- Email: [email protected]
Partner weight status and subject weight loss: Implications for cost-effective programs and public health
Addictive
Brhavior.s, Vol. 14, pp. 279-289, Printed in the USA. All rights reserved.
1989 Copyright
0306-4603/89 $3.00 + .OO 0 1989 Pergamon Press plc
PARTNER WEIGHT STATUS AND SUBJECT WEIGHT LOSS: IMPLICATIONS FOR COST-EFFECTIVE PROGRAMS AND PUBLIC HEALTH DAVID R. BLACK Purdue University
WILLIAM
E. THRELFALL
Stanford Univeristy Abstract - This study examined the difference in body changes of subjects with overweight partners and those with normal weight partners. The 9 subjects with normal weight partners lost significantly more weight than the 14 subjects with overweight partners by the end of the one-year treatment program (27.9 Ibs versus 20.2 Ibs, respectively) and at 3-month follow-up (29.6 Ibs vs. 17.3 Ibs, respectively). It was also found that both overweight and normal weight partners lost weight incidental to treatment but as expected, body changes of overweight partners significantly decreased over time whereas body changes of normal weight partners did not. Overweight partners and subjects lost statistical equivalent amounts of weight but there was a significant correlation at posttreatment between body changes of subjects in the overweight partner group and their partners. It was concluded that if body changes of both subjects and their overweight partners are considered, couples programs might be a cost-effective and important public health approach because two individuals lose weight as inexpensively as one and more people are treated. Researchers are encouraged to report weight status and body changes of both subjects and partners in order to understand more fully the relationship between partner weight status and subject weight loss.
Obesity researchers and care providers need to establish treatment approaches that are both effective and sensitive to cost (cf. Elder, Hovell, Lasater, Wells, & Carleton, 1985; Yates, 1985). Relevant to these issues are concerns about cost-effectiveness, treating a larger number of people, and long-term maintenance or continuation of weight loss. Cost-effective approaches for health promotion programs, and specifically for the treatment of obesity, are of major concern in light of the continued increase in costs for health care services (Farquhar et al., 1985; U.S. Department of Health and Human Resources, 1984). The weight loss to cost ratio is an important consideration in evaluating weight management programs (e.g., Yates, 1978). One cost-saving alternative and an inducement for treatment is to offer one person treatment for weight control and to provide services to another at no additional charge. Provision of such services is feasible depending on the “vehicle” for treatment. For example, the application of a bibliotherapy program shows potential in this respect because program materials and clinic appointments for evaluation can be shared (Black & Threlfall, 1986). Another related concern is treating a larger number of people. The National Institutes of Health (1985) estimates that about one-third of adult Americans are in need of treatment for obesity. One way to meet this demand is a minimal intervention program (Black & Friesen, 1983). A minimal intervention is the simplest and least costly treatment that works Preparation of this article was supported in part by a grant from the Purdue University Research Foundation and a Purdue University Library Scholars Grant to David R. Black. The authors thank John T. Hultsman, Richard A. Petosa, and Wojteck Chodzko-Zajko for reviewing earlier drafts of this manuscript. Requests for reprints should be sent to David R. Black, Health Promotion, Purdue University, 107B Lambert Building, West Lafayette, IN 47907-1899.
279
280
DAVID
R. BLACK
and WILLIAM
E. THRELFALL
for the person (Black & Hultsman, 1988, 1989; DeLucia, Black, Longhead, & Hultsman, in press; Hultsman, Black, Seehafer, & Hovell, 1987). Minimal intervention program would enable a care provider to reach more people because of the reduction in counselor time required for each participant. Another option is to enroll individuals in pairs in programs, in which one person participates in a more intensive weight control program or a sequence of programs (cf. Black & Hultsman, 1988, 1989; DeLucia et al., in press; Hultsman et al., 1987) and the other receives a minimal intervention treatment. By following this protocol, twice as many participants are treated at less expense than with more resource intensive programs. A third problem that continues to challenge obesity researchers is long-term maintenance or continuation of weight loss after treatment (Hovell, Black, Dellinger, Hofstetter, & Mott, 1989; Jeffery, 1987). One innovative approach to long-term weight maintenance is the involvement of a spouse or partner in the subject’s weight control program. Examples of ways partners may augment treatment efficacy include reinforcing skills acquired during treatment and extending therapy time throughout the week. The weight status of the participating partner, however, may mediate the results of therapy. Black, Gleser, and Kooyers (1989) noted in a statistical meta-analysis of behavioral “couples” studies that groups with a significant other involved in treatment produced statistically significant weight loss results at posttreatment and short-term follow-up compared with a group within the same study where subjects attended treatment alone. The authors also noted extreme inter-subject variability in weight loss and suggested that partner weight status might be one factor that accounts for the variability. To date, only two couples studies have provided data pertaining to partner weight status and weight loss. Dubbert and Wilson (1984) reported weight losses of partners but did not analyze differences between partners who were initially overweight or normal weight. Brownell and Stunkard (1981) found that the majority of spouses (approximately 66%) were moderately but not extremely overweight (M = 47.7% overweight). Both normal weight and overweight spouses in their study lost weight (M = 6.5, 15.6 lbs, respectively) and overweight spouses reduced as much as subjects (15.8 lbs). Subjects did not statistically differ in inital weight regardless of their partners’ weight status and there was not a significant difference in weight loss between subjects based on partner weight status. The purpose of this study was to compare body changes of subjects with overweight partners and subjects with normal partners. A number of other questions were also investigated: (a) did partners lose weight incidental to treatment, (b) what was the relationship of partner weight status and weight loss to initial measures of subjects and subject weight loss, and (c) did body changes of overweight and normal weight partners differ. METHOD
Subjects There were 26 subjects, 8 men and 18 women. They ranged in age from 28 to 62 years old (M = 45.4), weighed 146.5 to 232.0 pounds (M = 186.4), and were 16.6% to 59.8% overweight (M = 37.2%). Recruitment was conducted in three ways; advertisements in two local community newspapers, appearances by the first author on radio and television newscasts, and letters sent to previous inquirers about weight loss programs offered at Stanford University. Enrollment eligibility requirements included being 10% to 75% overweight, having medical clearance to participate, being married or living with a member of the opposite sex, and having a spouse/partner who was willing to assist. Twenty-five subjects were married and one couple had been cohabitating for 3 years.
Partner weight status and subject weight loss
281
Partners There were 5 women and 11 men who were 20% or more overweight; 3 women and 7 men were normal weight. Partners ranged in age from 29 to 68 years old (M = 47.0), weighed 123.8 to 28 1 .O pounds (M = 192.82), and were 1.5% to 84.6% overweight (M = 34.8%). PROCEDURE
Program overview All subjects and their partners enrolled in the same program offered by the Stanford Heart Disease Prevention Program (SHDPP) and attended a follow-up meeting 3 months after the one-year treatment program. Treatment was a graduated two-step program in which the intensity of the intervention was increased depending on the subject’s weight loss progress. Partners were encouraged to assist subjects in losing weight but were neither required nor expected to lose weight. If partners wished to lose weight, they were told to follow the same verbal guidelines for weight reduction that were given to subjects during the first meeting (see Step 1 below). Subjects paid a $75 program fee and a $100 refundable deposit; $25 of the deposit was returned to couples for attendance at each of four assessment meetings during the study. Step 1 All subjects were initially enrolled in Step 1, the minimal intervention (MI) program. During the first meeting, subjects were given a few verbal guidelines about how to lose weight safely and were asked to self-monitor the number of calories consumed and expended daily. They were also to weigh themselves daily, self-record their weight, and mail weight and calorie records to SHDPP weekly. Some subjects continued Step 1 for the entire treatment year. Others began Step 2 based on lack of weight loss progress and concurrence with their counselor that a more intensive treatment would have therapeutic effects. Step 2 Step 2 was a bibliotherapy problem-solving (PS) program and a faded treatment phase. Subjects in the PS program received 10 packets by mail at the rate of approximately one packet per week. Subjects read behavioral weight loss and problem-solving materials, completed and mailed quizzes to SHDPP weekly, and kept the same self-recording calorie intake and expenditure records as well as body weight forms as during step 1. In addition, subjects completed practice exercises to learn and apply problem-solving skills to weight management (Black, 1986). Subjects also rated the type and frequency of weight-related problems they encountered that were associated with overeating and under-exercising. Partners were encouraged to assist subjects in losing weight and to help solve weight-related problems. Faded treatment began after subjects returned their last bibliotherapy packet and continued for the remainder of the treatment year. During this time period, subjects continued to self-monitor and to mail their self-recording forms to SHDPP each week. Follow-up A follow-up meeting was scheduled 3 months after the one-year treatment program. Subjects did not complete self-recording forms during this interim. Follow-up procedures were the same for subjects who completed only Step I and for those who completed both Steps 1 and 2. Counselors Initially, there were five counselors,
each of whom was assigned five or six subjects. Each
282
DAVID R. BLACK and WILLIAM E. THRELFALL
counselor volunteered for different periods of time which meant that subjects were assisted by two or three different counselors during the study. A total of 14 counselors participated during the 15month study. All counselors were student volunteers with limited or no background in behavioral weight management counseling. Counselors were trained by the authors and met every two weeks throughout the project. Counselors assisted subjects during the one-year treatment program by periodically (approximately every 3 to 6 weeks) telephoning or writing subjects to inquire about progress and to answer questions. During the PS program, counselors also provided written feedback regarding the forms, quizzes, and practice exercises returned to SHDPP. No contact with counselors occurred during follow-up. Counselors did not directly assist partners at any time during the study. Dependent measures Pounds lost and percentage overweight lost were the dependent measures for subjects and partners. Body weight for subjects as well as partners was measured to the nearest quarter pound on a balance-beam scale at each assessment meeting. Pounds lost was calculated as the amount of weight lost since the initial weigh-in. Percentage overweight lost was the portion of percentage overweight lost since the initial meeting. Percentage overweight, initially and for each time period, was calculated using the following formula: (actual weight/ideal weight - 1) X 100. Ideal weight for subjects, used to compute percentage overweight, was based on a formula for ideal percentage body fat for men and women and was estimated from skinfold measurements (Rogers, Mahoney, Mahoney, Straw, & Kenigsberg, 1980). Skinfold thicknesses were assessed with Harpenden calipers (model 3496, Quinton Instruments) in accordance with procedures outlined by Allsen, Harrison, and Vance (1984). Reliability of skinfold measurements was evaluated by the two skinfold technicians measuring several of the same subjects at the beginning and end of each meeting. For consistency between and within technicians, coefficients of stability were calculated for each assessment period. The within correlation coefficients and the between coefficients were each averaged according to procedures described by McNemar ( 1969). The weighted averages for within coefficients were .96 and .94 for the two technicians respectively, and the between technician coefficient was .95. Ideal weight for partners, used in the formula for percentage overweight, was computed differently than it was for subjects. Ideal weight for partners was based on the median desirable weight for gender and height on the Metropolitan Life Insurance Table (1959) and not on skinfold measurements. The Metropolitan Life Insurance Table was used in order to compare results with the two other couples studies reporting weight status of partners (cf. Brownell & Stunkard, 1981; Dubbert & Wilson, 1984). Group division The criterion for considering partners as overweight was the same as that reported by Brownell and Stunkard (1981). The 16 subjects whose partners were 20% or more above ideal weight were assigned to the overweight partner group, and the 10 subjects whose partners were less than 20% overweight were assigned to the normal weight partner group. RESULTS
Analyses for subjects Initial measures. Subjects with overweight partners and subjects with normal weight partners did not differ statistically on initial measures. A simple chi-square test (Bruning & Kintz, 1977) indicated no significant difference between groups in gender. The t test for small sample sizes (Hays, 1981) indicated no significant differences in initial weight,
Partner weight status and subject weight loss
percentage
283
overweight,
subjects’ ages, and partners’ ages. An analysis was calculated to determine if there was a significant difference in program utilization. Three subjects in the overweight partner group (18.8%) and one subject in the normal weight partner group (10.0%) used only Step 1. The Fisher exact probability test indicated that subjects in both groups were equivalent in the number of subjects who used one or both steps of the treatment program. Another analysis was caculated to determine if there was a difference between groups in program compliance. Return of the 10 bibliotherapy problem-solving packets was used to estimate compliance. The mean number of packets returned by subjects in the overweight partner group was 9.38 (SD = 1.50) and the mean number for subjects in the normal weight partner group was 8.56 (SD = 2.79). The Mann-Whitney U-Test indicated that there was no significant difference between the groups in the number of packets completed and returned. Attrition. An analysis was completed to assess differences in attrition rates. All 26 subjects completed the one-year program. Two subjects from the overweight partner group (12.5%) and one subject from the normal weight partner group (10%) did not attend the follow-up meeting. The Fisher exact probability test indicated that the groups did not significantly differ in attrition. The 3 subjects who did not attend the follow-up meeting were eliminated from subsequent repeated measures analyses. Body changes. Body changes of subjects in the normal weight partner group and subjects in the overweight partner group were compared. Table 1 shows individual data for subjects over time. Figure 1 graphically displays the mean cumulative weight loss during treatment and follow-up of subjects with overweight partners and subjects with normal weight partners as well as the weight loss of their respective partners. A 2 (Treatments) x 2 (Trials) repeated measures Analysis of Variance (ANOVA) was computed for the difference scores between pretreatment and posttreatment and posttreatment and follow-up for pounds lost and percentage overweight lost for subjects in both groups. A significant group difference was found for pounds lost, F (1, 21) = 4.57, p < .05, and the difference for percentage overweight lost approached significance, F (1, 21) = 2.96, p = .lO. Time and the time by treatment interaction was not significant for either dependent measure. Table 1 and Figure 1 support the significant difference between groups because subjects with normal weight partners lost approximately one-fourth more weight at postreatment and one-third more weight at follow-up than subjects with overweight partners. Program utilization and compliance.
Analyses for Partners Relationship of initial measures and body changes.
Partner and subject initial weight status and partner and subject body changes were correlated. The only significant relationships were noted at posttreatment for subjects in the overweight partner group and their partners. The weight loss and percentage overweight lost of subjects in the overweight partner group and their partners were moderately but significantly correlated, r (12) = .66, p < .Ol ; .54, p < .05, respectively. Body changes. Table 2 shows individual data for overweight and normal weight partners. A one-way repeated measures ANOVA showed that overweight partners lost a significant amount of weight over time, F (1, 26) = 17.29, p < .OOOl. A second one-way repeated measures ANOVA indicated that overweight partners also lost a significant amount of percentage overweight over time, F (1, 26) = 18.23, p < .OOOl. Two other one-way repeated measures ANOVA’s showed that normal weight partners did not significantly change on either dependent measure over time. Body changes of both overweight and normal weight partners were combined and compared with those of subjects in both groups. The results showed that subjects lost more weight than partners. A 2 (Treatments) X 2 (Trials) repeated measures ANOVA was
284
DAVID R. BLACK and WILLIAM
E. THRELFALL
Table I. Body changes of subjects whose partners were either overweight One-year
Subject
1
2 3 4 5 6 7 8 9 10 II I2 13 14 I5 I6
M/F
Initial assessment Age (yrs) Wt. (lb) Percent overwth
Sex”
I 2 3 4 5 6 7 8 9 10
MIF
Subjects with overweight partners 9.3 17.2 28.5 26.9 29.1 47.5 34.3 34.5 17.5 56.4 28.0 53.1 5.8 16.6 59.8 17.0 26.3 29.3 34.1 48.3 22.1 42.0 22.4 +3.0 36.7 20.9 33.3 27.7 9.5 44.6 10.5 30.5
M F F M F M F F F F F F M M F F
61 51 34 51 61 48 41 38 35 56 55 45 55 52 47 31
218.0 151.0 182.3 173.5 172.0 231.0 214.5 198.0 162.0 175.0 157.0 191.0 197.8 200.5 201 .o 146.5
5Mil IF
47.6
185.7
37.1
9.5
24.9
13.6
SD
Cumulative lb lost‘
nrogram Cumulative percent overwt. lost
or normal weight Three-month Cumulative lb lost
follow-up Cumulative percent overwt lost
3.8 24.0 23.5 26.6 15.9 18.5 3.1 13.7 21.0 28.5 19.9 +5.5 14.4 21.2 6.8 9.3
3.0 32.0 18.3 31.8 16.5 23.2 5.3 6.0 29.3 34. I 18.1 +3.5 17.1 II.5 _ _
I .6 26.9 14.8 24.6 15.0 15.4 2.9 4.9 23.4 28.5 16.3 t5.8 11.8 7.3 -
20.2
15.3
17.3
13.4
11.3
9.6
11.9
10.3
15.1 33.0 27.4 12.9 24.9 25.6 22.1 10.3 16.9 12.5
24.3 44.1 36.8 18.8 39.5 28.8 31.1 23.8 19.1 -
15.1 33.5 27.8 13.5 23.4 22.0 20.8 20.5 18.1 _
Subjects with normal weight partners 24.3 47.9 45.0 43.3 36.3 30.8 18.0 34.6 42.0 24.9 33.6 29.6 33.1 32.4 12.0 37.7 17.9 56.8 17.5 34.8
M F F F M F F F F M
28 31 50 36 39 57 28 31 57 62
232.0 190.0 173.0 188.0 211.0 169.8 197.0 159.8 165.0 189.5
3Ml7F
41.9
187.5
37.5
27.8
20.1
29.6
21.6
13.3
22.2
9.7
11.3
7.6
9.0
6.2
SD
“F = Female. M = Male. ‘Percent overwt. = (Actual wt.)/(IdeaI wt.) - I x 100. Ideal weight was based on a formula for ideal percentage body fat and estimated skinfold measurements (Rogers et al., 1980.) ‘Cumulative lb lost and cumulative percent overwt. lost were calculated by subtracting actual weight or percentage overweight at each time period from initial values.
computed for difference scores between pretreatment and posttreatment and posttreatment and follow-up for pounds lost and percentage overweight lost for subjects and partners. There was a significant difference between subjects and partners in both weight lost and percentage overweight lost, F (1, 44) = 17.92, p < .003; 23.24, p < .OOOl; respectively. No significant time effect or time by treatment interaction was noted. This indicated that subjects lost more weight and percentage overweight than partners at posttreatment and follow-up. Body changes of subjects in the overweight partner group and their partners were compared. A 2 (Treatment) x 2 (Trials) repeated measures ANOVA indicated that there was not a significant difference between groups in pounds lost but the value approached significance, F (1, 26) = 2.45, p = .12. The 2 X 2 repeated measures ANOVA for percentage overweight lost indicated a significant group difference, F (1, 26) = 4.33, p < .04. There was a significant finding for time for pounds lost and percentage overweight lost,
285
Partner weight status and subject weight loss
Overweight
Overweight Partners
/
Normal Weight Partners
9)
(n=
I
I
I
I
Initial
I- Year
Program
ASSESSMENT
3-Month Follow-up
TIMES
Fig. 1. Mean cumulative weight loss for partners who were either normal weight or overweight, subjects whose partners were either normal weight or overweight.
and
F (1, 26)
= 10.49, 10.82, p < .004, respectively but there was no significant time by treatment interaction. The significant time effect most likely occurred because both subjects and partners lost a substantial amount of weight throughout the study. The significant difference between groups was substantiated because subjects with overweight partners reduced almost twice as much in percentage overweight than their partners at both time periods. Normal weight partners versus subjects in the normal weight partner group: Body changes of subjects in the normal weight partner group and those of their partners were compared. Results of both 2 X 2 repeated measure ANOVA’s showed a significant difference between subjects and their normal weight partners in pounds lost and percentage overweight lost, F (1, 16) = 35.18, p < .OOOl; 42.93, p < .OOOOl, respectively. The time effect and the time by treatment interaction for both dependent measures were not significant, indicating that subjects in the normal weight partner group lost more weight and percentage overweight than their partners at both time periods. DISCUSSION
The findings of the present study indicate that subjects with normal weight partners in a two-step couples weight control program lost significantly more weight than subjects with overweight partners. Also, body changes of all subjects were significantly greater than those of all partners. As expected, overweight partners lost more weight and percentage overweight than normal weight partners. In addition, overweight partners lost an amount
286
DAVID R. BLACK and WILLIAM
Table 2. Body changes of overweight One-year
Partner
Initial assessment wt. Percent Age overwtd (yrs.) (lb)
1 2 3 4 5 6 7 8 9 10 I1 12 13 14 15 16
55 60 30 51 64 48 40 37 36 58 55 44 54 68 41 39
198.5 195.0 188.0 209.0 172.5 223.0 223.0 260.5 180.5 281.0 187.5 191 .o 168.0 275.0 163.0 199.3
65.4 27.4 26.2 74.2 23.2 68.9 45.8 48.0 28.9 73.5 33.9 36.4 75.0 84.6 44.2 23.0
M SD
49.1 10.9
207.2 36.6
4X.7 21.6
I 2 3 4 5 6 7 8 9 10
30 33 50 38 39 57 29 36 64 60
150.5 171.8 178.5 205.0 138.0 176.8 169.8 206.0 178.3 123.8
10.7 12.3 13.0 16.5 I .5 19.5 7.5 17.0 19.7 9.6
M
43.6 13.0
169.8 26.4
12.7 5.7
SD
Cumulative lb losti’ Overweight 4.5 28.5 7.8 19.3 2.5 20.7 f1.3 16.3 1.2 31 .o 14.5 6.0 10.0 36.0
E. THRELFALL
or normal weight partners program Cumulative percent overwt. lost
Three-month Cumulative lb lost
follow-up Cumulative percent overwt. lost
partners +4.5 28.5 5.3 16.5 4.x 16.9 +.3 16.3 1.3 27.5 14.5 6.0 4.2 2X.0
_
3.7 X.8 5.3 16.1 1.8 15.6 +.X 9.3 +.2 19.2 10.3 4.3 10.4 24.2 -
+3.n 8.X 3.6 13.8 3.4 12.8 t.1 9.3 .9 16.9 10.3 3.3 4.4 18.X -
14.0 Il.9
9.1 7.5
11.8 11.0
7.4 6.7
4.0 18.8
3.0 12.3 4.8 +4.0 8.4 +6.5 +10.5 10.5 7.7 -
Normal weight partners 4.0 3.0 8.5 5.6 7.5 4.8 +4.0 i2.3 9.1 6.1 +1.7 i3.6 +4.3 +6.X +3.0 +1.X 5.7 3.9 2.1 6.0
1.3 4.3
7.5 +7.0 I I.4 +5.9 + 16.6 18.5 11.4 4.7 12.2
2.9 8.1
“Percent overwt. = (Actual wt.)/(Ideal wt.) - 1 X 100. Ideal weight was based on the median desirable weight for gender and height on the Metropolitan Life Insurance Table (1959). ‘Cumulative lb lost and cumulative percent overwt. lost were calculated by subtracting actual body weight or percentage overweight at each time period from initial values.
equivalent to subjects in the overweight partner group and there was a significant relationship between body changes of subjects in the overweight partner group and their partners at posttreatment. The authors recognize that conclusions from studies using an ex post facto design, such as this one, are not as definitive as those from experimental designs (cf. Rubinson & Neutens, 1987). It should be noted, though, that initial measures and process variables used to evaluate the homogeniety of groups indicate that the groups are statistically equivalent with the exception of the desired difference in initial weight of partners. Another question that might be raised about the study concerns the length of treatment, and perhaps the length of follow-up (cf. Brownell, 1982). The 15-month time period of the study, is however, comparable to other behavioral weight loss studies (cf. Stuart, Mitchell. & Jensen, 1981) and time was used differently in an effort to produce greater body changes and to promote long-term weight loss. One other concern might be because the two dependent measures differ slightly in outcome. It has been noted, however, that slight discrepancies between
Partner weight status and subject weight loss
287
measures of body change are common in the behavioral weight loss literature (e.g., Brownell, 1982; Hall, Bass, & Monroe, 1978; Hanson, Borden, Hall, & Hall, 1976). The reason subjects with normal weight partners are more successful in losing weight than subjects with overweight partners is not entirely clear. One possible explanation is related to the treatment procedures used. In the present study, a two-stepped less intensive approach was used and partners were asked to assist in the problem-solving process. Other studies indicated that a stepped procedure and a problem-solving strategy were important components of a successful weight control program (e.g., Black, 1987; Black & Scherba, 1983). Interestingly too, Brownell and Stunkard (1981), using a more traditional face-to-face counseling procedure, found no significant difference in the weight loss of subjects according to partner weight status. Another reason subjects with normal weight partners may have lost more weight, based on verbal report, may be that they did not need to compete with their partners to lose weight. Zitter and Fremouw (1978) found that competition with a weight-reducing partner was not advantageous to subjects in losing weight. Further research is needed to ascertain the reasons for the weight loss advantage of subjects with normal weight partners. Several important issues for the provision of service are raised by this study. When body changes of subjects are examined alone, the changes for subjects in the normal weight partner group are superior and approximately one-third greater than subjects in the overweight partner group. When body changes of subjects and partners are considered together, however, the outcome and conclusions are quite different. From this perspective, couples programs are a cost-effective approach because subjects as well as partners lose weight without any additional expenditure of human and economic resources. Couples programs may also provide an important alternative from a public health perspective because more people lose weight if partners are involved than if participants receive treatment alone. Losing weight together becomes even more appealing because oveweight partners lose as much weight as subjects. The amount that partners lost is impressive and is comparable to the average weight loss reported in intensive behavioral programs (cf. Black et al., 1989; Brownell & Jeffery, 1987; Hovel1 et al., 1989; Stuart et al., 1981). The incidental weight loss of partners regardless of their weight status, and the 14-pound weight loss of overweight partners using minimal instruction and counselor assistance, also reaffirms the efficacy of a minimal intervention program (cf. Black et al., 1984; Black & Friesen, 1983). The reason partners were able to lose weight using a minimal intervention program may be because they felt they were receiving a “bargain” by not having to pay for additional services and, therefore, their expectation of benefit increased (Stanton, 1976). Partners also participated to a lesser degree than subjects which may have decreased their feelings of therapeutic demand (Loro, Fisher, & Levenkron, 1979) and increased their feelings of perceived freedom and intrinsic motivation (Lepper, 1981; Neulinger, 1981). Future research is needed to identify what motivational and weight loss methods partners use to lose weight and why some partners lose weight and others do not. This study has heuristic value for researchers and clinicians. The weight status and weight loss of partners seem to be factors that may have affected outcome in previous couples studies. Weight status of partners might, therefore, be considered a blocking variable or covariate and body changes of both subjects and partners should be reported in couples studies in order to understand more fully the relationship between partner weight status and subject weight loss. The potential for long-term effectiveness of couples studies is still in question (Black et al., 1989) but partner weight status and weight loss seem to be related to subjects’ weight loss and may affect maintenance as well. In addition, when both subject and partner weight loss are considered, the possible cost advantage and public health benefit of
288
DAVID R. BLACK and WILLIAM
E. THRELFALL
expending the same human and economic resources to serve more people may be a valuable and, as yet, untapped advantage of couples programs for weight control.
REFERENCES Allsen. P.E.. Harrison, J.M., &Vance, B. (1984). Fitnessfor /$e: An individurrli~edrrpprouch (3rd ed.). Dubuque. IA: William C. Brown. Black, D.R. (1986. November). Problem solving in a stepped-approach for weight control. In A.M. Nezu (Chair). Social problem solving: Recent advances in theory, research, and clinical applications. Symposium conducted at the meeting of the Association for Advancement of Behavior Therapy, Chicago, IL. Black. D.R. (1987). A minimal intervention and a problem-solving program for weight control. Cognitive Therap) and Research, 11, 107-I 19. Black, D.R., Coe. W.C., Friesen, J.G., & Wurzmann. A.G. (1984). Minimal intervention for weight control: A cost-effective alternative. Addictive Behaviors, 9, 279-285. Black, D.R., & Friesen. J.G. (1983). Deposit money: A component in a self-directed minimal intervention program for weight control. Behavior Themop?, 14. 333-340. Black, D.R.,Gleser, L.J., & Kooyers, K.J. (1989). A meta-analytic evcdurztion of‘ couples byeight 10s~ programs. Manuscript submitted for publication. Black, D.R., & Hultsman, J.T. (1989). The Purdue Stepped Approach Model: h4arsholling existent communit\ health resources to reduce cardiovnscular risk factors. Manuscript submitted for publication. Black, D.R., & Hultsman, J.T. (1989). The Purdue Approach Model: A heuristic application to health counseling. The Counseling Psychologist, 16, 647-667. Black, D.R.. & Scherba, D.S. (1983). Contracting to problem solve versus contracting to practice behavioral weight loss skills. Behavior Therapy, 14, 100-109. Black, D.R., & Threlfall, W.E. (1986). A stepped approach to weight control: A minimal intervention and a bibliotherapy problem-solving program. Behavior Therap!. 17, 144-l 57. Brownell, K.D. (1982). Obesity: Understanding and treating a serious. prevalent. and refractory disorder. Journnl c>fConsulting and C/in&d PsJchology, 50, 820-840. Brownell, K.D., & Jeffery, R.W. (1987). Improving long-term weight loss: Pushing the hmits of treatment. Behavior Therap?, 18, 353-374. Brownell, K.D.. & Stunkard, A.J. (1981). Couples training pharmacotherapy, and behavior in the treatment of obesity. Archir,es of General Ps,vchiutry, 38. 1224-1229. Bruning. J.L., & Kin&. B.L. (1977). Computotionnl handbook of .stntistics (2nd ed.). Glenview. IL: Scott. Foresman. DeLucia. J.L.. Black, D.R., Loughead, T.A.. & Hultsman, J.T. (in press). Purdue Stepped Approach Model: Groups as a symbiosis of career development and mental health counseling. Journal of Career Development. Dubbert. P.M.. & Wilson, G.T. (1984). Goal-setting and spouse involvement in the treatment of obesity. Behnviour. Research, & Therapy, 22, 227-242. Elder, J.P.. Hovell. M.F., Lasater, T.M., Wells, B.L., & Carleton. R.A. (1985). The application of behavior modification to community health education: The case of heart disease prevention. Health Educarion Quarterly, 12. 151-168. Farquhar. J. W., Fortman, S.P., Maccoby, N., Haskell, W.L., Williams, P.T., Flora, J.A., Taylor. C.B., Brown, B.W., Solomon, D.S.. & Hulley, S.B. (1985). The Stanford Five-City Project: Deisgn and methods. American Journal of Epidemiology, 122, 323-324. Hall, S.M.. Baas. A., & Monroe, J. (1978). Continued contact and monitoring as follow-up strategies: A long-term study of obesity treatment. Addictive Behaviors. 3, 139-147. Hanson, R.W., Borden, B.L., Hall, S.M., & Hall, R.G. (1976). Use of programmed instruction in teaching self-management skills to overweight adults. Behavior Therupy, 7, 366-373. Hays. W.L. (1981). Statistics. New York: CBS College Publishing. Hovell. M.F.. Black, D.R.. Dellinger. A., Hofstetter, R., & Mott, M. (1989). Weight loss maintenuncet A.s.se.ssment of beholaiorol rrnining am/ very Ion cnlorie mod$ed @sting diets. Manuscript submitted for publication. Hultsman. J.T., Black, D.R., Seehafer, R.W., & Hovell, M.F. (1987). The Purdue Stepped Approach Model: Application to leisure counseling service delivery. Therapeutic Recreation Journd, 21(4). 9-22. Jeffery. R.W. (1987). Behavioral treatment of obesity. Annals of Behavioral Medicine, 9. 20-24. Lepper, M.R. (1981). Intrinsic and extrinsic motivation in children: Detrimental effects of superfluous social control. In W.A. Collins (Ed.), Aspects of the development of competence: The Minnesota Symposium on Child Psychology (Vol. 14). Hillsdale, NJ: Lawrence Erlbaum. Loro. A.D., Fisher, E.B., & Levenkron, J.C. (1979). Comparison of established and innovative weight-reduction treatment procedures. Journal @Applied Behavior AncdFsis, 12, 141-155. McNemar, Q. (1969). Psychological statistics (4th ed.). New York: John Wiley & Sons. Metropolitan Life Insurance Company. (1959). New weight standards for men and women. Staristical Bulletin, 40, l--1. National Institutes of Health Consensus Development Panel on the Health Implications of Obesity. (1985). Health
Partner weight status and subject weight loss
289
implications of obesity: NIH consensus development statement. Annals ojlnrernal Medicine, 103, 1073-1077. Neulinger, J. (1981). The psychology of leisure(2nd ed.). Springfield, IL: Charles C. Thomas. Rogers, T., Mahoney, M.J., Mahoney, B.K., Straw, M.K., & Kenigsberg, MI. (1980). Clinical assessment of obesity: An empirical evaluation of diverse techniques. Behavioral Assessmenf, 3, 161-181. Rubinson, L., & Neutens, J.J. (1987). Research techniques for the health sciences. New York: Macmillan. Stanton, H.E. (1976). Fee-paying and weight loss: Evidence for an interesting interaction. American Journal of Clinical Hypnosis, 19, 47-49. Stuart, R.B., Mitchell, C., & Jensen, J.A. (1981). Therapeutic options in the management of obesity. In L.A. Bradley & C.K. Prokop (Eds.), Medical Psychology: A new perspective (pp. 321-3.53). New York: Academic Press. U.S. Department of Health and Human Resources. (1984). Health, Unired Stares, 1984. Washington, DC: U.S. Government Printing Office. Yates, B.T. (1978). Improving the cost-effectiveness of obesity programs: Three basic strategies for reducing the cost per pound. International Journal ojObesiry, 2, 249-266. Yates, B.T. (1985). Cost-effective analysis and cost-benefit analysis: An introduction. Behavioral Assessmenf, 7, 207-234. Zitter, R.E., & Fremouw, W.J. (1978). Individual versus partner consequation for weight loss. Behavior Therapy, 9. 808-813.
Brhavior.s, Vol. 14, pp. 279-289, Printed in the USA. All rights reserved.
1989 Copyright
0306-4603/89 $3.00 + .OO 0 1989 Pergamon Press plc
PARTNER WEIGHT STATUS AND SUBJECT WEIGHT LOSS: IMPLICATIONS FOR COST-EFFECTIVE PROGRAMS AND PUBLIC HEALTH DAVID R. BLACK Purdue University
WILLIAM
E. THRELFALL
Stanford Univeristy Abstract - This study examined the difference in body changes of subjects with overweight partners and those with normal weight partners. The 9 subjects with normal weight partners lost significantly more weight than the 14 subjects with overweight partners by the end of the one-year treatment program (27.9 Ibs versus 20.2 Ibs, respectively) and at 3-month follow-up (29.6 Ibs vs. 17.3 Ibs, respectively). It was also found that both overweight and normal weight partners lost weight incidental to treatment but as expected, body changes of overweight partners significantly decreased over time whereas body changes of normal weight partners did not. Overweight partners and subjects lost statistical equivalent amounts of weight but there was a significant correlation at posttreatment between body changes of subjects in the overweight partner group and their partners. It was concluded that if body changes of both subjects and their overweight partners are considered, couples programs might be a cost-effective and important public health approach because two individuals lose weight as inexpensively as one and more people are treated. Researchers are encouraged to report weight status and body changes of both subjects and partners in order to understand more fully the relationship between partner weight status and subject weight loss.
Obesity researchers and care providers need to establish treatment approaches that are both effective and sensitive to cost (cf. Elder, Hovell, Lasater, Wells, & Carleton, 1985; Yates, 1985). Relevant to these issues are concerns about cost-effectiveness, treating a larger number of people, and long-term maintenance or continuation of weight loss. Cost-effective approaches for health promotion programs, and specifically for the treatment of obesity, are of major concern in light of the continued increase in costs for health care services (Farquhar et al., 1985; U.S. Department of Health and Human Resources, 1984). The weight loss to cost ratio is an important consideration in evaluating weight management programs (e.g., Yates, 1978). One cost-saving alternative and an inducement for treatment is to offer one person treatment for weight control and to provide services to another at no additional charge. Provision of such services is feasible depending on the “vehicle” for treatment. For example, the application of a bibliotherapy program shows potential in this respect because program materials and clinic appointments for evaluation can be shared (Black & Threlfall, 1986). Another related concern is treating a larger number of people. The National Institutes of Health (1985) estimates that about one-third of adult Americans are in need of treatment for obesity. One way to meet this demand is a minimal intervention program (Black & Friesen, 1983). A minimal intervention is the simplest and least costly treatment that works Preparation of this article was supported in part by a grant from the Purdue University Research Foundation and a Purdue University Library Scholars Grant to David R. Black. The authors thank John T. Hultsman, Richard A. Petosa, and Wojteck Chodzko-Zajko for reviewing earlier drafts of this manuscript. Requests for reprints should be sent to David R. Black, Health Promotion, Purdue University, 107B Lambert Building, West Lafayette, IN 47907-1899.
279
280
DAVID
R. BLACK
and WILLIAM
E. THRELFALL
for the person (Black & Hultsman, 1988, 1989; DeLucia, Black, Longhead, & Hultsman, in press; Hultsman, Black, Seehafer, & Hovell, 1987). Minimal intervention program would enable a care provider to reach more people because of the reduction in counselor time required for each participant. Another option is to enroll individuals in pairs in programs, in which one person participates in a more intensive weight control program or a sequence of programs (cf. Black & Hultsman, 1988, 1989; DeLucia et al., in press; Hultsman et al., 1987) and the other receives a minimal intervention treatment. By following this protocol, twice as many participants are treated at less expense than with more resource intensive programs. A third problem that continues to challenge obesity researchers is long-term maintenance or continuation of weight loss after treatment (Hovell, Black, Dellinger, Hofstetter, & Mott, 1989; Jeffery, 1987). One innovative approach to long-term weight maintenance is the involvement of a spouse or partner in the subject’s weight control program. Examples of ways partners may augment treatment efficacy include reinforcing skills acquired during treatment and extending therapy time throughout the week. The weight status of the participating partner, however, may mediate the results of therapy. Black, Gleser, and Kooyers (1989) noted in a statistical meta-analysis of behavioral “couples” studies that groups with a significant other involved in treatment produced statistically significant weight loss results at posttreatment and short-term follow-up compared with a group within the same study where subjects attended treatment alone. The authors also noted extreme inter-subject variability in weight loss and suggested that partner weight status might be one factor that accounts for the variability. To date, only two couples studies have provided data pertaining to partner weight status and weight loss. Dubbert and Wilson (1984) reported weight losses of partners but did not analyze differences between partners who were initially overweight or normal weight. Brownell and Stunkard (1981) found that the majority of spouses (approximately 66%) were moderately but not extremely overweight (M = 47.7% overweight). Both normal weight and overweight spouses in their study lost weight (M = 6.5, 15.6 lbs, respectively) and overweight spouses reduced as much as subjects (15.8 lbs). Subjects did not statistically differ in inital weight regardless of their partners’ weight status and there was not a significant difference in weight loss between subjects based on partner weight status. The purpose of this study was to compare body changes of subjects with overweight partners and subjects with normal partners. A number of other questions were also investigated: (a) did partners lose weight incidental to treatment, (b) what was the relationship of partner weight status and weight loss to initial measures of subjects and subject weight loss, and (c) did body changes of overweight and normal weight partners differ. METHOD
Subjects There were 26 subjects, 8 men and 18 women. They ranged in age from 28 to 62 years old (M = 45.4), weighed 146.5 to 232.0 pounds (M = 186.4), and were 16.6% to 59.8% overweight (M = 37.2%). Recruitment was conducted in three ways; advertisements in two local community newspapers, appearances by the first author on radio and television newscasts, and letters sent to previous inquirers about weight loss programs offered at Stanford University. Enrollment eligibility requirements included being 10% to 75% overweight, having medical clearance to participate, being married or living with a member of the opposite sex, and having a spouse/partner who was willing to assist. Twenty-five subjects were married and one couple had been cohabitating for 3 years.
Partner weight status and subject weight loss
281
Partners There were 5 women and 11 men who were 20% or more overweight; 3 women and 7 men were normal weight. Partners ranged in age from 29 to 68 years old (M = 47.0), weighed 123.8 to 28 1 .O pounds (M = 192.82), and were 1.5% to 84.6% overweight (M = 34.8%). PROCEDURE
Program overview All subjects and their partners enrolled in the same program offered by the Stanford Heart Disease Prevention Program (SHDPP) and attended a follow-up meeting 3 months after the one-year treatment program. Treatment was a graduated two-step program in which the intensity of the intervention was increased depending on the subject’s weight loss progress. Partners were encouraged to assist subjects in losing weight but were neither required nor expected to lose weight. If partners wished to lose weight, they were told to follow the same verbal guidelines for weight reduction that were given to subjects during the first meeting (see Step 1 below). Subjects paid a $75 program fee and a $100 refundable deposit; $25 of the deposit was returned to couples for attendance at each of four assessment meetings during the study. Step 1 All subjects were initially enrolled in Step 1, the minimal intervention (MI) program. During the first meeting, subjects were given a few verbal guidelines about how to lose weight safely and were asked to self-monitor the number of calories consumed and expended daily. They were also to weigh themselves daily, self-record their weight, and mail weight and calorie records to SHDPP weekly. Some subjects continued Step 1 for the entire treatment year. Others began Step 2 based on lack of weight loss progress and concurrence with their counselor that a more intensive treatment would have therapeutic effects. Step 2 Step 2 was a bibliotherapy problem-solving (PS) program and a faded treatment phase. Subjects in the PS program received 10 packets by mail at the rate of approximately one packet per week. Subjects read behavioral weight loss and problem-solving materials, completed and mailed quizzes to SHDPP weekly, and kept the same self-recording calorie intake and expenditure records as well as body weight forms as during step 1. In addition, subjects completed practice exercises to learn and apply problem-solving skills to weight management (Black, 1986). Subjects also rated the type and frequency of weight-related problems they encountered that were associated with overeating and under-exercising. Partners were encouraged to assist subjects in losing weight and to help solve weight-related problems. Faded treatment began after subjects returned their last bibliotherapy packet and continued for the remainder of the treatment year. During this time period, subjects continued to self-monitor and to mail their self-recording forms to SHDPP each week. Follow-up A follow-up meeting was scheduled 3 months after the one-year treatment program. Subjects did not complete self-recording forms during this interim. Follow-up procedures were the same for subjects who completed only Step I and for those who completed both Steps 1 and 2. Counselors Initially, there were five counselors,
each of whom was assigned five or six subjects. Each
282
DAVID R. BLACK and WILLIAM E. THRELFALL
counselor volunteered for different periods of time which meant that subjects were assisted by two or three different counselors during the study. A total of 14 counselors participated during the 15month study. All counselors were student volunteers with limited or no background in behavioral weight management counseling. Counselors were trained by the authors and met every two weeks throughout the project. Counselors assisted subjects during the one-year treatment program by periodically (approximately every 3 to 6 weeks) telephoning or writing subjects to inquire about progress and to answer questions. During the PS program, counselors also provided written feedback regarding the forms, quizzes, and practice exercises returned to SHDPP. No contact with counselors occurred during follow-up. Counselors did not directly assist partners at any time during the study. Dependent measures Pounds lost and percentage overweight lost were the dependent measures for subjects and partners. Body weight for subjects as well as partners was measured to the nearest quarter pound on a balance-beam scale at each assessment meeting. Pounds lost was calculated as the amount of weight lost since the initial weigh-in. Percentage overweight lost was the portion of percentage overweight lost since the initial meeting. Percentage overweight, initially and for each time period, was calculated using the following formula: (actual weight/ideal weight - 1) X 100. Ideal weight for subjects, used to compute percentage overweight, was based on a formula for ideal percentage body fat for men and women and was estimated from skinfold measurements (Rogers, Mahoney, Mahoney, Straw, & Kenigsberg, 1980). Skinfold thicknesses were assessed with Harpenden calipers (model 3496, Quinton Instruments) in accordance with procedures outlined by Allsen, Harrison, and Vance (1984). Reliability of skinfold measurements was evaluated by the two skinfold technicians measuring several of the same subjects at the beginning and end of each meeting. For consistency between and within technicians, coefficients of stability were calculated for each assessment period. The within correlation coefficients and the between coefficients were each averaged according to procedures described by McNemar ( 1969). The weighted averages for within coefficients were .96 and .94 for the two technicians respectively, and the between technician coefficient was .95. Ideal weight for partners, used in the formula for percentage overweight, was computed differently than it was for subjects. Ideal weight for partners was based on the median desirable weight for gender and height on the Metropolitan Life Insurance Table (1959) and not on skinfold measurements. The Metropolitan Life Insurance Table was used in order to compare results with the two other couples studies reporting weight status of partners (cf. Brownell & Stunkard, 1981; Dubbert & Wilson, 1984). Group division The criterion for considering partners as overweight was the same as that reported by Brownell and Stunkard (1981). The 16 subjects whose partners were 20% or more above ideal weight were assigned to the overweight partner group, and the 10 subjects whose partners were less than 20% overweight were assigned to the normal weight partner group. RESULTS
Analyses for subjects Initial measures. Subjects with overweight partners and subjects with normal weight partners did not differ statistically on initial measures. A simple chi-square test (Bruning & Kintz, 1977) indicated no significant difference between groups in gender. The t test for small sample sizes (Hays, 1981) indicated no significant differences in initial weight,
Partner weight status and subject weight loss
percentage
283
overweight,
subjects’ ages, and partners’ ages. An analysis was calculated to determine if there was a significant difference in program utilization. Three subjects in the overweight partner group (18.8%) and one subject in the normal weight partner group (10.0%) used only Step 1. The Fisher exact probability test indicated that subjects in both groups were equivalent in the number of subjects who used one or both steps of the treatment program. Another analysis was caculated to determine if there was a difference between groups in program compliance. Return of the 10 bibliotherapy problem-solving packets was used to estimate compliance. The mean number of packets returned by subjects in the overweight partner group was 9.38 (SD = 1.50) and the mean number for subjects in the normal weight partner group was 8.56 (SD = 2.79). The Mann-Whitney U-Test indicated that there was no significant difference between the groups in the number of packets completed and returned. Attrition. An analysis was completed to assess differences in attrition rates. All 26 subjects completed the one-year program. Two subjects from the overweight partner group (12.5%) and one subject from the normal weight partner group (10%) did not attend the follow-up meeting. The Fisher exact probability test indicated that the groups did not significantly differ in attrition. The 3 subjects who did not attend the follow-up meeting were eliminated from subsequent repeated measures analyses. Body changes. Body changes of subjects in the normal weight partner group and subjects in the overweight partner group were compared. Table 1 shows individual data for subjects over time. Figure 1 graphically displays the mean cumulative weight loss during treatment and follow-up of subjects with overweight partners and subjects with normal weight partners as well as the weight loss of their respective partners. A 2 (Treatments) x 2 (Trials) repeated measures Analysis of Variance (ANOVA) was computed for the difference scores between pretreatment and posttreatment and posttreatment and follow-up for pounds lost and percentage overweight lost for subjects in both groups. A significant group difference was found for pounds lost, F (1, 21) = 4.57, p < .05, and the difference for percentage overweight lost approached significance, F (1, 21) = 2.96, p = .lO. Time and the time by treatment interaction was not significant for either dependent measure. Table 1 and Figure 1 support the significant difference between groups because subjects with normal weight partners lost approximately one-fourth more weight at postreatment and one-third more weight at follow-up than subjects with overweight partners. Program utilization and compliance.
Analyses for Partners Relationship of initial measures and body changes.
Partner and subject initial weight status and partner and subject body changes were correlated. The only significant relationships were noted at posttreatment for subjects in the overweight partner group and their partners. The weight loss and percentage overweight lost of subjects in the overweight partner group and their partners were moderately but significantly correlated, r (12) = .66, p < .Ol ; .54, p < .05, respectively. Body changes. Table 2 shows individual data for overweight and normal weight partners. A one-way repeated measures ANOVA showed that overweight partners lost a significant amount of weight over time, F (1, 26) = 17.29, p < .OOOl. A second one-way repeated measures ANOVA indicated that overweight partners also lost a significant amount of percentage overweight over time, F (1, 26) = 18.23, p < .OOOl. Two other one-way repeated measures ANOVA’s showed that normal weight partners did not significantly change on either dependent measure over time. Body changes of both overweight and normal weight partners were combined and compared with those of subjects in both groups. The results showed that subjects lost more weight than partners. A 2 (Treatments) X 2 (Trials) repeated measures ANOVA was
284
DAVID R. BLACK and WILLIAM
E. THRELFALL
Table I. Body changes of subjects whose partners were either overweight One-year
Subject
1
2 3 4 5 6 7 8 9 10 II I2 13 14 I5 I6
M/F
Initial assessment Age (yrs) Wt. (lb) Percent overwth
Sex”
I 2 3 4 5 6 7 8 9 10
MIF
Subjects with overweight partners 9.3 17.2 28.5 26.9 29.1 47.5 34.3 34.5 17.5 56.4 28.0 53.1 5.8 16.6 59.8 17.0 26.3 29.3 34.1 48.3 22.1 42.0 22.4 +3.0 36.7 20.9 33.3 27.7 9.5 44.6 10.5 30.5
M F F M F M F F F F F F M M F F
61 51 34 51 61 48 41 38 35 56 55 45 55 52 47 31
218.0 151.0 182.3 173.5 172.0 231.0 214.5 198.0 162.0 175.0 157.0 191.0 197.8 200.5 201 .o 146.5
5Mil IF
47.6
185.7
37.1
9.5
24.9
13.6
SD
Cumulative lb lost‘
nrogram Cumulative percent overwt. lost
or normal weight Three-month Cumulative lb lost
follow-up Cumulative percent overwt lost
3.8 24.0 23.5 26.6 15.9 18.5 3.1 13.7 21.0 28.5 19.9 +5.5 14.4 21.2 6.8 9.3
3.0 32.0 18.3 31.8 16.5 23.2 5.3 6.0 29.3 34. I 18.1 +3.5 17.1 II.5 _ _
I .6 26.9 14.8 24.6 15.0 15.4 2.9 4.9 23.4 28.5 16.3 t5.8 11.8 7.3 -
20.2
15.3
17.3
13.4
11.3
9.6
11.9
10.3
15.1 33.0 27.4 12.9 24.9 25.6 22.1 10.3 16.9 12.5
24.3 44.1 36.8 18.8 39.5 28.8 31.1 23.8 19.1 -
15.1 33.5 27.8 13.5 23.4 22.0 20.8 20.5 18.1 _
Subjects with normal weight partners 24.3 47.9 45.0 43.3 36.3 30.8 18.0 34.6 42.0 24.9 33.6 29.6 33.1 32.4 12.0 37.7 17.9 56.8 17.5 34.8
M F F F M F F F F M
28 31 50 36 39 57 28 31 57 62
232.0 190.0 173.0 188.0 211.0 169.8 197.0 159.8 165.0 189.5
3Ml7F
41.9
187.5
37.5
27.8
20.1
29.6
21.6
13.3
22.2
9.7
11.3
7.6
9.0
6.2
SD
“F = Female. M = Male. ‘Percent overwt. = (Actual wt.)/(IdeaI wt.) - I x 100. Ideal weight was based on a formula for ideal percentage body fat and estimated skinfold measurements (Rogers et al., 1980.) ‘Cumulative lb lost and cumulative percent overwt. lost were calculated by subtracting actual weight or percentage overweight at each time period from initial values.
computed for difference scores between pretreatment and posttreatment and posttreatment and follow-up for pounds lost and percentage overweight lost for subjects and partners. There was a significant difference between subjects and partners in both weight lost and percentage overweight lost, F (1, 44) = 17.92, p < .003; 23.24, p < .OOOl; respectively. No significant time effect or time by treatment interaction was noted. This indicated that subjects lost more weight and percentage overweight than partners at posttreatment and follow-up. Body changes of subjects in the overweight partner group and their partners were compared. A 2 (Treatment) x 2 (Trials) repeated measures ANOVA indicated that there was not a significant difference between groups in pounds lost but the value approached significance, F (1, 26) = 2.45, p = .12. The 2 X 2 repeated measures ANOVA for percentage overweight lost indicated a significant group difference, F (1, 26) = 4.33, p < .04. There was a significant finding for time for pounds lost and percentage overweight lost,
285
Partner weight status and subject weight loss
Overweight
Overweight Partners
/
Normal Weight Partners
9)
(n=
I
I
I
I
Initial
I- Year
Program
ASSESSMENT
3-Month Follow-up
TIMES
Fig. 1. Mean cumulative weight loss for partners who were either normal weight or overweight, subjects whose partners were either normal weight or overweight.
and
F (1, 26)
= 10.49, 10.82, p < .004, respectively but there was no significant time by treatment interaction. The significant time effect most likely occurred because both subjects and partners lost a substantial amount of weight throughout the study. The significant difference between groups was substantiated because subjects with overweight partners reduced almost twice as much in percentage overweight than their partners at both time periods. Normal weight partners versus subjects in the normal weight partner group: Body changes of subjects in the normal weight partner group and those of their partners were compared. Results of both 2 X 2 repeated measure ANOVA’s showed a significant difference between subjects and their normal weight partners in pounds lost and percentage overweight lost, F (1, 16) = 35.18, p < .OOOl; 42.93, p < .OOOOl, respectively. The time effect and the time by treatment interaction for both dependent measures were not significant, indicating that subjects in the normal weight partner group lost more weight and percentage overweight than their partners at both time periods. DISCUSSION
The findings of the present study indicate that subjects with normal weight partners in a two-step couples weight control program lost significantly more weight than subjects with overweight partners. Also, body changes of all subjects were significantly greater than those of all partners. As expected, overweight partners lost more weight and percentage overweight than normal weight partners. In addition, overweight partners lost an amount
286
DAVID R. BLACK and WILLIAM
Table 2. Body changes of overweight One-year
Partner
Initial assessment wt. Percent Age overwtd (yrs.) (lb)
1 2 3 4 5 6 7 8 9 10 I1 12 13 14 15 16
55 60 30 51 64 48 40 37 36 58 55 44 54 68 41 39
198.5 195.0 188.0 209.0 172.5 223.0 223.0 260.5 180.5 281.0 187.5 191 .o 168.0 275.0 163.0 199.3
65.4 27.4 26.2 74.2 23.2 68.9 45.8 48.0 28.9 73.5 33.9 36.4 75.0 84.6 44.2 23.0
M SD
49.1 10.9
207.2 36.6
4X.7 21.6
I 2 3 4 5 6 7 8 9 10
30 33 50 38 39 57 29 36 64 60
150.5 171.8 178.5 205.0 138.0 176.8 169.8 206.0 178.3 123.8
10.7 12.3 13.0 16.5 I .5 19.5 7.5 17.0 19.7 9.6
M
43.6 13.0
169.8 26.4
12.7 5.7
SD
Cumulative lb losti’ Overweight 4.5 28.5 7.8 19.3 2.5 20.7 f1.3 16.3 1.2 31 .o 14.5 6.0 10.0 36.0
E. THRELFALL
or normal weight partners program Cumulative percent overwt. lost
Three-month Cumulative lb lost
follow-up Cumulative percent overwt. lost
partners +4.5 28.5 5.3 16.5 4.x 16.9 +.3 16.3 1.3 27.5 14.5 6.0 4.2 2X.0
_
3.7 X.8 5.3 16.1 1.8 15.6 +.X 9.3 +.2 19.2 10.3 4.3 10.4 24.2 -
+3.n 8.X 3.6 13.8 3.4 12.8 t.1 9.3 .9 16.9 10.3 3.3 4.4 18.X -
14.0 Il.9
9.1 7.5
11.8 11.0
7.4 6.7
4.0 18.8
3.0 12.3 4.8 +4.0 8.4 +6.5 +10.5 10.5 7.7 -
Normal weight partners 4.0 3.0 8.5 5.6 7.5 4.8 +4.0 i2.3 9.1 6.1 +1.7 i3.6 +4.3 +6.X +3.0 +1.X 5.7 3.9 2.1 6.0
1.3 4.3
7.5 +7.0 I I.4 +5.9 + 16.6 18.5 11.4 4.7 12.2
2.9 8.1
“Percent overwt. = (Actual wt.)/(Ideal wt.) - 1 X 100. Ideal weight was based on the median desirable weight for gender and height on the Metropolitan Life Insurance Table (1959). ‘Cumulative lb lost and cumulative percent overwt. lost were calculated by subtracting actual body weight or percentage overweight at each time period from initial values.
equivalent to subjects in the overweight partner group and there was a significant relationship between body changes of subjects in the overweight partner group and their partners at posttreatment. The authors recognize that conclusions from studies using an ex post facto design, such as this one, are not as definitive as those from experimental designs (cf. Rubinson & Neutens, 1987). It should be noted, though, that initial measures and process variables used to evaluate the homogeniety of groups indicate that the groups are statistically equivalent with the exception of the desired difference in initial weight of partners. Another question that might be raised about the study concerns the length of treatment, and perhaps the length of follow-up (cf. Brownell, 1982). The 15-month time period of the study, is however, comparable to other behavioral weight loss studies (cf. Stuart, Mitchell. & Jensen, 1981) and time was used differently in an effort to produce greater body changes and to promote long-term weight loss. One other concern might be because the two dependent measures differ slightly in outcome. It has been noted, however, that slight discrepancies between
Partner weight status and subject weight loss
287
measures of body change are common in the behavioral weight loss literature (e.g., Brownell, 1982; Hall, Bass, & Monroe, 1978; Hanson, Borden, Hall, & Hall, 1976). The reason subjects with normal weight partners are more successful in losing weight than subjects with overweight partners is not entirely clear. One possible explanation is related to the treatment procedures used. In the present study, a two-stepped less intensive approach was used and partners were asked to assist in the problem-solving process. Other studies indicated that a stepped procedure and a problem-solving strategy were important components of a successful weight control program (e.g., Black, 1987; Black & Scherba, 1983). Interestingly too, Brownell and Stunkard (1981), using a more traditional face-to-face counseling procedure, found no significant difference in the weight loss of subjects according to partner weight status. Another reason subjects with normal weight partners may have lost more weight, based on verbal report, may be that they did not need to compete with their partners to lose weight. Zitter and Fremouw (1978) found that competition with a weight-reducing partner was not advantageous to subjects in losing weight. Further research is needed to ascertain the reasons for the weight loss advantage of subjects with normal weight partners. Several important issues for the provision of service are raised by this study. When body changes of subjects are examined alone, the changes for subjects in the normal weight partner group are superior and approximately one-third greater than subjects in the overweight partner group. When body changes of subjects and partners are considered together, however, the outcome and conclusions are quite different. From this perspective, couples programs are a cost-effective approach because subjects as well as partners lose weight without any additional expenditure of human and economic resources. Couples programs may also provide an important alternative from a public health perspective because more people lose weight if partners are involved than if participants receive treatment alone. Losing weight together becomes even more appealing because oveweight partners lose as much weight as subjects. The amount that partners lost is impressive and is comparable to the average weight loss reported in intensive behavioral programs (cf. Black et al., 1989; Brownell & Jeffery, 1987; Hovel1 et al., 1989; Stuart et al., 1981). The incidental weight loss of partners regardless of their weight status, and the 14-pound weight loss of overweight partners using minimal instruction and counselor assistance, also reaffirms the efficacy of a minimal intervention program (cf. Black et al., 1984; Black & Friesen, 1983). The reason partners were able to lose weight using a minimal intervention program may be because they felt they were receiving a “bargain” by not having to pay for additional services and, therefore, their expectation of benefit increased (Stanton, 1976). Partners also participated to a lesser degree than subjects which may have decreased their feelings of therapeutic demand (Loro, Fisher, & Levenkron, 1979) and increased their feelings of perceived freedom and intrinsic motivation (Lepper, 1981; Neulinger, 1981). Future research is needed to identify what motivational and weight loss methods partners use to lose weight and why some partners lose weight and others do not. This study has heuristic value for researchers and clinicians. The weight status and weight loss of partners seem to be factors that may have affected outcome in previous couples studies. Weight status of partners might, therefore, be considered a blocking variable or covariate and body changes of both subjects and partners should be reported in couples studies in order to understand more fully the relationship between partner weight status and subject weight loss. The potential for long-term effectiveness of couples studies is still in question (Black et al., 1989) but partner weight status and weight loss seem to be related to subjects’ weight loss and may affect maintenance as well. In addition, when both subject and partner weight loss are considered, the possible cost advantage and public health benefit of
288
DAVID R. BLACK and WILLIAM
E. THRELFALL
expending the same human and economic resources to serve more people may be a valuable and, as yet, untapped advantage of couples programs for weight control.
REFERENCES Allsen. P.E.. Harrison, J.M., &Vance, B. (1984). Fitnessfor /$e: An individurrli~edrrpprouch (3rd ed.). Dubuque. IA: William C. Brown. Black, D.R. (1986. November). Problem solving in a stepped-approach for weight control. In A.M. Nezu (Chair). Social problem solving: Recent advances in theory, research, and clinical applications. Symposium conducted at the meeting of the Association for Advancement of Behavior Therapy, Chicago, IL. Black. D.R. (1987). A minimal intervention and a problem-solving program for weight control. Cognitive Therap) and Research, 11, 107-I 19. Black, D.R., Coe. W.C., Friesen, J.G., & Wurzmann. A.G. (1984). Minimal intervention for weight control: A cost-effective alternative. Addictive Behaviors, 9, 279-285. Black, D.R., & Friesen. J.G. (1983). Deposit money: A component in a self-directed minimal intervention program for weight control. Behavior Themop?, 14. 333-340. Black, D.R.,Gleser, L.J., & Kooyers, K.J. (1989). A meta-analytic evcdurztion of‘ couples byeight 10s~ programs. Manuscript submitted for publication. Black, D.R., & Hultsman, J.T. (1989). The Purdue Stepped Approach Model: h4arsholling existent communit\ health resources to reduce cardiovnscular risk factors. Manuscript submitted for publication. Black, D.R., & Hultsman, J.T. (1989). The Purdue Approach Model: A heuristic application to health counseling. The Counseling Psychologist, 16, 647-667. Black, D.R.. & Scherba, D.S. (1983). Contracting to problem solve versus contracting to practice behavioral weight loss skills. Behavior Therapy, 14, 100-109. Black, D.R., & Threlfall, W.E. (1986). A stepped approach to weight control: A minimal intervention and a bibliotherapy problem-solving program. Behavior Therap!. 17, 144-l 57. Brownell, K.D. (1982). Obesity: Understanding and treating a serious. prevalent. and refractory disorder. Journnl c>fConsulting and C/in&d PsJchology, 50, 820-840. Brownell, K.D., & Jeffery, R.W. (1987). Improving long-term weight loss: Pushing the hmits of treatment. Behavior Therap?, 18, 353-374. Brownell, K.D.. & Stunkard, A.J. (1981). Couples training pharmacotherapy, and behavior in the treatment of obesity. Archir,es of General Ps,vchiutry, 38. 1224-1229. Bruning. J.L., & Kin&. B.L. (1977). Computotionnl handbook of .stntistics (2nd ed.). Glenview. IL: Scott. Foresman. DeLucia. J.L.. Black, D.R., Loughead, T.A.. & Hultsman, J.T. (in press). Purdue Stepped Approach Model: Groups as a symbiosis of career development and mental health counseling. Journal of Career Development. Dubbert. P.M.. & Wilson, G.T. (1984). Goal-setting and spouse involvement in the treatment of obesity. Behnviour. Research, & Therapy, 22, 227-242. Elder, J.P.. Hovell. M.F., Lasater, T.M., Wells, B.L., & Carleton. R.A. (1985). The application of behavior modification to community health education: The case of heart disease prevention. Health Educarion Quarterly, 12. 151-168. Farquhar. J. W., Fortman, S.P., Maccoby, N., Haskell, W.L., Williams, P.T., Flora, J.A., Taylor. C.B., Brown, B.W., Solomon, D.S.. & Hulley, S.B. (1985). The Stanford Five-City Project: Deisgn and methods. American Journal of Epidemiology, 122, 323-324. Hall, S.M.. Baas. A., & Monroe, J. (1978). Continued contact and monitoring as follow-up strategies: A long-term study of obesity treatment. Addictive Behaviors. 3, 139-147. Hanson, R.W., Borden, B.L., Hall, S.M., & Hall, R.G. (1976). Use of programmed instruction in teaching self-management skills to overweight adults. Behavior Therupy, 7, 366-373. Hays. W.L. (1981). Statistics. New York: CBS College Publishing. Hovell. M.F.. Black, D.R.. Dellinger. A., Hofstetter, R., & Mott, M. (1989). Weight loss maintenuncet A.s.se.ssment of beholaiorol rrnining am/ very Ion cnlorie mod$ed @sting diets. Manuscript submitted for publication. Hultsman. J.T., Black, D.R., Seehafer, R.W., & Hovell, M.F. (1987). The Purdue Stepped Approach Model: Application to leisure counseling service delivery. Therapeutic Recreation Journd, 21(4). 9-22. Jeffery. R.W. (1987). Behavioral treatment of obesity. Annals of Behavioral Medicine, 9. 20-24. Lepper, M.R. (1981). Intrinsic and extrinsic motivation in children: Detrimental effects of superfluous social control. In W.A. Collins (Ed.), Aspects of the development of competence: The Minnesota Symposium on Child Psychology (Vol. 14). Hillsdale, NJ: Lawrence Erlbaum. Loro. A.D., Fisher, E.B., & Levenkron, J.C. (1979). Comparison of established and innovative weight-reduction treatment procedures. Journal @Applied Behavior AncdFsis, 12, 141-155. McNemar, Q. (1969). Psychological statistics (4th ed.). New York: John Wiley & Sons. Metropolitan Life Insurance Company. (1959). New weight standards for men and women. Staristical Bulletin, 40, l--1. National Institutes of Health Consensus Development Panel on the Health Implications of Obesity. (1985). Health
Partner weight status and subject weight loss
289
implications of obesity: NIH consensus development statement. Annals ojlnrernal Medicine, 103, 1073-1077. Neulinger, J. (1981). The psychology of leisure(2nd ed.). Springfield, IL: Charles C. Thomas. Rogers, T., Mahoney, M.J., Mahoney, B.K., Straw, M.K., & Kenigsberg, MI. (1980). Clinical assessment of obesity: An empirical evaluation of diverse techniques. Behavioral Assessmenf, 3, 161-181. Rubinson, L., & Neutens, J.J. (1987). Research techniques for the health sciences. New York: Macmillan. Stanton, H.E. (1976). Fee-paying and weight loss: Evidence for an interesting interaction. American Journal of Clinical Hypnosis, 19, 47-49. Stuart, R.B., Mitchell, C., & Jensen, J.A. (1981). Therapeutic options in the management of obesity. In L.A. Bradley & C.K. Prokop (Eds.), Medical Psychology: A new perspective (pp. 321-3.53). New York: Academic Press. U.S. Department of Health and Human Resources. (1984). Health, Unired Stares, 1984. Washington, DC: U.S. Government Printing Office. Yates, B.T. (1978). Improving the cost-effectiveness of obesity programs: Three basic strategies for reducing the cost per pound. International Journal ojObesiry, 2, 249-266. Yates, B.T. (1985). Cost-effective analysis and cost-benefit analysis: An introduction. Behavioral Assessmenf, 7, 207-234. Zitter, R.E., & Fremouw, W.J. (1978). Individual versus partner consequation for weight loss. Behavior Therapy, 9. 808-813.