Obesity is one of the largest modifiable threats to public health in developed countries(1). It affects a large proportion of the population in developed countries and is associated with chronic diseases such as CVD, type 2 diabetes and some cancers(Reference Pi-Sunyer2). The rates of overweight and obesity have been steadily increasing in countries such as the USA, Australia and the UK(3). A modest weight loss of 5 % in those obese has been shown to be beneficial in improving blood sugar control, CVD-related biomarkers and overall quality of life(Reference Pi-Sunyer4).
Primary-care physicians provide first-line health care in many countries. In Australia, more than 80 % of the population consult their primary-care physician at least once per annum(Reference Britt, Miller and Charles5). The average primary-care physician consultation rate in the UK rose from 3·9 consultations per person in 1995 to 5·3 in 2006(Reference Hippisley-Cox, Fenty and Heaps6). While more women and older people present for care(Reference Britt, Miller and Charles5), primary-care physicians still have access to a large proportion of the general population. Both patients(Reference Tan, Zwar and Dennis7) and physicians(Reference Thuan and Avignon8) perceive weight management to be part of a primary-care physician's role. Primary-care physicians have reported being interested in helping patients manage their weight, but face practical constraints in doing so(Reference Ruelaz, Diefenbach and Simon9). Those who have been advised by their primary-care physician to lose weight are more likely to try to do so(Reference Sciamanna, Tate and Lang10). Primary-care physicians are also likely to have multiple opportunities to identify excess weight and deliver ongoing weight-management care required for sustained weight loss.
Despite the advantages of using primary care for interventions targeting obesity, the effectiveness of interventions in this setting has not been widely evaluated. Previous systematic reviews have identified bariatric surgery(Reference Colquitt, Picot and Loveman11) and pharmacological treatments(Reference Norris, Zhang and Avenell12) as potentially effective methods for weight reduction; however, these interventions are costly and are usually indicated for the morbidly obese or those obese with coexisting conditions(Reference Colquitt, Picot and Loveman11). Behavioural, non-pharmacological interventions promoting dietary restrictions show some promise in producing moderate, short-term weight loss and are associated with fewer adverse events than pharmacological or surgical interventions(Reference Norris, Zhang and Avenell13). However, most studies have evaluated behavioural interventions in selected patient groups or in community groups, with few specifically targeting primary-care patients.
UK(14) and Australian(15) preventive guidelines recommend that primary-care physicians assess patients for overweight and obesity and develop appropriate weight-management plans. The US Preventive Services Task Force recommends that ‘intensive counselling and behavioural interventions’ be offered to all obese primary-care patients; with high intensity being defined as more frequent than monthly contact offered in the first 3 months of treatment(Reference McTigue, Harris and Hemphill16). A review by Tsai et al., which included studies conducted only in the USA, reported that the use of pharmacological treatment (i.e. sibutramine and orlistat) accompanied by brief physician counselling or the use of meal replacements with dietitian-delivered counselling were potentially effective strategies for weight reduction in primary-care patients(Reference Tsai and Wadden17). As their review was limited just to studies conducted in the USA, there is a need to examine weight-loss interventions in other countries so that findings are relevant to practitioners located outside the US health-care system. With the recent removal of sibutramine from the European, US and Australian markets, findings regarding the effectiveness of this drug may no longer be relevant to practitioners. Further, consideration of the methodological rigour of studies is important to ensure that valid conclusions are drawn. The present review aims to describe the number, methodological rigour and effectiveness of behavioural intervention studies involving primary-care physicians that targeted weight loss in overweight or obese adult primary-care patients, met the Cochrane Effective Practice and Organisation of Care Group (EPOC) study design criteria(18) and were published between 1999 and 2011.
Methods
The MEDLINE, EMBASE, Cochrane, PsycINFO and SCOPUS databases were searched using the following search terms: ‘obesity OR overweight OR weight loss’ AND ‘primary health care OR family practice OR general practice OR general practitioner OR physician patient relations OR guideline adherence’. The search was limited to completed studies, published in English from 1999 until December 2011. This time frame was selected because Tsai et al.'s review examining interventions in primary-care patients identified few studies published before 1999. The reference lists of relevant systematic reviews and studies were manually searched to identify additional studies. No additional studies were identified.
Inclusion criteria
Participants
Adult primary-care patients (aged ≥18 years) who were overweight or obese (defined as BMI ≥ 25·0 kg/m2) were included. Studies of interventions targeting specific patient groups (i.e. diabetes, hypertension) were included if the study specified overweight or obesity as an inclusion criteria.
Interventions
Studies aimed at reducing weight in overweight and obese primary-care patients were included. This encompassed behavioural interventions delivered by primary-care physicians alone or in conjunction with other personnel. Comparative trials where another intervention was compared with intervention(s) delivered by primary-care physicians were also included. Surgical and pharmacological interventions as well as studies where primary-care physicians were not involved in any component of the intervention were excluded.
Outcomes
Eligible studies included weight loss or/and reduction in BMI as an outcome. Weight/BMI change were chosen as the main outcomes as studies focused on other outcomes (such as physical activity levels, nutrition changes, biochemistry data) may not provide an adequate basis for identifying effective approaches for directly addressing overweight and obesity.
Study design
The following study designs that met the EPOC research criteria were included: randomised controlled trial (RCT), controlled clinical trial (CCT), controlled before-and-after study (CBA) and interrupted time series (ITS)(18).
Quality assessment
The EPOC quality criteria for RCT, CCT and CBA were used to assess the methodological adequacy of included studies(19). For each criterion a score of ‘yes’ was assigned if the study met the criterion, ‘no’ if it did not and ‘unclear’ if there was insufficient information to adequately decide if the criterion was met. A score out of nine for each study was reported.
Data extraction
The following were extracted by two authors independently.
Participants and intervention
Participant characteristics (including percentage of females, age, ethnicity and mean BMI) were extracted. Information related to the intervention, number of participants in each group, retention rate, mean weight change and whether statistically significant weight loss was achieved was also extracted. Whether a larger percentage of participants in the intervention group achieved clinically significant weight loss (for the purpose of the present review, this was defined as having a weight loss of more than 5 % of initial body weight) compared with the control group was recorded. A weight loss of 5 % or more of initial body weight has been shown to result in improvements in weight-related comorbidities(Reference Wing, Lang and Wadden20, 21). Where two intervention arms existed, comparisons between intervention and control group were reported.
Intensity
Intensity of interventions were coded as ‘low’, ‘moderate’ or ‘high’ based on frequency of contact in the first 3 months. An intervention was defined as high intensity if there was more than monthly contact, moderate if monthly contact and low if less than monthly contact occurred in the first 3 months of the intervention(Reference McTigue, Harris and Hemphill16). Where there was insufficient information, intensity was coded as ‘unsure’.
Quality assurance
All abstracts were reviewed by one researcher (S.L.Y.) and full-text articles of potentially relevant articles were retrieved. As a quality assurance measure, 10 % of the abstracts were reviewed and coded independently by a second reviewer (M.C., A.G.). All coding for quality criteria and data extraction were carried out by two authors (S.L.Y., A.G.) and differences resolved by mutual discussion.
Results
A total of 1356 articles were obtained from the electronic search: Medline (n 933), Cochrane (n 105), SCOPUS (n 280) and PsycINFO (n 38).
Seventeen articles describing sixteen studies met the inclusion criteria (see Fig. 1). Martin et al. published findings from the same study at the end of the intervention(Reference Martin, Rhode and Dutton22) and 2 years’ follow-up(Reference Martin, Dutton and Rhode23). All included studies were RCT except for one, which was a CBA(Reference Schuster, Tasosa and Terwoord24). One study was included as an RCT, although only two out of the three study arms were randomised(Reference Munsch, Biedert and Keller25). Only findings from the randomised groups were reported. Two studies did not have a control group but compared different interventions(Reference Ashley, St Jeor and Schrage26, Reference Willaing, Ladelund and Jorgensen27). A study by Wadden et al. was included although it had an intervention arm that included the use of pharmacology (sibutramine). Only results from the brief intervention group, which did not involve medication, are reported here(Reference Wadden, Volger and Sarwer28).
Fig. 1 Selection of articles for inclusion in the present systematic review (EPOC, Effective Practice and Organisation of Care Group)
One study met all nine EPOC criteria(Reference Moore, Summerbell and Greenwood29) and six others met seven or eight of the nine criteria(Reference Martin, Rhode and Dutton22, Reference Wadden, Volger and Sarwer28, Reference Appel, Clark and Yeh30–Reference Logue, Sutton and Jarjoura33). Five studies met less than half the criteria(Reference Schuster, Tasosa and Terwoord24–Reference Ashley, St Jeor and Schrage26, Reference Bolognesi, Nigg and Massarini34, Reference Ely, Banitt and Befort35) (see Table 1). Adequate protection against contamination and study free from selective outcome reporting were the least likely criteria to have been sufficiently met. There was often inadequate information to determine whether mechanisms had been put in place to prevent contamination between the intervention and control groups. Only four studies had published protocols, thus allowing for assessment of whether selective outcome reporting had occurred(Reference Wadden, Volger and Sarwer28–Reference Appel, Clark and Yeh30, Reference Jebb, Ahern and Olson32).
Table 1 Methodological assessment of included intervention studies based on the EPOC risk of bias criteria
EPOC, Effective Practice and Organisation of Care Group; RCT, randomised controlled trial; CBA, controlled before-and-after study.
Legend: ✓ = yes; × = no; ? = unclear.
The interventions were broadly categorised into: (i) lifestyle counselling delivered primarily by primary-care physicians; (ii) lifestyle counselling delivered primarily by non-primary-care physicians; and (iii) multi-component intervention.
Six studies examined the effectiveness of lifestyle counselling delivered primarily by primary-care physicians (see Table 2). Of these, three tested the use of brief, tailored lifestyle counselling targeting dietary and/or exercise behaviour in changing patients’ weight compared with usual care(Reference Martin, Rhode and Dutton22, Reference Martin, Dutton and Rhode23, Reference Christian, Bessesen and Byers31, Reference Bolognesi, Nigg and Massarini34) and one examined the effect of a physician-delivered group weight-management programme(Reference Munsch, Biedert and Keller25). Two studies targeted providers, with one assessing the effectiveness of providing an educational intervention(Reference Moore, Summerbell and Greenwood29) and the other testing the use of a sticker in overweight/obese patients’ charts representing diagnosis and treatment or referral for the condition(Reference Schuster, Tasosa and Terwoord24). Of the six studies, three reported on low-intensity, one on moderate-intensity and two on high-intensity interventions.
Table 2 Weight-loss interventions in primary-care patients delivered by primary-care physicians
%F, percentage of females; Y, yes; N, no; DNR, did not report; n.s., not specified; T2DM, type 2 diabetes mellitus; IG, intervention group; CG, control group; NHLBI, National Heat, Lung, and Blood Institute; SES, socio-economic status; GP, general practitioner(s); ↑, weight gain; ↓, weight loss.
To convert lb to kg, multiply lb by 0·4534.
Comparison between IG and CG: **P < 0·01, ***P < 0·001.
†Total participants not equal to 122 as non-randomised intervention arm excluded.
‡Change in BMI reported as authors did not report change in weight.
None of the interventions targeting providers’ behaviour resulted in statistically significant weight loss in their patients. Three studies targeting patients(Reference Martin, Rhode and Dutton22, Reference Martin, Dutton and Rhode23, Reference Munsch, Biedert and Keller25, Reference Bolognesi, Nigg and Massarini34) reported a statistically significant difference in amount of weight loss between the intervention and control group at end of intervention, with Martin et al. reporting significant weight loss at 6 months(Reference Martin, Rhode and Dutton22), but no significant weight loss at 9 or 12 months(Reference Martin, Dutton and Rhode23). None of the studies reported that clinically significant weight loss was achieved.
Nine studies reported on the effectiveness of lifestyle counselling delivered by non-primary-care physicians, with support from primary-care physicians (see Table 3). The personnel delivering the intervention were allied health-care providers (including nurses or dietitians) or non-health-care providers. The types of interventions included meal replacements(Reference Ashley, St Jeor and Schrage26), nurse- or dietitian-delivered counselling(Reference Willaing, Ladelund and Jorgensen27, Reference ter Bogt, Bemelmans and Beltman36, Reference Pritchard, Hyndman and Taba37), weight-loss websites(Reference Appel, Clark and Yeh30, Reference Jebb, Ahern and Olson32) and counselling delivered by non-medical health coaches(Reference Wadden, Volger and Sarwer28, Reference Jebb, Ahern and Olson32, Reference Logue, Sutton and Jarjoura33, Reference Tsai, Wadden and Rogers38). Two studies compared two interventions, without a control group(Reference Ashley, St Jeor and Schrage26, Reference Willaing, Ladelund and Jorgensen27). Four of the studies were high intensity, three were moderate intensity and one was low intensity. The number of sessions delivered during the first 3 months of the intervention was unclear in one study(Reference Willaing, Ladelund and Jorgensen27), thus intensity could not be determined.
Table 3 Weight-loss interventions in primary-care patients delivered by non-primary-care physicians
%F, percentage of females; Y, yes; N, no; DNR, did not report; HT, hypertension; T2DM, type 2 diabetes mellitus; n.s., not specified; DysL, dyslipidaemia; HC, hypercholesterolaemia; IGT, impaired glucose tolerance; IFG, impaired fasting glucose; PCOS, polycystic ovary syndrome; OA, osteoarthritis; MetS, metabolic syndrome; IG, intervention group; CG, control group; GP, general practitioner(s); RD, registered dietitian(s); SOC, stages of change; NP, nurse practitioner(s); PCP, primary-care physician(s); MA, medical assistant(s); DPP, Diabetes Prevention Program; ↑, weight gain; ↓, weight loss.
Comparison between IG and CG: *P < 0·05, **P < 0·01, ***P < 0·001.
†Total participants not equal to 390 as intervention arm which included use of pharmacology excluded.
One of the two comparative effectiveness studies reported statistically significant findings. Ashley et al. found that the use of meal replacements in addition to lifestyle counselling by a dietitian produced greater weight loss than the other two interventions tested (dietitian counselling alone or counselling from primary-care physician and nurse practitioner plus meal replacements)(Reference Ashley, St Jeor and Schrage26). In Willaing et al.'s study, which compared nutrition counselling delivered by a dietitian with that delivered by a primary-care physician(Reference Willaing, Ladelund and Jorgensen27), participants in both groups lost significantly more weight from baseline; however, there were no differences in weight loss between the two intervention groups. Six studies compared the intervention with either a usual care or minimal care group(Reference Appel, Clark and Yeh30, Reference Jebb, Ahern and Olson32, Reference Logue, Sutton and Jarjoura33, Reference ter Bogt, Bemelmans and Beltman36–Reference Tsai, Wadden and Rogers38). Of them, four reported statistically significant results between usual care and intervention groups and that clinically significant weight loss was achieved(Reference Appel, Clark and Yeh30, Reference Jebb, Ahern and Olson32, Reference Pritchard, Hyndman and Taba37, Reference Tsai, Wadden and Rogers38). Three of the four effective studies involved weight-loss coaches delivering high-intensity behavioural counselling, with participants self-monitoring their dietary intake, physical activity and weight change.
Only one study(Reference Ely, Banitt and Befort35) examined a multi-component intervention involving a chronic care model (including electronic registry, decision support and patient self-management support; see Table 4). The intervention was high intensity with a health counsellor who utilised motivational interviewing techniques. The study found that statistically significant weight loss was achieved compared with the usual care group; however, this was not clinically significant.
Table 4 Multi-component weight-loss intervention in primary-care patients
%F, percentage of females; Y, yes; N, no; DNR, did not report; n.s., not specified; IG, intervention group; CG, control group; ↓, weight loss.
To convert lb to kg, multiply lb by 0·4534.
Comparison between IG and CG: **P < 0·01.
Discussion
The present review identified sixteen different intervention studies that met the specified inclusion criteria. The low number of studies identified is similar to a review conducted by Tsai and Wadden, where only ten studies targeting obesity in US primary-care settings were identified(Reference Tsai and Wadden17). In contrast to the review conducted by Tsai and Wadden, the current review included studies conducted outside the USA and excluded studies where pharmacological treatments were used. Due to the recent withdrawal of the weight-loss drug sibutramine from the market(Reference Sayburn39), a number of studies included in Tsai and Wadden's review may no longer be relevant to practitioners. The removal of sibutramine in 2010 has resulted in orlistat being the only weight-loss medication available for practitioners located in Europe(Reference Sayburn39). While other options exist for practitioners located in the USA, the overall limited availability and safety of weight-loss medications makes identifying effective behavioural interventions targeting excess weight an issue of critical importance. Given the high burden of illness associated with excess weight and the increasing discussion surrounding the use of primary care for weight management, the amount of research conducted is insufficient to inform practice.
Overall, the studies were of moderate to good quality. One study met all EPOC quality criteria(Reference Moore40). Two criteria which were poorly met across studies were selective outcome reporting and adequately protecting against contamination.
Only four studies included in the present review had published a study protocol(Reference Wadden, Volger and Sarwer28–Reference Appel, Clark and Yeh30, Reference Jebb, Ahern and Olson32). Selectively reporting positive or statistically significant findings can lead to overestimation of treatment effects, subsequently affecting conclusions drawn from systematic reviews and meta analyses(Reference Dwan, Altman and Cresswell41). Dwan et al. reported that discrepancies between protocol or trial registries and publications occur in a large proportion of studies, where at least one primary outcome was changed, introduced or omitted in 4–50 % of trial reports(Reference Dwan, Altman and Cresswell41). Where a protocol does not exist, it is unknown whether selective outcome reporting occurred. Therefore, for a large number of studies included in the present review, the criterion related to selective outcome assessment could not be adequately assessed.
All studies except two(Reference Schuster, Tasosa and Terwoord24, Reference Moore, Summerbell and Greenwood29) used patients or physicians within the same practices as the unit of randomisation, thus increasing the likelihood of contamination between experimental and control groups. Contamination may reduce the effect size of the intervention due to the unintentional provision of additional care to control groups(42). In order to improve the validity of findings, strategies need to be in place to ensure that the control group is not exposed to components of the intervention.
Selective outcome reporting and potential contamination may have affected findings from the included studies. Furthermore, poor reporting of study methodology in some studies made it difficult to assess study quality. These methodological and reporting shortcomings have been similarly reported in other reviews on weight loss(Reference Norris, Zhang and Avenell12, Reference McTigue, Harris and Hemphill16, Reference Tsai and Wadden17, Reference Harvey, Glenny and Kirk43).
Of studies examining lifestyle counselling delivered by primary-care physicians, interventions that produced statistically significant weight loss included the use of a structured and tailored protocol to assist physicians with delivery of weight-loss counselling(Reference Martin, Rhode and Dutton22, Reference Martin, Dutton and Rhode23, Reference Munsch, Biedert and Keller25, Reference Bolognesi, Nigg and Massarini34). Consistent with current evidence(Reference McTigue, Harris and Hemphill16), regular contact between patients and physicians was a key component in producing weight loss, with higher-intensity interventions reporting larger amounts of weight loss. This contact may not need to be one on one; one study reported that group counselling sessions were effective in producing significant weight loss(Reference Munsch, Biedert and Keller25). While one of the effective interventions(Reference Bolognesi, Nigg and Massarini34) was low intensity (one-off contact with physician), the amount of BMI change reported at 5–6 months’ follow-up was marginal. The authors reported that highly motivated patients were enrolled in the intervention group with a large proportion of patients being in the contemplation and preparation stages of change, and may not have been reflective of usual primary-care patients(Reference Bolognesi, Nigg and Massarini34).
The two studies targeting providers did not report achieving any significant weight loss in their patients. Of the two, one was a high-intensity intervention(Reference Moore, Summerbell and Greenwood29). Although classified as high intensity, the intervention relied on practitioners’ delivery of the proposed weight-loss model (this entailed that practitioners saw their patients about once every fortnight until they had lost 10 % of their initial body weight). The authors noted that practitioners’ adherence to the intervention protocol was low, thus intensity could not be accurately estimated. Provider-targeted interventions for weight loss have been discussed in detail in other reviews(Reference Harvey, Glenny and Kirk43–Reference Flodgren, Deane and Dickinson45).
While a structured protocol to assist practitioners with delivery of weight-loss counselling appeared effective in producing some weight loss in overweight or obese patients, none of the interventions reported achieving clinically significant weight loss, making it questionable whether physician-delivered interventions alone are worth implementing in primary care.
In studies where non-physicians delivered the intervention, lifestyle counselling was conducted by allied health-care providers (nurses, dietitians) or non-health-care providers (weight-loss counsellors, medical assistants).
Two studies included a web-based component in addition to intensity lifestyle counselling(Reference Appel, Clark and Yeh30, Reference Jebb, Ahern and Olson32). Of these two, one used the web-based component in combination with referral to a community-based weight-loss programme (WeightWatchers®)(Reference Tsai, Wadden and Rogers38) and the other with in-person or telephone support from weight-loss coaches(Reference Jebb, Ahern and Olson32). Both studies utilised similar high-intensity interventions, with regular contact with health coaches or group leaders and Internet-based systems to help with self-monitoring and provide peer support. For both studies, participants in the intervention group lost significantly more weight than the control group (mean weight loss of approximately 6·0 kg). Appel et al. reported no significant difference in amount of weight loss between face-to-face and telephone support, suggesting there is potential for telephone counselling to be delivered as part of weight-reduction programmes to minimise intervention cost(Reference Appel, Clark and Yeh30).
Findings from studies where non-health-care providers delivered weight-management counselling were mixed. Tsai et al.'s high-intensity intervention reported that significantly more weight loss was achieved in the intervention group compared with the control group(Reference Tsai, Wadden and Rogers38), whereas studies by Wadden et al.(Reference Wadden, Volger and Sarwer28) and Logue et al.(Reference Logue, Sutton and Jarjoura33) reported no significant difference in amount of weight loss between the intervention and control groups. Notably, the latter study compared the intervention with an ‘augmented usual care group’, where participants in the control group met with a dietitian for 10 min biannually(Reference Logue, Sutton and Jarjoura33). This could have affected the control group's behaviour, thus making it harder to demonstrate an intervention effect.
These findings tentatively suggest that high-intensity interventions delivered by non-health-care providers in adjunct to primary-care physician consult are effective in producing clinically significant weight loss.
In studies involving allied health-care providers, the way in which weight-loss counselling was conducted varied depending on the personnel delivering the intervention. Where the dietitian was involved, delivery of the intervention largely relied on the dietitian to provide individualised advice and weight-loss strategies(Reference Pritchard, Hyndman and Taba37). In contrast, nurse practitioners used a structured software program to assist with delivery of weight-loss counselling(Reference ter Bogt, Bemelmans and Beltman36). Both Pritchard et al.(Reference Pritchard, Hyndman and Taba37) and ter Bogt et al.(Reference ter Bogt, Bemelmans and Beltman36) reported significantly more weight loss in the intervention group than the control group; however, only the Pritchard study involving dietitian-delivered advice reported that clinically significant weight loss was achieved. Pritchard et al.'s study highlighted the advantage of physician involvement in addition to dietitian-delivered care in increasing retention rate and proportion attending all sessions of the intervention(Reference Pritchard, Hyndman and Taba37).
Other studies confirmed the effectiveness of dietitian-delivered interventions. Ashley et al. compared three interventions and found that dietitian-delivered advice coupled with meal replacements was effective in producing clinically significant weight loss compared with either receiving dietitian advice alone or using meal replacements coupled with primary-care physician and nurse practitioner counselling(Reference Ashley, St Jeor and Schrage26). Analysis was conducted only on participants who completed the intervention. Therefore, treatment effect may have been overestimated. Despite this limitation, the study suggests that the use of meal replacements in conjunction with dietitian advice is useful in producing significant weight loss. Willaing et al. found no difference in the effectiveness of dietary counselling delivered by a primary-care physician compared with dietary counselling delivered by a dietitian(Reference Willaing, Ladelund and Jorgensen27). Both groups had significant weight loss from baseline at 12 months, despite the primary-care physician spending less time during consultations than the dietitian.
Regardless of level of intervention intensity, dietitian-delivered counselling was effective in producing weight loss ranging from 3 to 6 kg. Dietitians receive specialist training in nutrition assessment and counselling for weight loss and may therefore be more equipped to provide weight-management advice.
Findings from these studies suggest that high-intensity interventions involving non-physicians, with primary-care physicians playing a supportive role of assessment and referral, may be more effective than advice delivered by primary-care physicians alone in producing significant weight loss in overweight and obese primary-care patients. Comparisons made here, however, are limited by differences in intensity of intervention, with most primary-care physician-delivered interventions being of low to moderate intensity and non-primary-care physician-delivered interventions being of moderate to high intensity. These differences are likely to reflect clinical practice as primary-care physicians often face the need to deal with more acute issues and have less time to spend on delivery of lifestyle advice. The involvement of dietitians, non-health professionals or commercial weight-loss programmes enables intensive targeted counselling specifically dealing with weight management to be delivered to patients.
One study examined the use of a multi-component intervention which included an electronic registry, decision support and motivational interviewing delivered via telephone by a Master's level weight-loss advisor(Reference Ely, Banitt and Befort35). That study reported no statistically significant weight loss between the intervention and control group. The small sample size (n 101), short follow-up length and high drop-out rate made it difficult for any conclusions to be drawn.
Practice implications
Findings reported here suggest that intensive interventions delivered by non-physician personnel in the primary-care setting are effective in achieving clinically significant weight loss. There is insufficient evidence to suggest that counselling delivered by primary-care physicians alone produces clinically significant reductions in weight. However, involvement of primary-care physicians appears to increase retention rates and uptake of interventions delivered by non-physicians(Reference Pritchard, Hyndman and Taba37). Approaches where non-physician providers play a more intensive role in delivery of behavioural interventions, accompanied by regular monitoring from primary-care physicians, could be a promising strategy to reduce obesity in primary-care patients. Given this finding, a review focused on assessing interventions solely delivered by non-primary-care physicians should be conducted to further inform weight management in this setting. The use of web-based interventions and meal replacements in adjunct with behavioural counselling (delivered by trained non-health providers or commercial centre weight-loss staff) appears promising. Additionally, delivery of interventions by dietitians appears effective regardless of intensity. With only few methodologically rigorous studies conducted, more studies evaluating the effectiveness of these interventions are needed. Future studies should also attempt to evaluate the acceptability, preference and uptake of these strategies among overweight and obese primary-care patients.
Limitations
The search terms used may not have identified all relevant studies. However, given the number of records extracted and the small proportion of relevant articles, it is likely that the majority of relevant articles were identified. The chance of missing relevant studies was further reduced by hand searching reference lists of relevant articles. Studies that examined behavioural interventions delivered in conjunction with medication were not examined as it was beyond the scope of the review.
Conclusions
Overall, the few studies identified and heterogeneity of interventions utilised made it difficult for conclusions to be drawn regarding what interventions are most effective in producing weight loss in overweight or obese primary-care patients. Given the burden of excess weight on the population and the advantage of using primary care to target weight loss, there is a need for more research exploring the use of this setting for delivery of weight-loss interventions. Results suggest that counselling delivered by non-physicians (face to face or telephone) with support from primary-care physicians is effective in producing weight loss. More studies assessing the effectiveness of these types of interventions are needed to confirm this.
Acknowledgements
Sources of funding: This work was supported by a grant from the beyondblue and National Heart Foundation of Australia Strategic Research Partnership Grant (G 08S 4042). M.C. is supported by Hunter Medical Research Institute Post-Doctoral Fellowships. Conflicts of interest: The authors declare no conflict of interest. Authors’ contribution: S.L.Y. was responsible for the overall design of the study, conducted the literature search, extracted relevant data, interpreted findings and drafted the initial version of the article. M.C. played a major role in design of the study, interpretation of findings and revision of the article. R.S.-F. designed the study and provided critical comments regarding the intellectual content of the article. A.G. assisted in extraction of relevant data and revision of the draft manuscript. Acknowledgements: The authors wish to thank Shiho Rose for her assistance with screening for relevant articles.
