight change	Yes	Patient weight: Standard measurement method by single clinician (Pg 414/ Col 1/ Para 4)
Protection against contamination?	Unclear	Clinicians in the intervention and control group worked at the same two clinics so communication likely

Methods	Design: RCT Unit of allocation: Patients: Immediately after screening, the study dietitian used a table of random numbers to allocate each consecutive patient (Pg 312/ Col 2/ Para 3) Unit of analysis: Patients: A Chi² test was used to compare the demographic composition of the study groups. Confidence intervals for differences in means were used to compare groups with respect to outcome measurements. (Pg 313/ Col 1/ Para 4) Power calculation: Based on an expected 5% weight reduction in the dietitian group and 10% in the doctor/ dietitian group, a minimum of 35 overweight patients per group was required to achieve a power of 0.9 that the null hypothesis would be rejected at the 0.5 level. (Pg 312/ Col 2/ Para 3) Number expected = 35 × 3 groups = 105 (Pg 312/ Col 2/ Para 3) Number recruited = 273 (Pg 313/ Col 2/ Para 4) Number overweight recruited = 270 (Table 1/ Pg 313)
Participants	The number randomised into the trial: *Episodes of care:* Dietitian group: 6 sessions (Pg 312/ Col 2/ Para 5) Dietitician/GP group: 6 sessions with dietitian plus 3 sessions with GP (Pg 312/ Col 2/ Para 7) Control group: Baseline and endpoint assessment plus standard care (Pg 313/ Col 1/ Para 1) *Patients:* 270 (plus 3 patients who had hypertension and/or diabetes but who were not overweight) (Table 1/ Pg 313) *Providers:* 1 Dietitcian (Pg 312/ Col 1/ Para 4). Unclear numbers for GPs *Practices:* 1 GP practice (Pg 312/ Col 1/ Para 4) *Hospitals:* N/A *Communities or regions:* N/A Characteristics of participating healthcare providers: Profession: Physicians (general practitioners) but numbers not stated. 1 Dietitician (Pg 312/ Col 1/ Para 4) Level of training: Fully trained general practitioners and nutritionist (Pg 312 / Col 1/ Para 4) Age of health professional: Unclear Years since graduation or in practice: Unclear Proportion of eligible providers (or allocation units) who participated in the evaluation: Unclear: Only a single site was used (a university general practice) but the number and proportion of general practitioners in this group practice that participated was not stated (Pg 312/ Col 1/ Para 4) Characteristics of the participating patients: Clinical problem(s) of participating patients: Overweight (BMI ≥ 25 but ≤ 30): Unclear: Patients with a body mass index (BMI) of more than 25 were diagnosed as overweight. (Pg 312/ Col 1/Para 6). Total n = 270 overweight participants (Table 1/ Pg 313). Number who were overweight vs obese was not stated. Obese (BMI ≥ 30): Unclear: Patients with a body mass index (BMI) of more than 25 were diagnosed as overweight (312/1/6). Total n = 270 overweight participants (Table 1/ Pg 313). But number overweight vs number obese not stated. *Diabetes:* Unclear: N = 17 (but some of these may not have been overweight) (Table 1/ Pg 313). *Ischemic heart disease:* IHD not stated, but 97 had hypertension. Some of these may not have been overweight (Table 1/ Pg 313) Other characteristics of participating patients: NB these data based only all 273 participating patients of whom 3 were not overweight and therefore not included in the results for this review *Age:* Unclear: 73% of patients were less than 50 years old (Pg 313/ Col 2/ Para 4) Baseline Weight (kg), mean (no SD provided): Doctor/dietitian group (n = 92): 91.7 Dietitian (n = 88): 85.5 Usual care (n = 90): 89.1 Baseline hypertension (mean blood pressure = diastolic BP + (systolic BP − diastolic BP)/3, in mm Hg), mean (no SD provided): Doctor/dietitian group (n = 33): 112 Dietitian (n = 30): 109 Usual care (n = 34):110 Baseline type 2 diabetes (% glycated haemoglobin), mean (no SD provided): Doctor/dietitian group: (n = 6) 8.0 Dietitian (n = 5): 8.2 Usual care (n = 6): 7.7 *Gender:* 75 men and 198 women (Pg 313/ Col 2/ Para 4) *Ethnicity:* Unclear *Other:* Socio-economic status quartile: 58% most disadvantaged, 20% more disadvantaged, 2% least disadvantaged. Occupation: 56% home duties (84% female), 20% driver/trade/labourer, 6% unemployed. 14% clerical/sales, 4% manager/professional. 22% without partners; 78% married or de facto. 31% had hypertension. (Pg 313/ Col 2/ Para 5-6) Setting : Reimbursement system: Unclear Setting of care: General practice (Pg 312/ Col 1/ Para 4) Academic status of the setting of care: Non-teaching or university affiliated: a university general practice (Pg 312/ Col 1/ Para 4) Country: Australia (Lockridge, near Perth, Western Australia (Pg 312/ Col 1/ Para 4)
Interventions	Professional intervention: Intervention groups: Dietitian group: Patients allocated to the dietitian group were invited to join the study by the dietitian at the time of screening. The dietitian conducted six individual counselling sessions, spaced equally, with the last session 12 months after recruitment. The initial session occupied 45 minutes, with 15 minutes for later sessions. Measurements were repeated at all sessions under similar conditions. Counselling focused on principles of good nutrition and exercise. The dietitian questioned life style and dietary patterns to identify problem areas. Counselling included advice on food shopping and cooking methods, food selection, meal planning, and exercise programmes. Patient kept food records and diet history was used in the counselling sessions to provide individual advice. Recommendations included restriction of total dietary energy, reduction of the fat component to no more than 30%, with carbohydrate contributing 50% or more and protein the balance. Smoking was discouraged. Alcohol consumption of no more than two standard drinks a day for women and four for men was recommended, with at least two alcohol-free days a week. (Pg 312/ Col 2/ Para 5-6) GP and dietitian group: After screening, the dietitian flagged the patient record to request the general practitioner, with whom the patient had made an appointment, to invite the patient to join the study. Patients saw the same general practitioner on two other occasions during the 12 months to encourage the patient and monitor progress. The dietitian coordinated the follow-up appointments and flagged the patient record with progress measurements to enable the general practitioner to discuss progress with the patient. Five minutes of general practitioner time was allocated to these tasks. Otherwise, treatment was the same as for the dietitian group. (Pg 312/ Col 2/ Para 7-8) Control group: Standard care group: The control group received the results of the initial measurements and if they had queries were advised to discuss these with the doctor with whom they had made an appointment. No counselling was given by the dietitian. If patients asked the doctor about the measurements, they were treated as any other patient attending the practice. The fact that they were in the control group did not prevent the doctor from providing care usually provided for such conditions. This could include monitoring, advice and prescriptions, but not referral to the study’s dietitian. After 12 months, they received one mailed invitation to attend for reassessment of the initial measurements. In accordance with protocol, doctors were never informed about who was in the control and the dietitian groups. If a patient who was not in the doctor/dietitian asked about screening results, the doctor would not know to which group, if any, the patient belonged. (Pg 312/ Col 2/ Para 9 - Pg 313/ Col 1/ Para 2) Timing of intervention: *Proximity to clinical decision-making:* Dietitian group: Patients allocated to the dietitian group were invited to join the study by the dietitian at the time of screening. The dietitian conducted six individual counselling sessions, spaced equally, with the last session 12 months after recruitment. (Pg 312/ Col 2/ Para 5-6) GP and dietitian group: After screening, the dietitian flagged the patient record to request the general practitioner, with whom the patient had made an appointment, to invite the patient to join the study. Patients saw the same general practitioner on two other occasions during the 12 months to encourage the patient and monitor progress. The dietitian coordinated the follow up appointments and flagged the patient record with progress measurements to enable the general practitioner to discuss progress with the patient. Five minutes of general practitioner time was allocated to these tasks. (Pg 312/ Col 2/ Para 7-8) *Frequency/number of intervention events:* Unclear: Dietitians saw the patients on 2 additional occasions but unclear how often the dietitian flagged the patient records for the clinicians attention (Pg 312/ Col 2/ Para 7-8) *Duration of intervention:* “Five minutes of general practitioner time was allocated to these tasks” (i.e. to read notes flagged by dietitian and to discuss them with the patient) (Pg 312/ Col 2/ Para 7-8). The course of the intervention ran over 12 months (Pg 312/ Col 2/ Para 5-6) Healthcare professional recipient: *Intervention group:* N/A - organisation of care intervention *Control group:* N/A - organisation of care intervention Intervention deliverer: *Intervention group:* N/A - organisation of care intervention *Control group:* N/A - organisation of care intervention Types of targeted behaviour of the health professionals: N/A - organisation of care intervention Development of the intervention: Consultation with professional recipients: Not specified Evidence base of intervention: Not specified Consumer involvement: Not specified Barriers to change: Not done Source of funding for study: The research was funded by a grant from the Western Australian Health Promotion Foundation. (Pg 315/ Col 2/ Para 7) Ethical approval: Ethics approval was obtained from the Committee of Human Rights, The University of Western Australia. (315/2/8)
Outcomes	Outcomes measured: Changes in weight Blood pressure Glycated haemoglobin Cardiovascular medications use Costs Length of time outcomes measured after initiation of the intervention: Baseline and 12 months (Pg 313/ Col 2/ Para 1) Ceiling effect: *Identified by investigator:* Unclear *Identified by reviewer:* No (potential for weight loss clear and demonstrated) Losses to follow up: Number randomised (Table 1/ Pg 314) *Intervention groups:* Dietitian: 88 GP + dietitian: 92 Control group: 90 Number (%) completing follow up at 12 months (Table 1/ Pg 314) *Intervention groups:* Dietitian: 48 (55%) GP + dietitian: 65 (71%) Control group: 64 (71%) Reasons for loss to follow up: *Intervention groups:* Unclear Control group: Unclear Economic variables: Costs of the intervention: Yes (Table 3/ Pg 314) Total cost per group Cost per patient Additional cost per patient Additional cost per kg lost Costs included were dietitian and clinician time, materials, room use and usual practice overheads. (Pg 313/ Col 1/ Para 6) Changes in direct healthcare costs as a result of the intervention: No Changes in non-healthcare costs as a result of the intervention: No Costs associated with the intervention linked with provider or patient outcomes in an economic evaluation: Yes (Table 3/ Pg 314): Additional cost per patient Additional cost per kg lost
Notes	No unit of analysis error: See page 313/Col1/“outcome and statistical methods” section
*Risk of bias*
Item	Authors’ judgement	Description
Adequate sequence generation?	Yes	Used a table of random numbers to allocate each consecutive patient (Pg 312/ Col 2/ Para 4)
Allocation concealment?	Unclear	Immediately after screening, the study dietitian used a table of random numbers to allocate each consecutive patient (Pg 312/ Col 2/ Para 4)
Blinding? Primary outcome	No	Changes in weight: “doctors were never informed about who was in the control and the dietitian groups” (Pg 313/ Col 2/ Para 1) However, doctors were obviously not blinded to those who were in the doctor/dietitian versus dietitian group. No reference is made to blinding of outcome assessors
Blinding? Secondary outcomes	No	Hypertension: “doctors were never informed about who was in the control and the dietitian groups” (Pg 313/ Col 2/ Para 1) However, doctors were obviously not blinded to those who were in the doctor/dietitian versus dietitian group. No reference is made to blinding of outcome assessors. Costs: The dietitian maintained a record of activities for two periods of two weeks during the study. Time spent on the study tasks of screening, arranging appointments, changing appointments, drawing patient files, data entry, and counselling was recorded. The time was coasted at $20 per hour for the dietitian. Time spent by the patient with the doctor was coasted at $82 per hour, which was the equivalent cost of bulk billing four standard consultations. Materials used by the dietitian, roomrental and usual practice overheads were coasted and distributed according to the number of counselling sessions taken up by patients in the three groups. The cost effectiveness analysis was used to determine a cost for each intervention in terms of weight change over and above that of the screening group. (Pg 313/Col 1/ Para 6)The assessor of time taken and equipment used was not blinded
Blinding? Secondary outcomes	Unclear	Random glucose level: For diabetic patients, an Ames pen was used to obtain a capillary sample that was read with an Ames-3 glucometer to obtain a random glucose glucose level. (Pg 312/ Col 1/ Para 8) Glycated haemoglobin: Venous blood was also taken from diabetic patients for glycated haemoglobin at the beginning and the end of the study (Pg 312/ Col 2/ Para 1) No reference is made to blinding of outcome assessors but obviously they could have easily been blinded for these two outcomes
Incomplete outcome data addressed? Primary outcome	Yes	Changes in weight: Missing data have been imputed using appropriate methods: The main outcomes evaluated were changes in weight and mean blood pressure (diastolic pressure + (systolic-diastolic pressure)/ 3) for each of the three groups. These outcomes were subjected to analysis by intention- to-treat, which assumed that a patient’s measurements remained unchanged after the patient dropped out of the study. Thus a patient’s last measurement was used to populate all subsequent missing data values (Pg 313/ Col 1/ Para 4)
Incomplete outcome data addressed? Secondary outcomes	Unclear	Changes in blood-pessure: Missing data have been imputed using appropriate methods i.e intention to treat analysis Changes in glycated haemoglobin: see above Changes in medication use:
Free of selective reporting?	Unclear	All of the study’s pre-specified primary outcomes (weight and blood pressure) have been reported (Table 1/ Pg 314):
Baseline characteristics similar?	Yes	Authors state that there were no significant differences between intervention and control groups with respect to sex or age or socioeconomic status quartiles or occupation. (Pg 313 / Col 2/ Para 4-5 and Table 1/ Pg 313)
Reliable primary outcome measures? Average weight change	Yes	Collected by individual: screened opportunistically by the study dietitian (312/1/4). Body weight and height were measured with patients wearing only light indoor clothing. Body weight was measured on digital balance scales to the nearest 0.1 kg with the patient wearing no shoes. (Pg 312/ Col 1/ Para 6)
Protection against contamination?	No	The same GP could have delivered care to patients in an intervention group and to patients receiving standard care

Methods	Design: RCT Unit of allocation: Unclear Of the eligible patients 484were randomly selected and assigned to either an intervention or control group (Pg 64/ Col 2/ Para 1). Physicians participating in the study were randomly divided into three groups: 1) those who were to see only patients with automated records available; 2) those who were to see patients without automated records; and 3) those whose patient load was approximately half with and half without automated records (Pg 64/ Col 2/ Para 2).The relationship between physician groupings and intervention and control group is not explained Unit of analysis: Unclear.“The analysis of variance and the analysis of covariance were used to compare the experimental and control conditions on blood pressure and weight measurements” (Pg 65/Col 2/Para 1) Power calculation: No power calculation.
Participants	The number randomised into the trial: Episodes of care: Not available (Table 1/ Pg 67) *Patients:* 147 obese patients (Table 1/ Pg 67) *Providers:* Unclear - number of physicians not stated (Pg 64/ Col 2/ Para 2) *Practices:* N/A *Hospitals:* 1 - The Cardiac Pulmonary and Renal Clinics of the Northwestern University (Pg 64/ Col 1/ Para 3) *Communities or regions:* N/A Characteristics of participating healthcare providers: Profession: Physicians but number not stated (Pg 64/ Col 2/ Para 2) Level of training: Not stated (Pg 64/ Col 2/ Para 2) Age of health professional: Not stated (Pg 64/ Col 2/ Para 2) Years since graduation or in practice: Not stated (Pg 64/ Col 2/ Para 2) Proportion of eligible providers (or allocation units) who participated in the evaluation: Not stated (Pg 64/ Col 2/ Para 2) Characteristics of the participating patients: Clinical problem(s) of participating patients: Overweight (BMI ≥ 25 but ≤ 30): Unclear: (Table 1/ Pg 66), Patients whose weight exceeded 20%of their ideal weight were classified as obese - but breakdown not available Obese (BMI ≥ 30): Unclear: (Table 1/ Pg 66), Patients whose weight exceeded 20% of their ideal weight were classified as obese - but breakdown not available *Diabetes:* (Table 1/ Pg 67): 48/147 obese i.e. 33.3% *Ischemic heart disease:* Not available Other characteristics of participating patients: *Age:* Not available (Table 1/ Pg 67) *Gender:* 88 F, 26M (77% female) (Table 5/ Pg 71) *Ethnicity:* Not available (Table 1/ Pg 67) *Other:* N/A Setting: Reimbursement system: Unclear Setting of care: The Cardiac Pulmonary and Renal Clinics of the Northwestern University (Pg 64/ Col 1/ Para 3) Academic status of the setting of care: University (teaching) hospital: The Cardiac Pulmonary and Renal Clinics of the Northwestern University (Pg 64/ Col 1/ Para 3). Country: USA: Michigan (Pg 63/ Col 1/ Para 5)
Interventions	Professional intervention: Intervention group: In the experimental group patients had available a computer printout of a current NUCRSS summary in addition to the traditional medical record. (Pg 64/ Col 2/ Para 1) A computerised medical record system (NUCRSS) was developed to provide physicians with concise and current information on patient’s problems, to identify omissions in recording of observations and treatment recommendations, to show ordered procedures that were not carried out, to record deficiencies in medical reasoning, and most importantly, to recommend corrective actions according to selected criteria. These criteria of “good care” were established by consensus of the physicians providing care at our university (Pg 64/ Col 1/ Para 3) Control group: The control group had available only the handwritten traditional medical record (Pg 64/ Col 2/ Para 1) Timing of intervention: Proximity to clinical decision-making: Immediately proximate to clinical decision making. In the experimental group patients had available a computer printout of a current NUCRSS summary in addition to the traditional medical record (Pg 64/ Col 2/ Para 1). The control group had available only the handwritten traditional medical record (Pg 64/ Col 2/ Para 1) *Frequency/number of intervention events:* Not stated *Duration of intervention:* N/A Intervention deliverer: *Intervention groups:* (i) Computer system (NUCRSS) (Pg 64/ Col 1/ Para 3) and (ii) 50% Computer system (NUCRSS) and 50% handwritten traditional medical record *Control group:* The control group had available only the handwritten traditional medical record (Pg 64/ Col 2/ Para 1) Types of targeted behaviour of the health professionals: The NUCRSS keeps track of weight loss progress and reminds physicians to review or change diets (Pg 72/ Col 2/ Para 1) Development of the intervention: Consultation with professional recipients: Yes: The criteria (used by NUCRSS) of good care were established by consensus of the physicians providing care at our university (Pg 64/ Col 1/ Para 3) Evidence-base of intervention: Explicitly not evidence-based: These criteria of “good care” were established by consensus of the clinicians providing care at our university (Pg 64/ Col 1/ Para 3) Consumer involvement: Not specified Barriers to change: Not done Source of funding for study: The major support for this project was provided by Grant- Number HS02649 from the National Centre for Health Services Researcgh, HRA. Initial data collection and analysis were made possible by DHEW Grant number H500674-04 and USPAS Grant number RR05370 (NIH). (Pg 63/ Col 1/ Para 4) Ethical approval: Not reported
Outcomes	Outcomes measured: Pounds overweight (patients) Failure to give or review diet (clinicians) Length of time outcomes measured after initiation of the intervention: Baseline Year 1 (10-15 months) Year 2 (22-24 months) (Table 4/ Pg 69) Ceiling effect: *Identified by investigator:* No *Identified by reviewer:* No - see Table 3/ Pg 68 Losses to follow up: NB - data not available for all patients at 1 year and 2 year follow up time points Number randomised: *Intervention group:* 68 (Table 1/ Pg 66) *Control group:* 79 (Table 1/ Pg 66) Number completing follow up: *Intervention group:* 62 (at end of study) (Table 1/ Pg 66) *Control group:* 62 (at end of study) (Table 1/ Pg 66) Reasons for loss to follow up: *Intervention group:* 1 dead, 5 moved (at end of study) (Table 1/ Pg 66) *Control group:* 7 dead, 10 moved (at end of study) (Table 1/ Pg 66) Economic variables: Costs of the intervention: Not reported Changes in direct healthcare costs as a result of the intervention: Not reported Changes in non-healthcare costs as a result of the intervention: Not reported Costs associated with the intervention linked with provider or patient outcomes in an economic evaluation: Not reported
Notes	Unclear whether there was a unit of analysis error: See Page 65/Col 2/Top para
*Risk of bias*
Item	Authors’ judgement	Description
Adequate sequence generation?	Unclear	Pg 62/ Col 2/ Para 1-2
Allocation concealment?	Unclear	Pg 62/ Col 2/ Para 1-2
Blinding? Primary outcome	Yes	Weight loss: Blind retrospective chart reviews were done for both experimental and control patients (Pg 64/ Col 2/ Para 3)
Blinding? Secondary outcomes	Yes	Dietary advice: Blind retrospective chart reviews were done for both experimental and control patients (Pg 64/ Col 2/ Para 3)
Blinding? Secondary outcomes	Unclear	N/A
Incomplete outcome data addressed? Primary outcome	No	Weight loss: Table 5/ Pg 71. Obese patients entered into study: 68 computer assisted, 79 handwritten notes Baseline: 55 computer group, 59 handwritten group 12 months: 55 computer group, 57 handwritten group 24 months: 46 computer group, 44 handwritten group Dropout info (Table 1/ Pg 66) does not even cover all of the dropouts by baseline
Incomplete outcome data addressed? Secondary outcomes	Unclear	Dietary advice: Table 3/ Pg 68 does not state how many patients still in study at each time point, just number who were given advice
Free of selective reporting?	No	Not all of the study’s pre-specified primary outcomes have been reported (Pg 64/ Col 2/ Para 3): The database consisted of: 1) items related to the utilisation of services and to the overall quality of care (e.g. number of clinic visits, yearly routine physical examinations etc) 2) more detailed information such as the presence or absence of recommended laboratory examinations for patients with obesity 3) answers to a questionnaire concerning patients’ views on their own health and on the care received i.e. only some of 1) is reported, its is unclear what tests 2) were to be ordered for obesity, and 3) is not reported at all
Baseline characteristics similar?	Unclear	The number of males and females is not given in the baseline Table 1/ Pg 67 but sex is used to divide the results later. Also information not given for the proportion of obese patients with diabetes and length of prior clinic attendance
Reliable primary outcome measures? Average weight change	Yes	Weight loss: Blind retrospective chart reviews were done for both experimental and control patients (Pg 64/ Col 2/ Para 3)
Protection against contamination?	Unclear	No information was provided to assure us that there were no mis-allocated patients, or that computerised records were always available

2.1.1 Unit of allocation (i.e. who or what was allocated to study groups, and was it cluster or individual randomisation)	Relevant supporting text and location. (page/column/paragraph)
Patient
Episode of care
Clinic Day
Provider
Firm
Practice
Institution
Community
Other: (Please specify)
UNCLEAR

2.1.2 Unit of analysis (e.g. results analysed as events per practice)	Relevant supporting text and location. (page/column/paragraph)
Patient
Episode of care
Clinic Day
Provider
Firm
Practice
Institution
Community
Other: (Please specify)
UNCLEAR

	2.2 Power calculation:	Relevant supporting text and location. (page/column/paragraph)
YES	Study has sufficient statistical power to detect clinically important effects as statistically significant
	Number expected to be recruited / number actually recruited
NO	No power calculation
UNCLEAR

2.3.1 SEQUENCE GENERATION Was the allocation sequence adequately generated? Describe the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups	Relevant supporting text and location. (page/column/paragraph)
The unit of allocation was health professional, patient or episode of care and the investigators describe a random component in the sequence generation process such as: · Referring to a random number table · Using a computer random number generator · Coin tossing · Shuffling cards or envelopes · Throwing dice · Drawing of lots · Minimization* *Minimization may be implemented without a random element, and this is considered to be equivalent to being random.
The unit of allocation was health professional, patient or episode of care and the investigators describe a quasi-random component in the sequence generation process such as: · Sequence generated by odd or even date of birth; · Sequence generated by some rule based on date (or day) of admission; · Sequence generated by some rule based on hospital or clinic record number
The investigators describe a non-random component in the sequence generation process. E.g · Allocation by judgement of the clinician; · Allocation by preference of the participant; · Allocation by availability of the intervention.
Insufficient information about the sequence generation

	2.3.2 ALLOCATION CONCEALMENT Was allocation adequately concealed? Describe the method used to conceal the allocation sequence in sufficient detail to determine whether intervention allocations could have been foreseen in advance of, or during, enrolment	Relevant supporting text and location. (page/column/paragraph)
YES	The unit of allocation was health professional, patient or episode of care and participants and investigators enrolling participants could not foresee assignment because one of the following, or an equivalent method, was used to conceal allocation: · Central allocation (including telephone, web-based, and pharmacy-controlled, randomization); · Sequentially numbered, opaque, sealed envelopes. For cluster randomisation where it is possible that randomisation of all units happens once. Id usually look to have some statement of allocation by an independent statistician
NO	Participants or investigators enrolling participants could possibly foresee assignments and thus introduce selection bias, such as allocation based on: · Using an open random allocation schedule (e.g. a list of random numbers); · Assignment envelopes were used without appropriate safeguards (e.g. if envelopes were unsealed or nonopaque or not sequentially numbered); · Alternation or rotation; · Date of birth; · Case record number; · Any other explicitly unconcealed procedure. Again for cluster randomisation the judgement is whether a study where allocation was performed by the study statistician is regarded as biased
UNCLEAR	Insufficient information to permit judgement of Yes or No. This is usually the case if the method of concealment is not described or not described in sufficient detail to allow a definite judgement? E.g. if the use of assignment envelopes is described, but it remains unclear whether envelopes were sequentially numbered, opaque and sealed

2.3.3.1 BLINDING OF OUTCOME ASSESSORS: Was knowledge of the allocated interventions adequately prevented during the study? Describe all measures used, if any, to blind the outcome assessors from knowledge of which intervention a participant received. Provide any information relating to whether the intended blinding was effective
YES	Any one of the following: · No blinding, but the review authors judge that the outcome and the outcome measurement are not likely to be influenced by lack of blinding; · Blinding of the outcome assessors ensured, and unlikely that the blinding could have been broken
NO	Any one of the following: · No blinding or incomplete blinding, and the outcome or outcome measurement is likely to be influenced by lack of blinding; · Blinding of the outcome assessors attempted, but likely that the blinding could have been broken
UNCLEAR	Any one of the following: · Insufficient information to permit judgement of Yes or No; · The study did not address this outcome.
Reported Outcome(s) (Add rows as necessary)		Low Risk of Bias: YES/NO/UNCLEAR

2.3.4.1 INCOMPLETE OUTCOME DATA: Were incomplete outcome data adequately addressed? Describe the completeness of outcome data for each main outcome, including attrition and exclusions from the analysis. State whether attrition and exclusions were reported, the numbers in each intervention group (compared with total randomized participants), reasons for attrition/ exclusions where reported, and any re-inclusions in analyses performed by the review authors
YES	Any one of the following: · No missing outcome data; · Reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); · Missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; · For dichotomous outcome data, the proportion of missing outcomes compared with observed event risk not enough to have a clinically relevant impact on the intervention effect estimate; · For continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size; · Missing data have been imputed using appropriate methods
NO	Any one of the following: · Reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; · For dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in intervention effect estimate; · For continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes enough to induce clinically relevant bias in observed effect size; · As-treated analysis done with substantial departure of the intervention received from that assigned at randomization; · Potentially inappropriate application of simple imputation
UNCLEAR	Any one of the following: Insufficient reporting of attrition/exclusions to permit judgement of Yes or No (e.g. number randomized not stated, no reasons for missing data provided); The study did not address this outcome.
Reported Outcome(s) (Add rows as necessary)		Low Risk of Bias: YES/NO/UNCLEAR

	2.3.5 SELECTIVE OUTCOME REPORTING Are reports of the study free of suggestion of selective outcome reporting? State how the possibility of selective outcome reporting was examined by the review authors, and what was found NB. KD will try to find study protocols if not present in your paper, you dont need to do this	Relevant supporting text and location. (page/column/paragraph)
YES	Any of the following: · The study protocol is available and all of the studies pre-specified (primary and secondary) outcomes that are of interest in the review have been reported in the prespecified way; · The study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were pre-specified (convincing text of this nature may be uncommon)
NO	Any one of the following: · Not all of the studies pre-specified primary outcomes have been reported; · One or more primary outcomes is reported using measurements, analysis methods or subsets of the data (e.g. sub-scales) that were not pre-specified; · One or more reported primary outcomes were not pre-specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect); · One or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta-analysis; · The study report fails to include results for a key outcome that would be expected to have been reported for such a study
UNCLEAR	Insufficient information to permit judgement of Yes or No. It is likely that the majority of studies will fall into this category

2.3.6 BASELINE MEASUREMENT	Relevant supporting text and location. (page/column/paragraph)
Performance or patient outcomes measured prior to the intervention, and no substantial differences present across study groups in main outcome measures and also in possible confounding variables (e.g. sex, age)
Differences at baseline in main outcome measures or confounding variables (e.g. sex, age) likely to undermine the post intervention differences, e.g. differences between groups before the intervention similar to those found post intervention or had extreme baseline imbalance
Baseline measures not reported, or unclear whether baseline measures are different across study groups

2.3.7 RELIABLE PRIMARY OUTCOME MEASURE(S)
YES	Two or more raters with agreement ≥ 90% or kappa ≥ 0.8 OR outcome assessment is objective, e. g. length of hospital stay, drug levels assessed by a standardised test
NO	Two or more raters with agreement < 90% or kappa < 0.8.
UNCLEAR	Reliability not reported for outcome measures obtained by chart extraction or collected by an individual
Reported Outcome(s) (Add rows as necessary)		Low Risk of Bias: YES/NO/UNCLEAR

PERMALINK

Interventions to change the behaviour of health professionals and the organisation of care to promote weight reduction in overweight and obese adults

Gerd Flodgren

Katherine Deane

Heather O Dickinson

Sara Kirk

Hugh Alberti

Fiona R Beyer

James G Brown

Tarra L Penney

Carolyn D Summerbell

Martin P Eccles

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors’ conclusions

BACKGROUND

Description of the condition

Description of the intervention

Why it is important to do this review

OBJECTIVES

METHODS

Criteria for considering studies for this review

Types of studies

Types of participants

Settings of studies

Types of interventions

i. Interventions targeting health professionals

ii. Interventions targeting the organisation of care

Comparators

Excluded studies

Types of outcome measures

Main outcomes

Other outcomes

Patient outcomes

Health professional outcomes

Search methods for identification of studies

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

RESULTS

Description of studies

Results of the search

Included studies

Table 1. Summary of methods and participants of included studies.

Interventions targeting health professionals

Interventions targeting the organisation of care

Excluded studies

Risk of bias in included studies

Figure 1. Risk of bias graph: review authors’ assessment of each risk of bias domain presented as percentages across all included studies.

Figure 2. Risk of bias summary: review authors’ assessment of the risk of bias of the individual domains for each included study.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Interventions targeting health professionals versus standard care

Interventions targeting the organisation of care versus standard organisation of care

DISCUSSION

Summary of main results

Interventions targeting health professionals

Interventions targeting the organisation of care

Overall completeness and applicability of evidence

Quality of the evidence

Potential biases in the review process

Agreements and disagreements with other studies or reviews

	2.3.8 PROTECTION AGAINST CONTAMINATION	Relevant supporting text and location. (page/column/paragraph)
YES	Allocation by community, institution or practice and unlikely that control group received the intervention
NO	Likely that control group received the intervention, e.g. cross-over trials or if patients rather than professionals were randomised
UNCLEAR	Professionals allocated within a clinic or practice and possible that communication between experimental and control group professionals could have occurred

2.1.1 Profession (mark all appropriate): Please state the numbers of each profession involved. Also please note if the numbers come from baseline, the remaining population at the endpoint, or other time period (e.g. sequential accrual)	Relevant supporting text and location. (page/column/paragraph)
Physicians
Nurses
Pharmacists
Physiotherapists
Dietitianss/Nutritionists
Psychologists
Other: (Please specify)
UNCLEAR

3.1.2 Level of training:	Relevant supporting text and location. (page/column/paragraph)
In post-graduate training (House Officer/Intern, Registrar/Resident)
Fully trained (Consultant/Attending)
Mixed
Other (Specify i.e. copy all information available in paper)
UNCLEAR (information not available)

3.1.3 Age of health professional:	Relevant supporting text and location. (page/column/paragraph)
Mean age
UNCLEAR (information not available)

3.1.4 Years since graduation or in practice:	Relevant supporting text and location. (page/column/paragraph)
Mean
UNCLEAR (information not available)

3.1.5 Proportion of eligible providers (or allocation units) who participated in the evaluation:	Relevant supporting text and location. (page/column/paragraph)
Report the numbers or the percentage of providers in target population who were allocated to study groups
UNCLEAR (information not available)

3.2.1 Clinical problem(s) of participating patients: Please give information on the authors definitions of the conditions e.g. over 5lbs over the recommended maximum weight for their height. Please also note the numbers with each condition and if they come from baseline, the remaining population at the endpoint, or other time period (e.g. sequential accrual)	Relevant supporting text and location. (page/column/paragraph)
3.2.1.1 Overweight (BMI over 25 but less than 30)
UNCLEAR (information not available)
3.2.1.2 Obese (BMI 30 or over)
UNCLEAR (information not available)
3.2.1.3 Diabetes
UNCLEAR (information not available)
3.2.1.4 Ischemic heart disease
UNCLEAR (information not available)

3.2.2 Other characteristics of participating patients: Please note if the numbers come from baseline, the remaining population at the endpoint, or other time period (e.g. sequential accrual)	Relevant supporting text and location. (page/column/paragraph)
3.2.2.1 Age:
Mean
Range
UNCLEAR (information not available)
3.2.2.2 Gender
UNCLEAR (information not available)
3.2.2.3 Ethnicity
UNCLEAR (information not available)
3.2.2.4 Other (Please specify)
UNCLEAR (information not available)

3.2.3 The number randomised into the trial (i.e. all those who actually entered the study)	Relevant supporting text and location. (page/column/paragraph)
3.2.3.1 Episodes of care:
UNCLEAR (information not available)
3.2.3.2 Patients
UNCLEAR (information not available)
3.2.3.3 Providers
UNCLEAR (information not available)
3.2.3.4 Practices
UNCLEAR (information not available)
3.2.3.5 Hospitals
UNCLEAR (information not available)
3.2.3.6 Communities or regions
UNCLEAR (information not available)

3.3.1 Reimbursement system:	Relevant supporting text and location. (page/column/paragraph)
Fee for service (provider paid for number and type of services delivered)
Capitation (provider paid set amount per patient for providing specific care)
Prospective payment
Global budget
Mixed
UNCLEAR