A telephonic coaching program has more impact when body mass index is over 35

Article Outline

• Summary
• Purpose
• Methods
  • Design
  • Sample
  • Measures
  • Intervention program
  • Analysis
• Results
• Discussion
  • Summary
  • Limitations
  • Significance
• Conflicts of interest
• References
• Copyright

Summary 

Purpose

The purpose of this study was to test the theory that two indicators of risk (body mass index (BMI) and overall medical risk at baseline) are correlated with weight change in a telephone employer-provided coaching program.

Design

A retrospective cohort study with assessments at baseline and six months after program completion.

Setting

A large manufacturing employer in the United States.

Subjects

Adult employees and dependents enrolled in a voluntary weight loss program.

Intervention

The weight program was based on the Self-Management of Care model. Coaching was based on collaborative goal-setting and included telephonic self-management health education. Clients were staged according to readiness to change.

Measures

Weight change (in kilograms), percent weight change, BMI, health risk indices, readiness to change, and demographic variables.

Analysis

Age, gender, race, education, income, total health risk, readiness to change, and baseline body mass index (BMI) were included as model covariates in a multiple linear regression analysis.

Results

Individuals with a BMI >35 at baseline lost more weight than those with normal weight (p=0.001). Total health risk at baseline was not significantly related to weight loss at p<0.05.

Conclusion

Our findings suggest that the greatest weight loss could be achieved in this telephone coaching program by targeting morbidly obese employees.

Keywords: Occupational health, Health promotion, Weight loss intervention, Health risk appraisal, Telephonic lifestyle coaching, Obesity

Back to Article Outline

Purpose 

Obesity is an independent risk factor for several chronic diseases and conditions and early death [1]. It also is related to pain arising from arthritis [2], migraine progression [3], and orthopedic disorders [4] as well as general pain. [5] Obesity affects the cost of medical care, especially when it results in bariatric surgery [6], [7] and also the number of medical visits consumed by primary care patients [8]. Referral to weight loss programs or nutritional counseling is a common course of action followed by primary care providers, though some do not make such referrals because of high failure rates.

Weight loss counselors work with a wide range of clients, some of whom are overweight and others whose health is compromised due to morbid obesity. If overweight individuals are more likely to succeed, while those who urgently need to lose weight are less likely to succeed, then a rational triage model might allocate visit slots to those who are at moderate risk. This is consistent with Rose's strategy of preventive medicine [9].

A competing hypothesis is derived from the Health Belief Model (HBM) [10]. When this model is applied to weight loss, it predicts that when people face elevated health risks, they may perceive a higher level of urgency that motivates them to lose more weight than those with lower risks [11].

We tested these hypotheses using data obtained from a telephonic coaching program provided to employees and dependents of a large manufacturing company.

Back to Article Outline

Methods 

Design 

This was a retrospective cohort study with assessments at baseline and six months after program completion. An intention-to-treat model was followed in which enrollees were interviewed at follow-up regardless of their level of participation in the program. Enrollees who had both baseline and follow-up weight reported were included in analysis and were considered to be participants of the program.

Sample 

A large, United States employer in the manufacturing industry offered a health-risk telephonic coaching program to its employees and dependents (ages 18 and over). Eligibility was determined by consent and the presence of four or more health risks. An on-line health assessment (HA) evaluated 11 possible health risks. 47,908 employees and dependents took the HA and 16,138 consented to participate in a lifestyle coaching program. Of those that consented, 10,489 were eligible based on HA risks and 2853 actually enrolled in one of the coaching programs. Of the 2853 program enrollees, 1112 of them elected to participate in the weight control program. Self-reported baseline and six-month follow-up weights were available for 936 weight-program participants.

Measures 

The dependent variable for this study was change in weight reported in pounds and analyzed in kilograms. To compute change, weight at initial program intake was subtracted from the six-month follow-up weight. Percent change in weight (change in weight divided by initial weight) was evaluated as a secondary endpoint.

Demographic information was obtained at program intake. Age group (20–39, 40–49, 50–59, and 60 and above), gender, race (White versus other), education group (some high school/diploma, some college, Bachelor's degree, and some graduate school/degree) and income group ($25,000–$49,999; $50,000–$74,999; $75,000–$99,999; $100,000–$124,999; and $125,000 and above) were included as covariates.

BMI and the numbers of medical risks, lifestyle risks, and total risks were the primary independent variables. BMI was determined from self-reported height and weight and divided into standard categories (underweight <18.5, overweight=25–29.9, obese=30–34.9 and morbidly obese ≥35). No one with baseline BMI <18.5 was included in our assessment of the weight management program.

Health risk indices were computed by summing the relevant responses from the HA. Medical risk consisted of high blood pressure (≥120/80mmHg or a self-reported estimate of high blood pressure); high blood sugar (fasting ≥110mg/dL or non-fasting ≥200mg/dL or self-reported estimate of high blood sugar); cholesterol (total cholesterol ≥200mg/dL or HDL <35mg/dL or LDL ≥130mg/dL or total cholesterol to HDL ratio ≥4.5 or a self-reported estimate of high cholesterol); triglycerides (fasting ≥200mg/dL or non-fasting ≥300 or a self-reported estimate of high triglycerides); and weight. Lifestyle risk was computed from responses concerning alcohol use (>7 drinks/week for women or >14 drinks/week for men); emotional health (indicated that stress was often or very often a problem and difficulty coping with stress, or depressed to extent that it affects job or personal life); aerobic exercise (<900kcal/week, calculated from self-reported frequency and duration of aerobic exercise); nutrition (<5 daily servings of fruits/vegetables and/or medium to very high fat diet); safety (does not always wear seat belt); and tobacco (cigarette smoking). Total risk was calculated by summing all medical risks and lifestyle risks, giving one point for each category for a potential maximum risk score of 11.

Readiness to change assessment at intake was guided by the Prochaska and DiClemente's stages of change model [12]. The readiness to change categories consists of the following: pre-contemplative (not interested in changing health behaviors), contemplative (ambivalent about changing health behaviors), preparation (actively planning to change an unhealthy behavior), action (practicing new health behaviors), or maintenance (sustaining new health behaviors) and relapse (resumption of old unhealthy behaviors). Due to low numbers in pre-contemplative and action groups, responders were reclassified into three groups for analysis: pre-contemplative/contemplative, preparation and action/maintenance.

Confidence in maintaining weight was assessed by asking participants at baseline and follow-up “On a scale from 1 to 10, how would rate your confidence in your ability to manage your weight?” (1=not at all confident and 10=totally confident).

Intervention program 

The weight program was based on the Self-Management of Care model [13]. Coaches were not licensed professional counselors, but were trained in the approach before beginning their work. Coaching was based on collaborative goal-setting and included self-management health education. The following goal categories were used: (1) understand principles of weight management; (2) incorporate more physical activity into daily routine; (3) increase exercise; (4) follow food pyramid guidelines; and (5) follow recommended portion sizes. Enrollees were contacted by telephone and a baseline assessment was completed. Following the assessment, the participant was contacted up to six more times over the next six months for discussion about goals and progress. The program was considered “complete” after the 5th phone call. A follow-up call was made six months after enrollment in the program to assess outcomes.

Analysis 

Two sample t-tests, Wilcoxon rank sum tests and chi-square tests were used to assess possible differences between participants and non-participants among the weight-program enrollees for continuous, skewed and categorical variables, respectively. Due to the skewness in the weight change distribution, the data were Winsorized (values less than the 1st percentile were set equal to the 1st percentile (−18.1kg) and values greater than the 99th percentile were set to that value (8.2kg)) prior to the multiple regression analysis to provide more robust assessment. Age, gender, race, education, income, total risk groups, and base BMI were included as model covariates. The impacts of stage of change and program participation (number of program phone calls) were evaluated by adding each variable to the reported regression model. Regression analysis was also performed with percent change in weight as the dependent variable.

Back to Article Outline

Results 

Of the 1112 enrollees, a total of 176 were excluded for missing data. Of those excluded, 174 did not report follow-up weight and 2 were missing baseline weight. These non-participants tended to be younger, have less education, and were more likely to be Black or Asian. There was no difference between participants and non-participants in respect to weight risk, medical risk, lifestyle risk, or total risk. Non-participants had a higher mean base BMI than participants: 33.4 versus 31.9, respectively (p=0.008).

Descriptive statistics of program participants are shown in Table 1. The mean age of the sample was 48.7 years with a standard deviation of 8.4; 36% were female and 83% reported their race as White. The majority of participants had a Bachelor's degree or higher and reported income at or above $75,000. For the 936 program participants, 97% met criteria for weight risk. The mean number of medical risks was 2.2 with a range of 0–5, and the mean number of lifestyle risks was 3.1 with a range of 0–6. The mean number of total health risks was 5.3 with a range of 4–9. Most participants were either in the preparation stage (54.1%) or action stage (32.3%).

Table 1. Demographics and baseline risk factors for lifestyle coaching participants in weight program.

	(N=936)
Male	600 (64.1%)
Age
20–39 years	124 (13.2%)
40–49 years	374 (40.0%)
50–59 years	352 (37.6%)
60+ years	86 (9.2%)
Race
White/Caucasian	778 (83.0%)
Hispanic	37 (4.0%)
Black/African American	45 (4.8%)
Asian	31 (3.3%)
Other	22 (2.4%)
Refused/not reported	23 (2.5%)
Education
High school graduate	106 (11.3%)
Some college/associate's degree	261 (27.9%)
Bachelor's degree	277 (29.6%)
Master's school and above	267 (28.5%)
Refused/not reported	25 (2.7%)
Income
$25,000–$49,999	74 (7.9%)
$50,000–$74,999	235 (25.1%)
$75,000–$99,999	242 (25.9%)
$100,000–$124,999	187 (20.0%)
$125,000 and above	123 (13.1%)
Refused/not reported	75 (8.0%)
BMI groups
Normal	25 (2.7%)
Overweight	380 (40.6%)
Obese	439 (46.9%)
Morbidly obese	92 (9.8%)
Medical risks
0 risks	8 (0.9%)
1 risk	207 (22.1%)
2 risks	393 (42.0%)
3 risks	253 (27.0%)
4 risks	62 (6.6%)
5 risks	13 (1.4%)
Lifestyle risks
0 risks	2 (0.2%)
1 risk	38 (4.1%)
2 risks	184 (19.7%)
3 risks	407 (43.5%)
4 risks	263 (28.1%)
5 risks	38 (4.1%)
6 risks	4 (0.4%)
Total risks
4 risks	262 (28.0%)
5 risks	316 (33.8%)
6 risks	222 (23.7%)
7 risks	100 (10.7%)
8 risks	26 (2.8%)
9 risks	10 (1.1%)
Readiness
Pre-contemplative	15 (1.6%)
Contemplative	107 (11.4%)
Preparation	506 (54.1%)
Action	302 (32.3%)
Maintenance	6 (0.6%)
Number of calls
0 calls	110 (11.8%)
1 call	115 (12.3%)
2 calls	119 (12.7%)
3 calls	89 (9.5%)
4 calls	89 (9.5%)
5 calls	386 (41.2%)
6+ calls	28 (3.0%)

Overall, the mean change in weight was a loss of 1.8kg (4.0pounds) with a range from a gain of 18kg to a loss of 48kg. Participants’ confidence in managing his or her weight increased 1.19 points on a 10-point scale (p<0.001). Unadjusted comparisons on weight loss across demographic, risk factor and participation categories are provided in Table 2. There were significant differences in weight loss by baseline BMI (p<0.001), the number of total risk factors (p=0.021), and program participation (p<0.001) with greater weight loss at higher BMI categories, more risk factors and program completion.

Table 2. Unadjusted results: weight loss (kg) by demographic, risk, and call groups.

	N	Percent change in weight	Change in weight (kg)			Kruskal–Wallis p-value
		Mean	Mean	SD	Median
Gender
Female	336	−1.87%	−1.742	4.563	−0.907	0.438
Male	600	−1.74%	−1.852	5.047	−1.361
Age
20–39 years	124	−1.08%	−0.838	4.030	−0.907	0.540
40–49 years	374	−1.76%	−1.810	4.559	−1.134
50–59 years	352	−1.92%	−1.975	4.617	−1.361
60+ years	86	−2.39%	−2.563	7.534	−1.361
Race
White/Caucasian	778	−1.94%	−1.977	5.012	−1.361	0.138
Black/African American	45	−0.62%	−0.847	5.533	0.000
Hispanic	37	−0.94%	−0.797	3.116	−0.454
Asian	31	−1.22%	−0.922	3.181	−0.907
Other	22	−2.29%	−2.000	3.657	−0.454
Education
Some high school/diploma	106	−2.77%	−2.683	4.306	−1.588	0.083
Some college	261	−1.68%	−1.752	5.255	−1.361
Bachelor's degree	277	−1.96%	−1.926	5.066	−1.361
Some graduate school/degree	267	−1.41%	−1.493	4.581	−0.907
Income
$25,000–$49,999	74	−2.63%	−2.617	4.612	−1.588	0.564
$50,000–$74,999	235	−1.74%	−1.774	5.464	−0.907
$75,000–$99,999	242	−1.88%	−1.929	5.203	−1.134
$100,000–$124,999	187	−1.38%	−1.368	3.987	−0.907
$125,000+	123	−1.98%	−1.969	4.335	−1.361
BMI
Normal	25	−0.51%	−0.327	2.584	0.000	0.001
Overweight	380	−1.42%	−1.196	3.118	−0.907
Obese	439	−1.81%	−1.872	4.940	−1.361
Morbidly obese	92	−3.56%	−4.477	8.675	−2.722
Total risks
Less than 5 risks	262	−1.05%	−1.070	4.136	−0.907	0.021
5 risks	316	−2.18%	−2.180	4.996	−1.361
6 risks	222	−1.60%	−1.655	5.372	−1.361
7 risks	100	−2.51%	−2.472	4.527	−2.268
8+ risks	36	−2.84%	−3.125	5.864	−2.268
Readiness
Pre/contemplative	122	−2.20%	−2.205	6.631	−1.134	0.563
Preparation	506	−1.91%	−1.987	4.870	−1.361
Action/maintenance	308	−1.42%	−1.370	3.972	−0.907
Number of calls
0 calls	110	−0.57%	−0.470	3.468	0.000	<0.001
1 call	115	−1.58%	−1.625	3.876	−0.907
2 calls	119	−1.16%	−1.349	7.058	0.000
3 calls	89	−1.61%	−1.580	3.980	−1.361
4 calls	89	−1.98%	−1.942	4.152	−0.907
5+ calls	414	−2.35%	−2.376	4.935	−1.588

Results of multiple regression analysis are shown in Table 3. Weight loss did not differ significantly between genders, age groups, race categories, or income groups after accounting for other factors. There was a tendency for more weight loss in less educated individuals, however it was not significant. Individuals with a higher BMI, as indicated by BMI >35, lost more weight (p=0.001). Results of the analysis of percent change in weight were consistent with the primary endpoint. These results are reported in Table 4. The only significant associations were with low education and morbid obesity. Table 5 reports the number and percent of individuals in each BMI category by various levels of weight loss. Those with higher total health risk did not experience greater weight loss after adjusting for other variables (p=0.102). After adjusting for health risks, reported stage of change at baseline was not associated with size of weight loss (p=0.573). Meanwhile, program participation was significantly associated with weight loss after adjusting for health risks (p=0.004) with larger weight loss with longer participation.

Table 3. Regression analysis of weight changes adjusting for demographics and risks.

Table 4. Regression analysis of percent of weight changes adjusting for demographics and risks.

Table 5. Unadjusted weight loss: percent of weight loss by BMI groups.

Back to Article Outline

Discussion 

Summary 

As expected, the average weight change across the population after six months was modest in this telephonic lifestyle coaching program. However, there was great individual variation with the maximum weight loss of 48kg. Weight loss in various programs and on various diets tends to approximate about 10pounds. Tate et al. studied a behavioral e-counseling program and found that among 92 subjects losing about 10pounds was typical for those in the experimental group [14]. Repetition of the study confirmed the findings [15]. Wadden, Butryn and Byrne studied the literature and concluded that lifestyle modification induces a loss of approximately 10% of initial weight in 16–26 weeks [16]. Jeffery et al. studied a telephonic employer-sponsored coaching program and found average weight loss to be 2.38kg at six months [17]. A six-month telephone intervention resulted in a mean weight loss of 1.5kg in comparison to a brochure-only (95% CI −2.2;−0.8, p<0.001) [18]. This degree of impact is typical of telephone interventions reported in the medical literature [19], [17], although telephonic interventions are not always effective [20]. Usually not reported is an assessment of how baseline body weight affects the odds or degree of weight loss. Our study addresses this gap in the literature.

In our sample, an elevated BMI (as indicated by BMI >35) was independently related to weight loss. Participants with more risks tended to lose more weight than those with fewer risks in unadjusted comparisons. However, weight appears to dominate the other risks, as the regression model identified only the highest BMI category as significantly related to weight loss. Furthermore, the observed greater weight loss with more complete participation appears to reflect on the effectiveness of the program.

However, our findings do not indicate that certain types of risk carry more motivating force than others. This lack of significance for the individual risk indices could reflect weak reliability or validity in these scales. Neither lifestyle nor medical risks were independently related to weight change after adjustment for total health risk and obesity. Apparently, employees and dependents do not assign more concern to medical risk than to other types of risk, when risk is measured using the indices employed in this study.

Finally, we note that the stages of change did not have a significant effect on weight loss after adjusting for other variables, including baseline obesity. If this finding is confirmed in other studies of employees and dependents, collection of stages of change information may not be necessary, thus simplifying baseline assessments.

Limitations 

This study used self-reported data. The studied population was composed of the employees of one large employer in one region of the country who were offered the intervention program. Furthermore, the study only examined one type of health promotion program, i.e., a telephone weight management coaching program. Since health risk information was obtained from the HA and weight information was obtained during program participation, it is not known how enrollees’ health risk might have changed during this time lapse. Finally, we note participation rates of eligible persons were low, introducing the possibility of selection bias.

Significance 

Despite these limitations, we believe our study is useful because it demonstrates that morbid obesity is positively related to weight loss in a health promotion program, suggesting that targeting these individuals might optimize net impact on population health.

Back to Article Outline

Conflicts of interest 

The authors have no financial conflicts of interest.

Back to Article Outline

References 

1. Patterson R, Frank L, Kristal A, White E. A comprehensive examination of health conditions associated with obesity in older adults. Am J Prev Med. 2004;27(December (5)):385–390
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (83 KB)
   • CrossRef
2. Okoro C, Hootman J, Strine T, Balluz L, Mokdad A. Disability, arthritis, and body weight among adults 45 years and older. Obes Res. 2004;12(May (5)):854–861
   • View In Article
   • MEDLINE
3. Bigal M, Lipton R. Modifiable risk factors for migraine progression. Headache. 2006;46(October (9)):1334–1343
   • View In Article
   • MEDLINE
   • CrossRef
4. Andersen R, Crespo C, Bartlett S, Bathon J, Fontaine K. Relationship between body weight gain and significant knee, hip, and back pain in older Americans. Obes Res. 2003;11(October (10)):1159–1162
   • View In Article
   • MEDLINE
5. Rohrer J, Adamson S, Barnes D, Herman R. Obesity and general pain in family medicine patients; should pain standards call for referral of obese patients to weight management programs?. Qual Manag Health Care. 2008;17(June (3)):204–209
   • View In Article
6. Bachman K. Obesity, weight management, and health care costs: a primer. Dis Manag. 2007;10(June (3)):129–137
   • View In Article
   • MEDLINE
   • CrossRef
7. Powers K, Rehrig S, Jones D. Financial impact of obesity and bariatric surgery. Med Clin North Am. 2007;91(May (3)):321–338ix
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (206 KB)
   • CrossRef
8. Rohrer JE, Adamson SC, Furst JW. Obesity and medical visits in family medicine. Dis Manag Health Outcomes. 2007;15(5):309–313
   • View In Article
9. Rose G. The strategy of preventive medicine. Oxford: Oxford Medical Publications; 1992;
   • View In Article
10. Baranowski T, Cullen KW, Nicklas T, Thompson D, Baranowski J. Are current health behavioral change models helpful in guiding prevention of weight gain efforts?. Obes Res. 2003;11(October (Suppl.)):23S–43S
    • View In Article
11. Daddario DK. A review of the use of the health belief model for weight management. MEDSURG Nurs. 2007;16(December (6)):363–366
    • View In Article
12. Prochaska JO, DiClemente CC. Stages and processes of self-change of smoking: toward an integrative model of change. J Consult Clin Psychol. 1983;51(June (3)):390–395
    • View In Article
    • MEDLINE
    • CrossRef
13. Bodenheimer T, Lorig K, Holman H, Grumbach K. Patient self-management of chronic disease in primary care. JAMA. 2002;288(19):2469–2475
    • View In Article
    • MEDLINE
    • CrossRef
14. Tate D, Wing R, Winett R. Using Internet technology to deliver a behavioral weight loss program. JAMA. 2001;285(March (9)):1172–1177
    • View In Article
    • MEDLINE
    • CrossRef
15. Tate D, Jackvony E, Wing R. Effects of Internet behavioral counseling on weight loss in adults at risk for type 2 diabetes: a randomized trial. JAMA. 2003;289(April (14)):1833–1836
    • View In Article
    • MEDLINE
    • CrossRef
16. Wadden T, Berkowitz R, Womble L, Sarwer DB, Phelan S, Cato RM, et al. Randomized trial of lifestyle modification and pharmacotherapy for obesity. N Engl J Med. 2005;353(November (20)):2111–2120
    • View In Article
    • CrossRef
17. Jeffery RW, Sherwood NE, Brelje K, Tate DF. Mail and phone interventions for weight loss in a managed-care setting: Weigh-To-Be one-year outcomes. Int J Obes Relat Metab Disord. 2003;27(12):1584–1592
    • View In Article
    • MEDLINE
    • CrossRef
18. Ariëns G, Dekkers JC, Hendriksen I, Smid T, van Mechelen W. Phone and e-mail counselling are effective for weight management in an overweight working population: a randomized controlled trial. BMC Public Health. 2009;9(1):6
    • View In Article
    • CrossRef
19. Vale MJ, Jelinek MV, Best JD, Dart AM, Griff LE, Hare DL, et al. Coaching patients On Achieving Cardiovascular Health (COACH): a multicenter randomized trial in patients with coronary heart disease. Arch Intern Med. 2003;163(December (22)):2775–2783
    • View In Article
    • MEDLINE
    • CrossRef
20. Eakin EG, Lawler SP, Vandelanotte C, Owen N. Telephone interventions for physical activity and dietary behavior change: a systematic review. Am J Prev Med. 1 2007;32(May (5)):419–434
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (153 KB)
    • CrossRef