You Are Here:
Behavioral Counseling to Promote Physical Activity and a Healthful Diet to Prevent Cardiovascular Disease in Adults
A Systematic Review for the U.S. Preventive Services Task Force
Release Date: December 2010
By Jennifer S. Lin, MD, MCR; Elizabeth O'Connor, PhD; Evelyn P. Whitlock, MD, MPH; and Tracy L. Beil, MS
The information in this report is intended to help clinicians, employers, policymakers, and others make informed decisions about the provision of health care services. This report is intended as a reference and not as a substitute for clinical judgment.
This report may be used, in whole or in part, as the basis for the development of clinical practice guidelines and other quality enhancement tools, or as a basis for reimbursement and coverage policies. AHRQ or U.S. Department of Health and Human Services endorsement of such derivative products may not be stated or implied.
This article was first published in Annals of Internal Medicine on December 7, 2010 (Annals of Internal Medicine 2010;153(11):736-750; www.annals.org). Select for copyright and source information.
Background: Poor diet and lack of physical activity can worsen cardiovascular health, yet most Americans do not meet diet and physical activity recommendations.
Purpose: To assist the U.S. Preventive Services Task Force in updating its previous recommendations by systematically reviewing trials of physical activity or dietary counseling to prevent cardiovascular disease.
Data Sources: MEDLINE, PsycINFO, Cochrane Central Register of Controlled Trials (2001 to January 2010), experts, and existing systematic reviews.
Study Selection: Two investigators independently reviewed 13,562 abstracts and 481 articles against a set of a priori inclusion criteria and critically appraised each study by using design-specific quality criteria.
Data Extraction and Analysis: Data from 73 studies (109 articles) were abstracted by one reviewer and checked by a second reviewer. Random-effects meta-analyses were conducted for multiple intermediate health and behavioral outcomes.
Data Synthesis: Long-term observational follow-up of intensive sodium reduction counseling showed a decrease in the incidence of cardiovascular disease; however, other direct evidence for reduction in disease morbidity is lacking. High-intensity dietary counseling, with or without physical activity counseling, resulted in changes of -0.3 to -0.7 kg/m2 in body mass index (adiposity), -1.5 mm Hg (95% CI, -0.9 to -2.1 mm Hg) in systolic blood pressure, -0.7 mm Hg (CI, -0.6 to -0.9 mm Hg) in diastolic pressure, -0.17 mmol/L (CI, -0.09 to -0.25 mmol/L) (-6.56 mg/dL [CI, -3.47 to -9.65 mg/dL]) in total cholesterol level, and -0.13 mmol/L (CI, -0.06 to -0.21 mmol/L) (-5.02 mg/dL [CI, -2.32 to -8.11 mg/dL]) in low-density lipoprotein cholesterol level. Medium- and high-intensity counseling resulted in moderate to large changes in self-reported dietary and physical activity behaviors.
Limitation: Meta-analyses for some outcomes had large statistical heterogeneity or evidence for publication bias. Only 11 trials followed outcomes beyond 12 months.
Conclusion: Counseling to improve diet or increase physical activity changed health behaviors and was associated with small improvements in adiposity, blood pressure, and lipid levels.
Primary Funding Source: Agency for Healthcare Research and Quality.
Diseases associated with physical inactivity and poor diet rank among the leading causes of illness and death in the United States (1, 2) and are well-established determinants in many chronic diseases, including cardiovascular and cerebrovascular disease, hypertension, dyslipidemia, and type 2 diabetes (1, 3-5). Although persons of all ages benefit from regular exercise (5), Americans generally do not meet recommended levels of physical activity (6, 7). Americans' diets are also drastically out of line with the generally accepted dietary recommendations published by many organizations, such as the U.S. Department of Health and Human Services (8, 9).
In 2002 and 2003, the USPSTF concluded that evidence was insufficient to recommend for or against behavioral counseling in primary care settings to promote physical activity (I recommendation) or behavioral counseling to promote a healthful diet in unselected patients in primary care (I recommendation). We therefore performed this review to assist in the updating of these recommendations. This review combines counseling for both physical activity and dietary change with a focus on the prevention of cardiovascular disease in adults without known diabetes, hypertension, dyslipidemia, or coronary heart disease. To conduct this review, we developed an analytic framework with 4 key questions (Appendix Figure) that included the effect of dietary or physical activity counseling on health outcomes (key question 1), intermediate cardiovascular disease-related outcomes (key question 2), behavioral outcomes (key question 3), and harms of the counseling interventions (key question 4).
The full report (10) provides a detailed description of our methods, including search strategies and excluded studies.
Data Sources and Searches
To identify literature published for each key question since the previous recommendations, we searched MEDLINE, PsycINFO, and the Cochrane Central Register of Controlled Trials from January 2001 to January 2010. We supplemented our searches with suggestions from experts and reference lists from other relevant publications, including the 2 previous USPSTF systematic reviews and 9 related existing reviews (11-21).
Two investigators independently reviewed 13,562 abstracts and 481 articles against the specified inclusion criteria (Figure). We included trials with primary care-relevant counseling on physical activity (for example, aerobic activities, such as walking, cycling, or swimming, or resistance training) or healthful diet interventions (for example, appropriate calorie intake; increased intake of fruits and vegetables, whole grains, and fiber; balanced intake of fats; or decreased sodium). We excluded interventions primarily aimed at weight loss or those that provided controlled diets or supervised physical activity. Primary care-relevant counseling included interventions that were conducted in, judged feasible to be conducted in (such as phone or electronic interventions), or potentially referable from a primary care setting. Interventions had to be compared with usual care, a minimal intervention, or an attention-control group. We excluded interventions that targeted persons with known hypertension, hyperlipidemia, diabetes, or cardiovascular disease and trials in which more than 50% of the population had known heart disease or any one or a combination of these risk factors. We required a minimum follow-up of 6 months after randomization. A priori outcomes included true health outcomes (morbidity or mortality related to cardiovascular disease); intermediate outcomes and physiologic changes associated with health outcomes (blood pressure, lipid profile, fasting glucose level and glucose tolerance, and adiposity); and behavioral outcomes (any self-reported change in physical activity or dietary intake). We did not include cost-effectiveness or cost-related outcomes. For harms, we included any observational studies that reported serious cardiovascular harms, such as acute cardiac events during or immediately after physical activity.
Data Extraction and Quality Assessment
Articles that met our inclusion criteria were critically appraised by 2 reviewers using the USPSTF and National Institute for Health and Clinical Excellence design-specific quality criteria (22). The reviewers were each blinded to the other's initial ratings, and discrepancies were resolved by consensus. Articles were rated as good, fair, or poor quality. Good-quality studies met all of the specified quality criteria, whereas fair-quality studies did not but had no fatal flaws in the design, execution, or reporting of the study. Poor-quality studies were excluded from this review.
For included studies, 1 investigator extracted data on study setting, populations, interventions, and prespecified outcomes into standardized evidence tables and a second investigator verified all extracted data.
Data Synthesis and Analysis
We conducted random-effects meta-analyses to estimate the effect size of counseling on all intermediate health outcomes and behavioral outcomes. We combined all trial with a given outcome and conducted separate analyses for each of the 3 intervention targets (physical activity, healthful diet, and combined) and, if applicable, for the specific dietary message (sodium reduction, focus on fruits and vegetables only, or general low-fat or heart-Healthful Dietary counseling). Analyses were stratified by estimated intervention intensity (low [≤30 minutes], medium [between 31 minutes and 6 hours of contact], or high [>6 hours of contact]). Trials were also categorized by population risk as being unselected or selected only on the basis of age; selected for suboptimal behavior (such as sedentary behavior or poor dietary intake); or selected for individual or population risk factors for increased incidence of cardiovascular disease (such as mildly elevated diastolic blood pressure or fasting glucose or serum lipid levels, obesity, or poverty or poor access to health care).
We assessed the presence of statistical heterogeneity among the studies by using standard chi-square tests and estimated the magnitude of heterogeneity by using the I2 statistic (23). Tests of publication bias to determine whether the distribution of the effect sizes was symmetric with respect to the precision measure were performed by using funnel plots and the Egger linear regression method (24, 25). Meta-regressions were performed on the basis of the random-effects models to examine the effect of 4 a priori variables of heterogeneity (intervention intensity, intervention target, study population risk, and recruitment method [volunteer vs. study-identified]) on effect size. To interpret effect sizes of standardized mean differences, we used the Cohen d statistic, in which an effect size of 0.2 to 0.3 generally represents a small effect; 0.5, a moderate effect; and 0.8, a large effect (26).
All analyses were performed by using Stata, version 10.0 (StataCorp, College Station, Texas).
Role of the Funding Source
We worked with 4 USPSTF liaisons at key points throughout the review process to develop and refine the scope and resolve other issues. The Agency for Healthcare Research and Quality funded this research under a contract to support the USPSTF. The funding source had no role in study selection, quality assessment, or synthesis, although they provided project oversight, reviewed the draft evidence synthesis, and assisted in external review of the draft evidence synthesis.
We identified 66 trials that reported the effects of counseling for physical activity, healthful diet, or both. Thirty trials (27-71), comprising 15,265 patients, tested the effect of counseling persons to increase physical activity, and 26 of these (14,172 patients) were included in our meta-analyses (27-29, 31, 33, 35-37, 42, 45, 47, 48, 50, 51, 53-57, 59, 61, 63-65, 67, 69, 70). Twenty-five trials (29, 30, 37, 38, 42, 43, 48, 72-103), comprising 71,267 patients, tested the effect of counseling persons to eat a healthful diet; 24 of these studies (70,969 patients) were included in our meta-analyses (37, 42, 48, 72-78, 81, 82, 85-93, 96, 99, 103). Seventeen trials (42-44, 76, 104-125), comprising 5105 patients, examined the effects of counseling persons for both physical activity and healthful diet; we included 15 of these (4475 patients) in our meta-analyses (42, 76, 104, 106, 108, 111, 113-118, 121, 124, 125). Some trials were not included in our meta-analyses because they did not provide necessary data; for example, some reported results only in figures without providing group means and others did not provide SEs or SDs. In addition, we identified 7 studies (126-132), comprising 4354 patients, that addressed the harms of physical activity. The tables in the Supplement (available at www.annals.org) provide study details for counseling trials.
Of the 66 counseling trials, only 13 were good-quality trials (31, 45, 53-55, 57, 59, 82, 96, 99, 116, 118, 125). Fair-quality trials often did not report randomization procedures, allocation concealment, or blinding of outcomes assessment. Many fair-quality trials also limited analyses to persons who completed the study. Some trials did not have allocation concealment of randomization or blinded outcome assessment. Some fair-quality trials, as opposed to good-quality trials, had higher attrition (for example, >20%), differential attrition, or differences in baseline characteristics between groups. Most trials, including the good-quality ones, did not report fidelity of or adherence to counseling interventions.
Trials included a representative range of persons. Men accounted for approximately 17% of all trial participants, or about 35% of trial participants excluding the largest trial that enrolled only women (99). Eleven trials included only women (48, 54, 61, 69, 70, 78, 92, 99, 115, 118, 121), and 5 trials included only men (42, 59, 64, 65, 103). Trials were primarily conducted in middle-aged adults (weighted mean age, 59 years [range, 20 to 78 years]). Nine trials (37, 39, 45, 53, 59, 64, 66, 75, 87) were explicitly conducted in older adults, and most of these trials involved only physical activity counseling (37, 39, 45, 53, 59, 64, 66). Among the 38 trials conducted in the United States, approximately 18% of participants were nonwhite (range in individual trials, 3% to 100%).
Counseling interventions ranged from low-intensity (such as those conducted in a single session or by mail alone) to very high-intensity (such as those with up to 20 sessions over 2 years). Trials were not evenly distributed across categories of study population risk (Table 1). Among diet counseling trials, low-intensity interventions were almost all conducted in general-risk persons and medium- to high-intensity interventions were mostly conducted in selected persons. In contrast, almost all of the physical activity counseling trials were conducted in persons who did not meet recommended levels of physical activity, and only 1 counseling intervention was high-intensity. The combined lifestyle counseling interventions showed a very different distribution, in that medium-intensity trials tended to target patients selected for risk factors related to cardiovascular disease, whereas high-intensity trials primarily involved unselected populations.
Key Question 1: Health Outcomes
Data on the morbidity or mortality related to cardiovascular disease were limited to 3 large, good-quality trials (Table 2). In the Women's Health Initiative Dietary Modification Trial (48,835 participants), postmenopausal women who were randomly assigned to intensive low-fat dietary counseling had no difference in major coronary heart disease events (hazard ratio [HR], 0.97 [95% CI, 0.90 to 1.06]), stroke (HR, 1.02 [CI, 0.90 to 1.15]), or death (HR, 0.98 [CI, 0.91 to 1.07]) after approximately 8.1 years of follow-up (100, 101). Long-term observational follow-up of TOHP (Trials of Hypertension Prevention) I and II (3126 participants) showed that persons with mildly elevated diastolic blood pressure (80 to 89 mm Hg) who were randomly assigned to intensive sodium restriction counseling had fewer cardiovascular events and revascularization (2415 participants; HR, 0.70 [CI, 0.53 to 0.94]), but no difference in total mortality (3126 participants; HR, 0.80 [CI, 0.51 to 1.26]) after 10 to 15 years of follow-up (134). A more conservative cardiovascular disease composite outcome (without revascularization) was not statistically significant, although the point estimate was similar (2415 participants; HR, 0.72 [CI, 0.50 to 1.03]). Visual inspection of cumulative incidence curves of cardiovascular disease events in TOHP I and II suggests that the intervention and control groups began to diverge at approximately 8 to 9 years of follow-up.
Key Question 2: Intermediate Outcomes
Medium- to high-intensity dietary interventions (with or without concomitant physical activity counseling) decreased body mass index at about 12 months. Statistical heterogeneity was high (I2 > 70%), making interpretation of pooled effect sizes questionable (Table 3). Two thirds of the trials of high-intensity dietary interventions, however, reported statistically significant group differences, suggesting that although the amount of weight change varied greatly from study to study, these interventions are likely to reduce weight (decrease in body mass index of approximately 0.3 to 0.7 kg/m2). Physical activity counseling trials were limited to primarily medium-intensity interventions for this outcome and generally did not reduce adiposity. Five trials evaluating high-intensity counseling had follow-up longer than 12 months; the reduction in body mass index persisted up to 72 months, although this result was slightly attenuated (99, 113, 117, 118, 121).
High-intensity diet and lifestyle interventions decreased systolic and diastolic blood pressure by 1.5 mm Hg (CI, 0.9 to 2.1 mm Hg) (78, 82, 91, 93, 96, 99, 104, 118, 121) and 0.7 mm Hg (CI, 0.6 to 0.9 mm Hg) (78, 82, 91, 93, 96, 99, 104, 117, 118, 121), respectively, at 12 months (Table 3). The largest reduction in blood pressures occurred in 3 intensive salt-restriction counseling interventions in persons with mildly elevated diastolic blood pressure, resulting in a change of approximately -1.9 mm Hg (CI, -1.2 to -2.5 mm Hg) in systolic blood pressure and -1.0 mm Hg (CI, -0.5 to -1.6 mm Hg) in diastolic blood pressure at 12 months (Table 3) (82, 93, 96). Reductions in blood pressure were still statistically significant, although slightly attenuated, at up to 36 months (82, 93, 96). Four other trials reported follow-up beyond 12 months, but none showed a statistically significant reduction by the end of the trials (99, 113, 118, 121). Meta-analyses of 6 trials showed that medium-intensity physical activity interventions did not improve blood pressure.
High-intensity trials targeting diet and combined lifestyle counseling decreased total and low-density lipoprotein cholesterol levels; changes were -0.17 mmol/L (CI, -0.09 to -0.25) (-6.56 mg/dL [CI, -3.47 to -9.65 mg/dL]) and -0.13 mmol/L (CI, -0.06 to -0.21) (-5.02 mg/dL [CI, -2.32 to -8.11 mg/dL]), respectively (Table 3) (72, 91, 104, 117, 118, 121). Of the 3 trials that reported follow-up beyond 12 months, 2 trials conducted exclusively in women demonstrated a persistent decrease in total cholesterol or low-density lipoprotein cholesterol levels at 18 or 54 months (118). There was no statistically significant increase in high-density lipoprotein cholesterol levels. Medium-intensity trials did not improve lipid levels on average. Very few trials that targeted physical activity alone reported lipid levels.
We found no evidence of statistically significant publication bias for any of the intermediate outcomes, on the basis of the Egger test and visual inspection of funnel plots. Trials that focused on physical activity–only counseling did not often report measures of adiposity, blood pressure, or lipids (8 studies; 3731 participants).
Key Question 3: Behavioral Outcomes
In general, counseling resulted in small increases in participants' physical activity levels, especially in trials that provided at least medium-intensity interventions. Twenty-four (27-29, 31, 33, 35-37, 42, 47, 48, 50, 51, 53-55, 57, 59, 61, 63, 64, 69, 70, 133) of the 30 trials of physical activity interventions (12,938 participants) and 13 (42, 104, 106, 108, 111, 113-116, 118, 121, 124, 125) of the 17 combined lifestyle trials (4150 participants) were included in the meta-analyses of physical activity level. Pooled standardized mean differences for medium- and high-intensity interventions at approximately 12 months were 0.18 (CI, 0.12 to 0.24) and 0.26 (CI, 0.14 to 0.37), respectively (Table 4). These pooled estimates represent small changes in physical activity levels: for example, an increase in physical activity level by 38.3 min/wk (CI, 25.9 to 50.7 min/wk) (Table 4). Longer-term follow-up was very rare, but changes in activity level were maintained in 1 trial of medium-intensity physical activity (71) and 1 trial of high-intensity physical activity plus lifestyle counseling that reported follow-up beyond 12 months (118). The data for low-intensity interventions were mixed. Although the meta-analysis of the trials of low-intensity physical activity showed no overall effect, 4 of the 11 trials of low-intensity interventions increased some measure of physical activity (35, 39, 55, 65); 2 of these 4 trials were not included in the meta-analysis because of limitations in reporting of outcomes at the individual study level. Only 1 of the combined lifestyle trials with this outcome involved a low-intensity intervention, and it showed no effect (114). The Egger test of small study effects was statistically significant for the medium-intensity trials; however, sensitivity analyses without the smallest trials showed similar results.
Meta-analyses showed statistically significant reductions in self-reported total fat intake at all levels of intervention intensity; however, statistical heterogeneity was high (I2 > 70%) (Table 4). For high-intensity trials, 7 of 10 trials showed statistically significant reductions in total fat, suggesting a robust finding of group differences. High-intensity counseling resulted in moderate to large reductions in self-reported fat intake, ranging from a 5.9% to 11% decrease in energy from total fat and a 2.8% to 3.7% decrease in energy from saturated fat (72, 78, 91, 99, 104, 117, 118). Four trials had longer-term follow-up (up to 72 months), during which reductions in total fat and saturated fat were still significant (99, 117, 118, 121). Low- and medium-intensity interventions resulted in smaller reductions in fat intake (42, 48, 74-77, 81, 86, 89, 92, 111, 114, 116, 124). None of the low- or medium-intensity counseling trials had follow-up beyond 12 months.
Healthful dietary counseling increased fruit and vegetable consumption at all levels of intervention intensity, although statistical heterogeneity was high in some subgroups (Table 4). Three fair-quality trials of behavioral counseling interventions focusing on fruit and vegetable consumption were successful in increasing fruit and vegetable intake (37, 85, 87). Meta-analyses of the low- and medium-intensity general low-fat/heart-healthy counseling trials had low to modest heterogeneity and were statistically significant (48, 75, 77, 86, 88, 90, 92). Although statistical heterogeneity was very large in the high-intensity trials, intervention group participants showed greater increases in fruit and vegetable consumption in 4 of the 5 trials (77, 78, 99, 104, 121). Overall, the mean change in fruit and vegetable intake ranged from 0.4 to 2 servings per day. Only 2 trials had follow-up beyond 12 months; these showed a persistent increase in fruit and vegetable intake at 24 and 72 months (37, 99).
Key Question 4: Harms
Adverse effects were rarely noted in physical activity counseling trials. Two trials reported minor muscular symptoms (27, 50), and 3 trials reported falls related to physical activity (31, 53, 54). We also found 7 case-crossover analyses, which showed that the risk for a cardiac event is increased 2-fold to 17-fold during vigorous exercise (126 -132). This increased risk was much greater for people with low baseline levels of regular activity. Nevertheless, the absolute risk for a cardiac event during physical activity was very low. One of these studies estimated that the incidence of sudden death associated with vigorous physical activity was 1 per 1.42 million person-hours of exercise (126).
We found no studies designed to assess the adverse effects of dietary counseling or the adverse effects of increasing intake of fruits and vegetables or fiber and decreasing intake of sodium or fat. None of the healthful diet counseling trials reported specific adverse events.