We evaluated whether telehealth counseling augments lifestyle change and risk factor decrease in subjects at high risk for primary or secondary cardiovascular events compared to a recommended guideline for brief preventive counseling. Subjects at high risk or with coronary heart disease (35 to 74 years of age, n = 680) were randomized to active control (risk factor feedback, brief advice, handouts) or telehealth lifestyle counseling (active control plus 6 weekly 1-hour teleconferenced sessions to groups of 4 to 8 subjects). Primary outcome was questionnaire assessment of adherence to daily exercise/physical activity and diet (daily vegetable and fruit intake and restriction of fat and salt) after treatment and at 6-month follow-up. Secondary outcomes were systolic and diastolic blood pressures, ratio of total to high-density lipoprotein cholesterol, and 10-year absolute risk for coronary disease. After treatment and at 6-month follow-up, adherence increased for telehealth versus control in exercise (29.3% and 18.4% vs 2.5% and 9.3%, respectively, odds ratio 1.60, 95% confidence interval 1.2 to 2.1) and diet (37.1% and 38.1% vs 16.7% and 33.3%, respectively, odds ratio 1.41, 95% confidence interval 1.1 to 1.9). Telehealth versus control had greater 6-month decreases in blood pressure (mean ± SE, systolic −4.8 ± 0.8 vs −2.8 ± 0.9 mm Hg, p = 0.04; diastolic −2.7 ± 0.5 vs −1.5 ± 0.6 mm Hg, p = 0.04). Decreases in cholesterol ratio and 10-year absolute risk were significant for the 2 groups. In conclusion, telehealth counseling augments therapeutic lifestyle change in subjects at high risk for cardiovascular events compared to a recommended guideline for brief preventive counseling.
The primary aim of this clinical trial was to evaluate whether a telehealth protocol that used motivational interviewing added therapeutic benefit for change in exercise, diet, or smoking in subjects at high risk for or with established cardiovascular disease. We used an active control intervention that represented a recommended guideline for brief preventive counseling. The extent to which telehealth augmented a decrease in cardiovascular risk factors after adjusting for medications was our secondary objective.
Subjects were at high risk for a cardiovascular event; they were 35 to 74 years of age and had a diagnosis of coronary heart disease or diabetes, or Framingham 10-year absolute risk for coronary heart disease ≥20, or ≥2 risk factors that included hypertension, dyslipidemia, smoking, or obesity (body mass index ≥30 kg/m 2 or waist circumference >88 cm in women or >102 cm in men). Exclusion was based on psychiatric illness, alcohol or drug dependence in the previous year, or residence in an institutional setting. Medications were not altered by design.
The Community Outreach Heart Health and Risk Reduction Trial (COHRT) was a 2 parallel-group, single-blind randomized controlled trial: active control versus telehealth counseling assessed at baseline, 2 weeks after treatment, and at 6-month follow-up. Subjects and their physicians were blinded to the research design. Research assistants providing the telehealth intervention had no involvement in outcome evaluations. Randomization was conducted by a computer program that stratified for gender, cardiovascular disease, diabetes, and depression (Beck Depression Inventory II ≥14). Randomization was blocked within our northern, rural, and urban recruitment sites in Ontario, Canada. A 2:3 sampling ratio was used for telehealth versus control groups to permit analysis of psychosocial determinants of lifestyle change (not reported in the present study). Randomization codes were concealed in opaque sealed envelopes that were sorted by stratification features. Approval was obtained from research ethics boards of each institution.
Recruitment initiatives included presentations to family medicine departments and patient groups, random-digit dialing within targeted area codes, and newspaper advertisements. Recruits were screened by telephone interview. Eligible subjects were mailed an information package and a patient profile form, which requested family physicians to confirm diagnoses and cardiovascular risk factors. We provided a prepaid requisition for assays of 12-hour fasting lipoprotein cholesterol if these were not assessed within the previous 6 months. Eligible subjects were scheduled for assessment at a local COHRT clinic after receipt of fasting blood tests and physician confirmation of diagnoses and cardiovascular risk factors. Randomization to telehealth versus control followed informed consent during the initial COHRT clinic visit.
COHRT clinics were held in family medicine outpatient clinics and a behavioral cardiology research unit in 2 tertiary care hospitals (urban site) and collaborating family medicine practices and offices of 2 public health units (rural and northern sites) in Ontario, Canada. COHRT clinics were scheduled between 8:00 a.m. and 12:00 p.m. Subjects were instructed to refrain from smoking and strenuous exercise for ≥4 hours before their appointment. Baseline measurements were taken by trained research assistants for height, weight, body mass index, waist circumference, and blood pressure (2 measurements, 30 minutes apart). Blood assays of 12-hour fasting cholesterol (total, high-density lipoprotein, total/high-density lipoprotein ratio, and low-density lipoprotein) were obtained for the COHRT clinic visit. Procedures for assaying blood samples were standardized across sites using a common laboratory network. Low-density lipoprotein cholesterol was calculated using the Friedewald formula when triglycerides were <4.52 mmol/L (400 mg/dl). The Framingham 10-year absolute risk index was estimated for subjects without established cardiovascular disease. Baseline exercise, diet, and smoking were measured by questionnaire ( Table 1 ). At 2 weeks after treatment, this questionnaire was administered by telephone. However, the 6-month follow-up assessment was conducted during a COHRT clinic visit that replicated the baseline protocol: anthropometric measurements, cardiovascular risk factors (with fasting blood assays obtained within 2 weeks of the visit), and questionnaires.
|Exercise||Planned exercise||I do a planned exercise for ≥20 minutes, 3–5 times/week (e.g., brisk walking, aerobics, jogging, swimming, skiing).|
|Daily activity||I keep active in my daily habits 5–6 days/week (e.g., walking and climbing stairs at home and work, cutting grass, shoveling snow, washing floors).|
|Diet||Vegetables||I eat 3–5 servings of vegetables each day (1/2 cup of raw or cooked vegetables is 1 serving).|
|Fruit||I eat 2–4 servings of fruit each day (1 piece of fruit, such as an apple, is 1 serving).|
|Fat intake||At each meal, I eat <30% of calories from fat by eating lean meat without the skin and avoiding fried foods (e.g., French fries) and high-fat comfort foods such as potato chips.|
|Sodium intake||At each meal, I avoid adding extra salt to my food, and I avoid eating salty foods such as chips, soy sauce, fast foods such as hamburgers, and prepared food mixes.|
|Smoking cessation||I have a smoke-free lifestyle everyday, which does not include even 1 puff of a cigarette.|
⁎ Each questionnaire item was scaled by the transtheoretical model : Precontemplation (not adhering to the target behavior and not ready to change within 6 months), contemplation (not adhering to the target behavior but ready to change within 6 months), preparation (not adhering to the target behavior but ready to change within 1 month), action (adhering to the target behavior for <6 months), and maintenance (adhering to the target behavior for ≥6 months).
Antihypertensive medications (angiotensin-converting enzyme inhibitors, calcium channel blockers, diuretics, β blockers, α-adrenergic blockers, or angiotensin II receptor blockers) or lipid-lowering agents were recorded at baseline and 6-month follow-up. Change in dosage or type of medications was coded for the baseline to 6-month follow-up interval.
During the initial COHRT clinic visit, controls received a 10-minute intervention that included a review and written summary of their cardiovascular risk factor profiles. Subjects without established cardiovascular disease were given their Framingham 10-year absolute risk score. Brief advice for therapeutic change in exercise, diet, and smoking was provided with accompanying educational handouts and a list of community programs for lifestyle change. The profile of each subject’s cardiovascular risk factors, 10-year absolute risk, and severity of depressive symptoms was mailed to their family physician.
The telehealth group received this intervention plus 6 weekly 1-hour sessions of lifestyle counseling by teleconference to small groups (n = 4 to 8). Subjects were matched according to a fixed schedule of weekly sessions. Access to each telehealth session was made by a toll-free number and a private access code. The protocol for telehealth sessions was standardized with a COHRT treatment manual that was a group-based application of motivational interviewing. Subjects identified their priority for lifestyle change (diet, exercise, or smoking) and they were taught to self-assess their stage of readiness for change: precontemplation, contemplation, preparation, action, and maintenance. Across sessions, stage-matched strategies to support lifestyle change were introduced and supported with take-home projects. Motivational interviewing guidelines were used to focus group discussion on (1) salient lifestyle goals identified by subjects, (2) progress in resolving ambivalence about lifestyle change, and (3) experiences of increased efficacy in initiating or maintaining change. Group facilitators included 2 public health nurses, 6 allied health professionals, and 8 PhD students in clinical psychology who completed a 3- to 4-day training program (with R.P.N.). Quality control of this intervention was maintained by weekly supervision by teleconference (with R.P.N. or K.C.).
Primary outcomes were defined as adherence to Health Canada guidelines for exercise (planned weekly exercise or daily activity), diet (daily intake of vegetables and fruit, and daily restriction of fat and salt), and smoke-free living ( Table 1 ). As in previous trials, adherence to each of these behaviors was defined by self-reported criteria for the action or maintenance stages of readiness for change. Questionnaire items for exercise and diet were previously validated according to grade of energy expenditure, food frequency questionnaires, and by body mass index and waist circumference. Secondary outcomes at 6-month follow-up included risk factors (systolic and diastolic blood pressures, total/high-density lipoprotein cholesterol) and the Framingham index of 10-year absolute risk for coronary heart disease.
Criterion validity of self-reported adherence versus nonadherence to exercise and diet at baseline and 6-month follow-up was measured by body mass index, weight decrease, and Health Canada guidelines for active living using analysis of variance and Pearson chi-square test. Baseline characteristics of telehealth subjects versus controls were evaluated with analysis of variance and Pearson chi-square test. Generalized estimation equations assessed whether a larger proportion of telehealth versus control subjects adhered to exercise and diet after treatment and at 6-month follow-up, controlling for baseline adherence to exercise and diet, age, gender, body mass index, and interval (after treatment and 6-month follow-up). Multivariable linear regression analyses evaluated whether telehealth versus control subjects had greater decreases at 6-month follow-up in systolic and diastolic blood pressures and total/high-density lipoprotein cholesterol adjusted for baseline values of each variable, age, gender, body mass index, antihypertensive or lipid-lowering medications at baseline, and change in antihypertensive or lipid-lowering medications up to 6-month follow-up. This regression model also assessed group differences in decrease of the Framingham index of 10-year absolute risk in subjects without cardiovascular disease.
COHRT was powered to detect increased adherence at 4-month follow-up after lifestyle counseling to the action/maintenance stage of readiness for change in dietary fat decrease (odds ratio 2.15, 95% confidence interval 1.3 to 3.6) and physical activity (odds ratio 1.89, 95% confidence interval 1.1 to 3.4) with 80% power, 5% type 1 error rate, and 25% attrition. This estimate used the method of Wang et al to account for a 2:3 sampling ratio for telehealth versus control. Analyses were performed using the intention-to-treat approach. Missing data were managed by multiple imputations using the Markov chain Monte Carlo method. Outcomes were similar across 5 imputations and each was consistent with the raw data. Statistical significance was defined by 2-tailed tests with a p value <0.05. Analyses were conducted with SAS 9.1 (SAS Institute, Cary, North Carolina).
The sample included 267 controls and 413 telehealth subjects ( Figure 1 ). Withdrawals did not differ significantly between groups (controls, n = 30, 11.2%; telehealth subjects, n = 32, 7.7%, p = 0.12). Table 2 presents baseline characteristics. Antihypertensive or lipid-lowering drugs were prescribed to 80% of subjects. There was a high prevalence of change in antihypertensive or lipid-lowering medications from baseline to 6-month follow-up (controls, n = 113, 42.3%; telehealth subjects, n = 171, 41.4%, p = 0.74; Appendix 1 , available on-line).
|Variables||Active Control (n = 267)||Telehealth (n = 413)||p Value|
|Age (years)||58.61 ± 0.53||59.27 ± 0.43||0.34|
|Women||132 (49.4%)||202 (48.9%)||0.89|
|Income groups (<$20,000 = 1, ≥$70,000 = 7)||4.91 ± 0.13||4.50 ± 0.10||0.02|
|Body mass index (kg/m 2 )||30.58 ± 0.38||31.60 ± 0.33||0.047|
|Waist circumference (>88 cm in women, >102 cm in men)||188 (70.4%)||321 (77.7%)||0.03|
|Hypertension||182 (68.2%)||300 (72.6%)||0.21|
|Dyslipidemia||181 (67.8%)||286 (69.2%)||0.69|
|Diabetes mellitus||138 (51.7%)||216 (52.3%)||0.88|
|Smoker||40 (15.0%)||53 (12.8%)||0.43|
|Coronary heart disease risk status||0.38|
|≥2 cardiovascular risk factors ⁎||69 (25.8%)||97 (23.5%)|
|Absolute 10-year coronary heart disease risk ≥20 or diabetes||132 (49.45%)||194 (47.0%)|
|Coronary heart disease||66 (24.75%)||122 (29.55%)|
|β Blocker||73 (27.3%)||114 (27.6%)||0.94|
|Calcium channel blocker||56 (21.0%)||94 (22.8%)||0.58|
|Angiotensin-converting enzyme inhibitor||94 (35.2%)||156 (37.8%)||0.50|
|Angiotensin II receptor blocker||32 (12.0%)||50 (12.1%)||0.96|
|α-Receptor blocker||1 (0.4%)||4 (0.6%)||0.38|
|Diuretic||75 (28.1%)||111 (26.9%)||0.73|
|Lipid-lowering agent||137 (51.3%)||217 (52.5%)||0.75|
|≥1 antihypertensive drug †||187 (70.0%)||298 (72.2%)||0.55|
|≥1 antihypertensive or lipid-lowering drug||212 (79.4%)||332 (80.4%)||0.75|
|Systolic blood pressure (mm Hg)||128.39 ± 0.97||131.30 ± 0.78||0.02|
|Diastolic blood pressure (mm Hg)||77.02 ± 0.56||77.48 ± 0.45||0.52|
|Total/high-density lipoprotein cholesterol (mmol/L) ‡||4.42 ± 0.09||4.35 ± 0.07||0.58|
|Total lipoprotein cholesterol (mmol/L) ‡||5.21 ± 0.09||5.30 ± 0.07||0.45|
|High-density lipoprotein cholesterol (mmol/L) ‡||1.26 ± 0.03||1.29 ± 0.02||0.48|
|Low-density lipoprotein cholesterol (mmol/L) ‡||3.05 ± 0.07||3.03 ± 0.05||0.78|
Criterion validity for questionnaires assessing adherence to diet and exercise was supported by corroborating physiologic data. Subjects who reported baseline adherence versus nonadherence to diet had a lower body mass index (30.1 ± 0.39 vs 31.8 ± 0.32 kg/m 2 , respectively, p <0.001). Body mass index was also lower in subjects who reported adherence versus nonadherence to exercise (30.1 ± 0.29 vs 32.7 ± 0.43 kg/m 2 , respectively, p <0.001). Weight loss from baseline to 6-month follow-up (assessed in COHRT clinics) was greater in subjects who reported adherence versus nonadherence to exercise (−l.56 ± 0.61 vs 0.96 ± 0.88 lb, respectively, p = 0.02) or to diet (−1.96 ± 0.70 vs 0.52 ± 0.71 lb, respectively, p = 0.01). Subjects reporting adherence to exercise at baseline were more likely to meet conventional criteria for active living as assessed by screening items obtained from the Canadian Community Health Survey ( Appendix 2 , available online).
The most prevalent self-selected goals for lifestyle change were exercise (n = 284, 41.8%) and diet (n = 280, 41.2%) or a combination of exercise and diet (n = 59, 8.7%). Smoking cessation was the sole primary goal for change for only 1 subject. At 6-month follow-up, smoking cessation was observed for 3 of 40 controls (7.5%) and 5 of 53 telehealth subjects (9.4%). Hence, additional analyses are not presented.
A larger proportion of telehealth subjects versus controls reported adherence to exercise and diet after treatment and at 6-month follow-up as observed for unadjusted outcomes ( Table 3 ) and after adjustment for covariates ( Table 4 ). All risk factors decreased significantly for telehealth subjects and controls at 6-month follow-up ( Table 5 ). Telehealth subjects demonstrated greater decreases in systolic and diastolic blood pressures but not total/high-density lipoprotein cholesterol or 10-year absolute risk of coronary heart disease ( Table 6 ).
|Variables||After Treatment||6-Month Follow-Up|
|Adherence to exercise ⁎|
|Active control (baseline, n = 161)||165 (2.5%)||176 (9.3%)|
|Telehealth (baseline, n = 239)||309 (29.3%)||283 (18.4%)|
|Adherence to diet †|
|Active control (baseline, n = 132)||154 (16.7%)||176 (33.3%)|
|Telehealth (baseline, n = 197)||270 (37.1%)||272 (38.1%)|
† Defined as daily intake of vegetables (3 to 5 servings per day) and fruit (2 to 4 servings per day) and restriction of fat (<30% of daily calories) and salt (no extra salt at meals and avoidance of prepared foods and salty snacks).
|Variables||Odds Ratio||95% Confidence Interval||p Value|
|Adherence to exercise ⁎|
|Baseline exercise: nonadherence versus adherence||4.77||3.59–6.32||<0.0001|
|Body mass index||0.97||0.95–0.99||0.001|
|Interval: after treatment versus 6-month follow-up ‡||1.13||0.92–1.40||0.28|
|Intervention: active control versus telehealth||1.60||1.20–2.12||0.0009|
|Adherence to diet ⁎|
|Baseline diet: nonadherence versus adherence||5.12||3.82–6.86||<0.0001|
|Body mass index||0.99||0.97–1.02||0.65|
|Interval: after treatment versus 6-month follow-up ‡||0.56||0.46–0.68||<0.0001|
|Intervention: active control versus telehealth||1.31||1.09–1.48||0.012|