Patient Education and Counseling 98 (2015) 412–419
Contents lists available at ScienceDirect
Patient Education and Counseling journal homepage: www.elsevier.com/locate/pateducou
Review
Factors contributing to the effectiveness of physical activity counselling in primary care: A realist systematic review Anna R. Gagliardi a,*, Guy Faulkner b, Donna Ciliska c, Audrey Hicks c a
University Health Network, 200 Elizabeth Street, Toronto, Canada University of Toronto, Canada c McMaster University, Canada b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 29 January 2014 Received in revised form 17 November 2014 Accepted 23 November 2014
Objective: Physical activity (PA) counselling in primary care increases PA but is not consistently practiced. This study examined factors that optimise the delivery and impact of PA counselling. Methods: A realist systematic review based on the PRECEDE–PROCEED model and RAMESES principles was conducted to identify essential components of PA counselling. MEDLINE, EMBASE, Cochrane Library, PsycINFO, and Physical Education Index were searched from 2000 to 2013 for studies that evaluated family practice PA counselling. Results: Of 1546 articles identified, 10 were eligible for review (3 systematic reviews, 5 randomised controlled trials, 2 observational studies). Counselling provided by clinicians or counsellors alone that explored motivation increased self-reported PA at least 12 months following intervention. Multiple sessions may sustain increased PA beyond 12 months. Conclusion: Given the paucity of eligible studies and limited detail reported about interventions, further research is needed to establish the optimal design and delivery of PA counselling. Research and planning should consider predisposing, reinforcing and enabling design features identified in these studies. Practice implications: Since research shows that PA counselling promotes PA but is not widely practiced, primary care providers will require training and tools to operationalize PA counselling. ß 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Physical activity Health promotion Primary care
Contents 1. 2.
3. 4.
Introduction . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . Approach . . . . . . . . . . . . . . . . . 2.1. 2.2. Scoping the literature . . . . . . . Search strategy . . . . . . . . . . . . 2.3. Screening process and criteria 2.4. Data collection and analysis . . 2.5. Results . . . . . . . . . . . . . . . . . . . . . . . . Discussion and conclusion. . . . . . . . . Discussion . . . . . . . . . . . . . . . . 4.1. 4.2. Conclusion. . . . . . . . . . . . . . . . Practice implications. . . . . . . . 4.3. References . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
* Corresponding author. Tel.: +1 416 340 4800. E-mail address: [email protected] (A.R. Gagliardi). http://dx.doi.org/10.1016/j.pec.2014.11.020 0738-3991/ß 2014 Elsevier Ireland Ltd. All rights reserved.
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
413 413 413 413 413 413 414 414 415 415 418 418 418
A.R. Gagliardi et al. / Patient Education and Counseling 98 (2015) 412–419
1. Introduction
2. Methods
Physical activity (PA) reduces the risk of chronic disease, premature mortality, and health system costs [1–6]. PA guidelines have been developed worldwide and in Canada, where PA is further promoted with Physical Activity Guides for all ages, the Children’s Fitness Tax Credit, and social marketing organizations such as ParticipACTION [7,8]. However, fitness levels have declined, while obesity and premature mortality have increased in Canada, Europe and the United States over the last 20 years [9–12]. The World Health Organization reports that physical inactivity is the fourth leading risk factor for global mortality causing an estimated 3.2 million deaths globally [1]. Therefore, it is imperative that additional strategies beyond issuing of guidelines and raising public awareness are identified to encourage PA. The primary care setting offers a convenient platform in which to promote PA. The majority of Canadians (77%) and Americans (80%) have at least one contact with a primary care physician annually during which discussions about health habits including PA are recommended [13,14]. Considerable research has shown that counselling more than any other intervention is an effective strategy for PA promotion in primary care [15]. PA counselling refers to advice and discussion about PA among primary care providers and individual patients. PA counselling in the primary care setting appears to be an efficient and effective means of increasing PA [16]. However, surveys of primary care physicians and/or patients report low rates of PA counselling worldwide [17–22]. Systematic reviews revealed numerous barriers to PA counselling in primary care including lack of time, knowledge, training, materials for learning, education and information, protocols or other system support, resources, incentives and reimbursement [23]. In particular, providers are uncertain about the effectiveness of counselling and uncomfortable providing detailed advice about PA [24]. PA counselling in primary care may have greater impact if providers were able to consistently counsel patients about PA. Systematic reviews already conducted did not examine factors that may have influenced the success of counselling such as the content or delivery of counselling, or of accompanying information or tools. Therefore further research is needed to understand how the design and impact of PA counselling could be optimised. Such information could be used to develop interventions that support PA counselling by primary care providers. A commonly used model for designing or evaluating health promotion interventions is the PRECEDE–PROCEED model [25]. PRECEDE refers to micro and meso level factors that may influence how an intervention works. Predisposing factors are those that a patient brings to the primary care setting (knowledge, attitudes, beliefs, values, age, health status). Reinforcing factors are those that primary care providers bring to the patient consultation (values, personal physical activity habits, health promotion and counselling practice). Enabling factors may include the availability of resources, protocols and service structures. PROCEED refers to policy or regulatory constructs, and involves implementing and evaluating the impact of selected interventions. The PRECEDE constructs provide a framework for understanding how various factors influence PA counselling and its impact. The primary purpose of this study was to identify the predisposing, reinforcing and enabling factors that optimise the effectiveness of PA counselling. The findings may provide insight on how to tailor the design, delivery and impact of counselling for PA promotion. That information could be used by policy-makers, educators, professional societies or primary care teams to develop and apply interventions or tools that support counselling for PA promotion in the primary care setting, ultimately leading to improved PA and associated improvements in the physical and mental health of patients.
2.1. Approach
413
A realist systematic review was conducted [26]. This approach is specifically used to describe the theoretical or contextual factors that contribute to the effectiveness of behavioural interventions. It is similar in rigor to a traditional systematic review but draws on a range of study designs to examine the interaction between context, intervention, outcome and underlying theory. RAMESES criteria guided the conduct and reporting of the review [27]. Data were publicly available so institutional review board approval was not necessary. 2.2. Scoping the literature To plan for the full-scale review a preliminary scan of relevant literature was undertaken. This refined the scope of the review and contributed to the development of screening criteria. Primary care was defined as office-based settings in which patients see family physicians or teams. This was distinguished from primary health care in which primary care services may be delivered to patients in a variety of other settings. The intervention of interest was counselling-alone, or counselling plus information or tools that were offered to patients in the office-based setting. This included print information, patient tools such as pedometers or diaries, or written prescription including the type, frequency and intensity of recommended exercise. This excluded exercise referral which was not found to be effective [16], and follow-up communication by telephone call or mailed information which would require time and resources beyond the counselling session. Counselling was initially broadly defined as one or more office-based instances of brief or more detailed advice or education provided to patients by one or more members of the primary care team including family physicians, nurses, physiotherapists, or exercise specialists either concurrently or consecutively. The outcome of interest was PA, either self-reported or objectively assessed through pedometer or other mechanism, or associated physiological functions such as blood pressure, body weight, serum lipid levels, glycaemic control or physical fitness (VO2max). Few studies were identified that evaluated PA promotion among children or youth, or older adults, so this study focused on adults aged 18–64 years. This study focused on PA counselling rather than lifestyle counselling, which also addresses diet and other health behaviours such as smoking, among relatively well individuals as a preventive health behaviour. 2.3. Search strategy MEDLINE, EMBASE, Cochrane Library, PsycINFO, and Physical Education Index were searched in April 2011 for the period of 2000 to 2013 inclusive. The search strategy employed was purposefully broad because the scoping exercise revealed that counselling was not consistently used to index relevant studies. MEDLINE, EMBASE and Cochrane Library search terms included [(exercise or exercise therapy or physical activity) and primary health care]. PsycINFO search terms were [(physical activity or physical fitness or exercise) and primary health care]. Physical Education Index search terms were [(physical activity or exercise) and (primary care or primary health care]. Searches were limited to English language systematic reviews, randomised controlled trials or observational cohort studies. 2.4. Screening process and criteria One research assistant and the principal investigator independently screened titles and abstracts. Studies were eligible if they
414
A.R. Gagliardi et al. / Patient Education and Counseling 98 (2015) 412–419
evaluated the impact of counselling-alone or combined with information, prescription or tools at the counselling session(s) by one or more members of the primary care team in family practice office settings on self-reported or objectively assessed PA or associated physiological functions among adults. Studies that focused on PA for secondary prevention; assessed the physiological effectiveness of different types of PA; involved special populations; largely evaluated referral or other interventions not part of the counselling visit; or were in the form of abstracts, letters or editorials were not eligible. All studies selected by at least one individual were retrieved since ultimate judgement must often be reserved until the full text can be examined. Systematic reviews were excluded if not mostly focused on PA counselling or family practice office settings. Individual studies included in eligible systematic reviews were excluded. The references of eligible studies were scanned but no additional studies were identified. 2.5. Data collection and analysis A data extraction form was developed to collect information on study design, number and type of participants, details about the design and delivery of the intervention, number and type of health professionals delivering the intervention, how outcome was assessed and impact, and other contextual issues if assessed. One research assistant and the principal investigator pilot tested the form on three articles through four iterations until data extraction was consistent. One research assistant extracted all data. Extracted data was confirmed independently by a second research assistant. Data could not be pooled due to heterogeneity in study design and interventions. The methodological quality of eligible studies was assessed with AMSTAR for systematic reviews, Cochrane Collaboration Risk of Bias tool for randomised controlled trials (RCTs), and a modified Downs and Black Quality Assessment Tool for observational studies [28–30]. Extracted data was further summarized to categorize details about intervention design and impact according to predisposing (patient knowledge, attitudes, beliefs), reinforcing (primary care professional knowledge, attitudes, beliefs, practices) and enabling factors (resources, protocols, service structures) if reported, and identify trends in these contextual details and intervention impact. 3. Results A total of 1546 articles were identified by search strategies across all sources of which 624 were duplicates leaving 922 items to be screened. Of these, 912 were excluded resulting in 10 studies eligible for review (Fig. 1). Three systematic reviews were eligible, of which two were judged as high and one as medium quality (Table 1). Counselling was brief in two studies that reported session length. Self-reported PA increased at 12 months in two studies [31,32] and at 6 but not 12 months in the third [33]. Orrow et al. [31] reviewed 15 RCTs of which 10 involved primary care counselling. Design of counselling including who delivered it and supplementary strategies or tools varied across studies. Multiple counselling sessions were offered in five studies, and motivational interviewing was used in seven studies. Tulloch et al. [32] reviewed 19 studies of which 15 were RCTs to assess which member of the primary care team should deliver counselling. Self-reported PA increased at both 6 and 12 months when delivered by physicians, allied health professionals or health educators only, or some combination of these. However, this finding was not consistent across all eligible studies, and interpretation was confounded by the fact that counselling was sometimes based on an assessment of patient motivation, and sometimes accompanied by information or PA prescription. Lawlor
Sources Searched (1,546) MEDLINE (419) EMBASE (531) Cochrane Library (91) PsycINFO (113) Physical Education Index (392) Duplicates removed (624) Screened (922)
Excluded after screening (n=912) Secondary prevention (288) Physiological effectiveness of PA (189) Publication type/study design (239) Population studied (33) Setting (111) Did not examine/focus on PA counselling (53)
Eligible for review (10) Fig. 1. Document flow diagram.
[33] reviewed eight studies of which two were RCTs and found that self-reported PA increased in four non-RCTs at less than 8 weeks follow-up. This was true of counselling-alone in one study, and counselling accompanied by information or prescription. Five RCTs were eligible, of which one was judged as high, two as unclear, and two as low risk of bias (Table 2). All involved physician counselling in a single session plus follow-up counselling by a physician or PA specialist, information, tools (log book, pedometer) and/or PA prescription. Length of counselling sessions was not specified. Self-reported PA increased at 3 months based on multiple sessions compared with a brief, single session and PA prescription [34]. In another study self-reported PA increased and was sustained at 24 months among those who received follow-up counselling sessions after a brief, single session compared with usual care [35]. Self-reported PA increased in one study based on counselling and prescription compared with patient self-monitoring with a pedometer and diary [36]. In another study self-reported PA increased but did not differ at 14 months between those who received counselling only and counselling plus information plus an offer of a subsidized session with a PA specialist [37]. Counselling plus both information and PA prescription resulted in greater selfreported PA at 6 months compared with those who received usual care in another study [38]. Two observational studies were eligible and judged as high quality (Table 3). In one study, a single counselling session by a trained lifestyle change facilitator with up to 12 follow-up counselling sessions at the patient’s discretion increased PA at 6 months, particularly among those who participated in more counselling sessions [39]. In a second observational study, a single physician counselling session of unspecified length was based on assessment of patient motivation, and supplemented with information and the option for additional counselling [40]. At 12 months, 37.2% self-reported increased PA but many said they did not use the leaflet or additional counselling. A key finding of this systematic review is that few studies have focused on evaluating PA promotion in family practice settings. Among the systematic reviews, PA counselling design and delivery were mixed so it is difficult to draw conclusions. In an attempt to elucidate the contextual factors that influenced counselling success, intervention details reported in RCTs and observational studies were categorized according to PRECEDE constructs of the PRECEDE–PROCEED model [25]. All but one of the studies assessed patient readiness for change but it was not clear how this was used
A.R. Gagliardi et al. / Patient Education and Counseling 98 (2015) 412–419
415
Table 1 Systematic reviews evaluating PA counselling. Study
Studies
Intervention design
Impact
Quality
Orrow 2012 UK [31]
10/15 studies on PA counselling in primary care (all RCTs) n = 5991 Up to 2010
Results reflect pooling of data across all 15 studies. PA promotion resulted in small to medium improvement in self-reported PA at 12 months: odds ratio 1.45 (95% CI 1.17–1.73), standardized mean difference 0.25 (0.11–0.38). Nature, timing and who delivered the intervention varied widely across studies and were not examined. No sub-group analyses for patient characteristics.
High
Tulloch 2006 Canada [32]
19 studies (15 RCTs) n = 9853 Up to 2003
Counselling by physician alone (1 study); physician plus exercise, PA or health specialist (2 studies); physician or nurse plus exercise, PA or health specialist (3 studies); nurse plus exercise, PA or health specialist (1 study); health or exercise specialist alone (2 studies); or physiotherapist alone (1 study). Motivational interviewing to assess readiness and barriers was used in seven studies. Counselling sometimes accompanied by educational material (5 studies), referral (3 studies), prescription (5 studies), or follow-up visit or phone calls (5 studies). Single counselling sessions were assessed in five studies and multiple counselling sessions in five studies. Length of counselling sessions not reported. Counselling by physician only (7 studies), physician plus allied health/educator (5 studies), or allied health/educators (7 studies). Sometimes accompanied by information or prescription and/or tailored based on assessment of patient motivation or follow-up counselling sessions in office setting. Most counselling was brief (5 min or less) when offered by a physician or physician plus other provider. Counselling ranged from 20 to 40 min when offered only by a PA specialist in four studies.
High
Lawlor 2001 UK [33]
8 studies (2 RCTs) n = 4747 1993–1999
Short term evaluation (6 months) found an increase in self-reported PA (2/7 physician only, 2/5 combined, and 5/7 AH/E only studies). Nature, timing and who delivered the intervention varied widely across studies. No meta-analysis of findings because interventions and measures were mixed. No sub-group analyses for patient characteristics. In four of six studies (none were RCTS) that evaluated short term results (
Contents lists available at ScienceDirect
Patient Education and Counseling journal homepage: www.elsevier.com/locate/pateducou
Review
Factors contributing to the effectiveness of physical activity counselling in primary care: A realist systematic review Anna R. Gagliardi a,*, Guy Faulkner b, Donna Ciliska c, Audrey Hicks c a
University Health Network, 200 Elizabeth Street, Toronto, Canada University of Toronto, Canada c McMaster University, Canada b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 29 January 2014 Received in revised form 17 November 2014 Accepted 23 November 2014
Objective: Physical activity (PA) counselling in primary care increases PA but is not consistently practiced. This study examined factors that optimise the delivery and impact of PA counselling. Methods: A realist systematic review based on the PRECEDE–PROCEED model and RAMESES principles was conducted to identify essential components of PA counselling. MEDLINE, EMBASE, Cochrane Library, PsycINFO, and Physical Education Index were searched from 2000 to 2013 for studies that evaluated family practice PA counselling. Results: Of 1546 articles identified, 10 were eligible for review (3 systematic reviews, 5 randomised controlled trials, 2 observational studies). Counselling provided by clinicians or counsellors alone that explored motivation increased self-reported PA at least 12 months following intervention. Multiple sessions may sustain increased PA beyond 12 months. Conclusion: Given the paucity of eligible studies and limited detail reported about interventions, further research is needed to establish the optimal design and delivery of PA counselling. Research and planning should consider predisposing, reinforcing and enabling design features identified in these studies. Practice implications: Since research shows that PA counselling promotes PA but is not widely practiced, primary care providers will require training and tools to operationalize PA counselling. ß 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Physical activity Health promotion Primary care
Contents 1. 2.
3. 4.
Introduction . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . Approach . . . . . . . . . . . . . . . . . 2.1. 2.2. Scoping the literature . . . . . . . Search strategy . . . . . . . . . . . . 2.3. Screening process and criteria 2.4. Data collection and analysis . . 2.5. Results . . . . . . . . . . . . . . . . . . . . . . . . Discussion and conclusion. . . . . . . . . Discussion . . . . . . . . . . . . . . . . 4.1. 4.2. Conclusion. . . . . . . . . . . . . . . . Practice implications. . . . . . . . 4.3. References . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
* Corresponding author. Tel.: +1 416 340 4800. E-mail address: [email protected] (A.R. Gagliardi). http://dx.doi.org/10.1016/j.pec.2014.11.020 0738-3991/ß 2014 Elsevier Ireland Ltd. All rights reserved.
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
413 413 413 413 413 413 414 414 415 415 418 418 418
A.R. Gagliardi et al. / Patient Education and Counseling 98 (2015) 412–419
1. Introduction
2. Methods
Physical activity (PA) reduces the risk of chronic disease, premature mortality, and health system costs [1–6]. PA guidelines have been developed worldwide and in Canada, where PA is further promoted with Physical Activity Guides for all ages, the Children’s Fitness Tax Credit, and social marketing organizations such as ParticipACTION [7,8]. However, fitness levels have declined, while obesity and premature mortality have increased in Canada, Europe and the United States over the last 20 years [9–12]. The World Health Organization reports that physical inactivity is the fourth leading risk factor for global mortality causing an estimated 3.2 million deaths globally [1]. Therefore, it is imperative that additional strategies beyond issuing of guidelines and raising public awareness are identified to encourage PA. The primary care setting offers a convenient platform in which to promote PA. The majority of Canadians (77%) and Americans (80%) have at least one contact with a primary care physician annually during which discussions about health habits including PA are recommended [13,14]. Considerable research has shown that counselling more than any other intervention is an effective strategy for PA promotion in primary care [15]. PA counselling refers to advice and discussion about PA among primary care providers and individual patients. PA counselling in the primary care setting appears to be an efficient and effective means of increasing PA [16]. However, surveys of primary care physicians and/or patients report low rates of PA counselling worldwide [17–22]. Systematic reviews revealed numerous barriers to PA counselling in primary care including lack of time, knowledge, training, materials for learning, education and information, protocols or other system support, resources, incentives and reimbursement [23]. In particular, providers are uncertain about the effectiveness of counselling and uncomfortable providing detailed advice about PA [24]. PA counselling in primary care may have greater impact if providers were able to consistently counsel patients about PA. Systematic reviews already conducted did not examine factors that may have influenced the success of counselling such as the content or delivery of counselling, or of accompanying information or tools. Therefore further research is needed to understand how the design and impact of PA counselling could be optimised. Such information could be used to develop interventions that support PA counselling by primary care providers. A commonly used model for designing or evaluating health promotion interventions is the PRECEDE–PROCEED model [25]. PRECEDE refers to micro and meso level factors that may influence how an intervention works. Predisposing factors are those that a patient brings to the primary care setting (knowledge, attitudes, beliefs, values, age, health status). Reinforcing factors are those that primary care providers bring to the patient consultation (values, personal physical activity habits, health promotion and counselling practice). Enabling factors may include the availability of resources, protocols and service structures. PROCEED refers to policy or regulatory constructs, and involves implementing and evaluating the impact of selected interventions. The PRECEDE constructs provide a framework for understanding how various factors influence PA counselling and its impact. The primary purpose of this study was to identify the predisposing, reinforcing and enabling factors that optimise the effectiveness of PA counselling. The findings may provide insight on how to tailor the design, delivery and impact of counselling for PA promotion. That information could be used by policy-makers, educators, professional societies or primary care teams to develop and apply interventions or tools that support counselling for PA promotion in the primary care setting, ultimately leading to improved PA and associated improvements in the physical and mental health of patients.
2.1. Approach
413
A realist systematic review was conducted [26]. This approach is specifically used to describe the theoretical or contextual factors that contribute to the effectiveness of behavioural interventions. It is similar in rigor to a traditional systematic review but draws on a range of study designs to examine the interaction between context, intervention, outcome and underlying theory. RAMESES criteria guided the conduct and reporting of the review [27]. Data were publicly available so institutional review board approval was not necessary. 2.2. Scoping the literature To plan for the full-scale review a preliminary scan of relevant literature was undertaken. This refined the scope of the review and contributed to the development of screening criteria. Primary care was defined as office-based settings in which patients see family physicians or teams. This was distinguished from primary health care in which primary care services may be delivered to patients in a variety of other settings. The intervention of interest was counselling-alone, or counselling plus information or tools that were offered to patients in the office-based setting. This included print information, patient tools such as pedometers or diaries, or written prescription including the type, frequency and intensity of recommended exercise. This excluded exercise referral which was not found to be effective [16], and follow-up communication by telephone call or mailed information which would require time and resources beyond the counselling session. Counselling was initially broadly defined as one or more office-based instances of brief or more detailed advice or education provided to patients by one or more members of the primary care team including family physicians, nurses, physiotherapists, or exercise specialists either concurrently or consecutively. The outcome of interest was PA, either self-reported or objectively assessed through pedometer or other mechanism, or associated physiological functions such as blood pressure, body weight, serum lipid levels, glycaemic control or physical fitness (VO2max). Few studies were identified that evaluated PA promotion among children or youth, or older adults, so this study focused on adults aged 18–64 years. This study focused on PA counselling rather than lifestyle counselling, which also addresses diet and other health behaviours such as smoking, among relatively well individuals as a preventive health behaviour. 2.3. Search strategy MEDLINE, EMBASE, Cochrane Library, PsycINFO, and Physical Education Index were searched in April 2011 for the period of 2000 to 2013 inclusive. The search strategy employed was purposefully broad because the scoping exercise revealed that counselling was not consistently used to index relevant studies. MEDLINE, EMBASE and Cochrane Library search terms included [(exercise or exercise therapy or physical activity) and primary health care]. PsycINFO search terms were [(physical activity or physical fitness or exercise) and primary health care]. Physical Education Index search terms were [(physical activity or exercise) and (primary care or primary health care]. Searches were limited to English language systematic reviews, randomised controlled trials or observational cohort studies. 2.4. Screening process and criteria One research assistant and the principal investigator independently screened titles and abstracts. Studies were eligible if they
414
A.R. Gagliardi et al. / Patient Education and Counseling 98 (2015) 412–419
evaluated the impact of counselling-alone or combined with information, prescription or tools at the counselling session(s) by one or more members of the primary care team in family practice office settings on self-reported or objectively assessed PA or associated physiological functions among adults. Studies that focused on PA for secondary prevention; assessed the physiological effectiveness of different types of PA; involved special populations; largely evaluated referral or other interventions not part of the counselling visit; or were in the form of abstracts, letters or editorials were not eligible. All studies selected by at least one individual were retrieved since ultimate judgement must often be reserved until the full text can be examined. Systematic reviews were excluded if not mostly focused on PA counselling or family practice office settings. Individual studies included in eligible systematic reviews were excluded. The references of eligible studies were scanned but no additional studies were identified. 2.5. Data collection and analysis A data extraction form was developed to collect information on study design, number and type of participants, details about the design and delivery of the intervention, number and type of health professionals delivering the intervention, how outcome was assessed and impact, and other contextual issues if assessed. One research assistant and the principal investigator pilot tested the form on three articles through four iterations until data extraction was consistent. One research assistant extracted all data. Extracted data was confirmed independently by a second research assistant. Data could not be pooled due to heterogeneity in study design and interventions. The methodological quality of eligible studies was assessed with AMSTAR for systematic reviews, Cochrane Collaboration Risk of Bias tool for randomised controlled trials (RCTs), and a modified Downs and Black Quality Assessment Tool for observational studies [28–30]. Extracted data was further summarized to categorize details about intervention design and impact according to predisposing (patient knowledge, attitudes, beliefs), reinforcing (primary care professional knowledge, attitudes, beliefs, practices) and enabling factors (resources, protocols, service structures) if reported, and identify trends in these contextual details and intervention impact. 3. Results A total of 1546 articles were identified by search strategies across all sources of which 624 were duplicates leaving 922 items to be screened. Of these, 912 were excluded resulting in 10 studies eligible for review (Fig. 1). Three systematic reviews were eligible, of which two were judged as high and one as medium quality (Table 1). Counselling was brief in two studies that reported session length. Self-reported PA increased at 12 months in two studies [31,32] and at 6 but not 12 months in the third [33]. Orrow et al. [31] reviewed 15 RCTs of which 10 involved primary care counselling. Design of counselling including who delivered it and supplementary strategies or tools varied across studies. Multiple counselling sessions were offered in five studies, and motivational interviewing was used in seven studies. Tulloch et al. [32] reviewed 19 studies of which 15 were RCTs to assess which member of the primary care team should deliver counselling. Self-reported PA increased at both 6 and 12 months when delivered by physicians, allied health professionals or health educators only, or some combination of these. However, this finding was not consistent across all eligible studies, and interpretation was confounded by the fact that counselling was sometimes based on an assessment of patient motivation, and sometimes accompanied by information or PA prescription. Lawlor
Sources Searched (1,546) MEDLINE (419) EMBASE (531) Cochrane Library (91) PsycINFO (113) Physical Education Index (392) Duplicates removed (624) Screened (922)
Excluded after screening (n=912) Secondary prevention (288) Physiological effectiveness of PA (189) Publication type/study design (239) Population studied (33) Setting (111) Did not examine/focus on PA counselling (53)
Eligible for review (10) Fig. 1. Document flow diagram.
[33] reviewed eight studies of which two were RCTs and found that self-reported PA increased in four non-RCTs at less than 8 weeks follow-up. This was true of counselling-alone in one study, and counselling accompanied by information or prescription. Five RCTs were eligible, of which one was judged as high, two as unclear, and two as low risk of bias (Table 2). All involved physician counselling in a single session plus follow-up counselling by a physician or PA specialist, information, tools (log book, pedometer) and/or PA prescription. Length of counselling sessions was not specified. Self-reported PA increased at 3 months based on multiple sessions compared with a brief, single session and PA prescription [34]. In another study self-reported PA increased and was sustained at 24 months among those who received follow-up counselling sessions after a brief, single session compared with usual care [35]. Self-reported PA increased in one study based on counselling and prescription compared with patient self-monitoring with a pedometer and diary [36]. In another study self-reported PA increased but did not differ at 14 months between those who received counselling only and counselling plus information plus an offer of a subsidized session with a PA specialist [37]. Counselling plus both information and PA prescription resulted in greater selfreported PA at 6 months compared with those who received usual care in another study [38]. Two observational studies were eligible and judged as high quality (Table 3). In one study, a single counselling session by a trained lifestyle change facilitator with up to 12 follow-up counselling sessions at the patient’s discretion increased PA at 6 months, particularly among those who participated in more counselling sessions [39]. In a second observational study, a single physician counselling session of unspecified length was based on assessment of patient motivation, and supplemented with information and the option for additional counselling [40]. At 12 months, 37.2% self-reported increased PA but many said they did not use the leaflet or additional counselling. A key finding of this systematic review is that few studies have focused on evaluating PA promotion in family practice settings. Among the systematic reviews, PA counselling design and delivery were mixed so it is difficult to draw conclusions. In an attempt to elucidate the contextual factors that influenced counselling success, intervention details reported in RCTs and observational studies were categorized according to PRECEDE constructs of the PRECEDE–PROCEED model [25]. All but one of the studies assessed patient readiness for change but it was not clear how this was used
A.R. Gagliardi et al. / Patient Education and Counseling 98 (2015) 412–419
415
Table 1 Systematic reviews evaluating PA counselling. Study
Studies
Intervention design
Impact
Quality
Orrow 2012 UK [31]
10/15 studies on PA counselling in primary care (all RCTs) n = 5991 Up to 2010
Results reflect pooling of data across all 15 studies. PA promotion resulted in small to medium improvement in self-reported PA at 12 months: odds ratio 1.45 (95% CI 1.17–1.73), standardized mean difference 0.25 (0.11–0.38). Nature, timing and who delivered the intervention varied widely across studies and were not examined. No sub-group analyses for patient characteristics.
High
Tulloch 2006 Canada [32]
19 studies (15 RCTs) n = 9853 Up to 2003
Counselling by physician alone (1 study); physician plus exercise, PA or health specialist (2 studies); physician or nurse plus exercise, PA or health specialist (3 studies); nurse plus exercise, PA or health specialist (1 study); health or exercise specialist alone (2 studies); or physiotherapist alone (1 study). Motivational interviewing to assess readiness and barriers was used in seven studies. Counselling sometimes accompanied by educational material (5 studies), referral (3 studies), prescription (5 studies), or follow-up visit or phone calls (5 studies). Single counselling sessions were assessed in five studies and multiple counselling sessions in five studies. Length of counselling sessions not reported. Counselling by physician only (7 studies), physician plus allied health/educator (5 studies), or allied health/educators (7 studies). Sometimes accompanied by information or prescription and/or tailored based on assessment of patient motivation or follow-up counselling sessions in office setting. Most counselling was brief (5 min or less) when offered by a physician or physician plus other provider. Counselling ranged from 20 to 40 min when offered only by a PA specialist in four studies.
High
Lawlor 2001 UK [33]
8 studies (2 RCTs) n = 4747 1993–1999
Short term evaluation (6 months) found an increase in self-reported PA (2/7 physician only, 2/5 combined, and 5/7 AH/E only studies). Nature, timing and who delivered the intervention varied widely across studies. No meta-analysis of findings because interventions and measures were mixed. No sub-group analyses for patient characteristics. In four of six studies (none were RCTS) that evaluated short term results (
