Effectiveness of a Peer Support Intervention Program on Obesity Control Among Women in a Rural Area of Turkey


 BackgroundObesity in women is one of the leading public health problems globally. Peer support interventions have been effective in many areas of health promotion, and they have also been successful in obesity prevention and treatment.MethodsThis is a quasi-experimental obesity peer-led intervention study called the Leading Woman Model. At baseline, the obesity prevalence was found to be 60.5% among women aged 18–64 (n = 655) living in a rural district of Turkey. Of the participants (n = 137), 86.9% completed the 3rd month and 78.1% the 6th month of the intervention. Leading women (n = 11) were recruited from the community to supervise and monitor their own target groups of women during the intervention, which included supervised balanced nutrition and regular physical activity. ResultsThe mean age of the target intervention group was 42.8 ± 12.2. Significant improvements were observed in the body weight of the participating in the 3rd (-1.15 ± 2.51 kg) and 6th month (-1.13 ± 4.15 kg) of the intervention (p <0.05). Of the women, 10.9% lost at least 5% of their weight by the 3rd month and 13.1% by the 6th month of intervention, and 8.4% and 11.2% of the women achieved a better BMI category in the 3rd and 6th month of intervention, respectively.ConclusionCommunity-based obesity interventions are challenging but much more promising than those based at a facility. We suggest the Leading Women Model for community-based obesity interventions in women.

supported interventions that have been carried out in a clinical context (7,8). The main difference between community-based and hospital-based trials is the target population's accessibility to health services.
Our aim was to develop a new peer-based obesity intervention model that could be integrated into routine primary care services. We set out a semi-structured model that aimed to help women lose weight with the help of their peers and services readily provided by all primary health-care organizations in our country.
Therefore, the purpose of this research was to reduce the Body Mass Index (BMI), fat ratio, and fat amount of the targeted women through the support of their peers (leading women) by regulating nutrition and improving physical activity in a rural community.

Study design
This quasi-experimental eld intervention study was conducted in a rural district of Manisa province in Turkey. The peer-led training intervention model used in this study was named the Leading Woman Model.
The study consisted of two main stages: population weight screening and eld intervention. To start with, all women living in the district aged 18-64 (n = 655) underwent weight and height measurement. Of them, 396 women were found to be overweight or obese and subject to intervention. Forty women were excluded from the study due to medical conditions, and 219 women did not join the intervention program after being informed that it would include nutrition and exercise requirements. Eventually, 137 women were accepted for participation in the intervention program, which is the second stage of the study. The ow diagram of the study stages is presented in Figure 1.

Subjects
The women were either national or international migrants, with the majority being Slavic Muslims known as Pomaks. Two distinct participant groups were identi ed: Leading Women: Leading women (n = 11) were chosen to act as trustworthy mentors and guides for the women. They were recruited from the community based on the discussions of a panel of community health professionals including staff at a family health center and local pharmacists.
Target women: Overweight and obese women with a BMI range of 25.00 to 39.99 (n = 126) were the targets of the intervention. They were all willing to participate in the peer intervention program in the district.
Assigning the target group to the leading women Each of the 11 leading women were asked to select the women with whom they had social contact, and preferably those who lived on neighboring streets or roads, to enable good communication and easy access. The range of the number of women assigned to a leading woman was 5 to 17.

Intervention infrastructure
Of the 137 women who agreed to participate in the intervention, 86.9% (n = 119) completed the rst 3 months of the intervention and 78.1% (n = 107) completed its 6-month target.
The intervention infrastructure was divided into three main items.

Anthropometric and biometric device preparation
Each leading woman was given electronic scales sensitive to 100 grams to measure the weight of the women in her group. A stadiometer already present at the family health center was used to measure their height. Body analyses of the women, such as for body fat mass, body fat ratio, trunk fat mass, trunk fat ratio, body muscle mass, and body fat-free mass, were recorded using a bioelectrical impedance analyzer (InBody 230). A pedometer was distributed to all participants during the intervention period.

2.Questionnaire battery
This battery consisted of the women's demographic, anthropometric, and biometric characteristics and their medical conditions, exercise status, and self-assessment of compliance with the nutrition program.
The battery also included three supplementary scales, namely the International Physical Activity Questionnaire (IPAQ), to assess the women's physical activity; Attitudes Towards Obese People (ATOP), to measure their attitudes towards obesity; and the Hospital Anxiety and Depression Scale (HADS), to evaluate the depressive mood of the women. IPAQ classi ed respondents into three activity subcategories (high, moderate, and low) (9). ATOP scores were treated as a continuous variable in the analysis: the higher the score, the better the attitude towards obese people (10). HADS scores were recoded as dichotomous variables with a cut-off value of 7.0 (11). The HADS and IPAQ forms were lled in at baseline and at the 3rd and 6th month of intervention. ATOP was only tested at baseline.

Leading Woman Training
The leading women were initially given a 3-day training courses on obesity, nutrition, and physical activity and exercise. The courses were conducted by a community health dietitian and exercise experts working at Celal Bayar University.

Intervention procedure
The intervention lasted a total of 6 months ( Figure 1). The inclusion criteria were having a BMI range of 25.00-39.99 kg/sqm and volunteering for the intervention. The exclusion criteria were pregnancy, breastfeeding, physical or intellectual disability (including cancer, post-myocardial infarct, multiple sclerosis, and chronic neurologic diseases), morbid obesity (BMİ ≥40.00 kg/sqm), and unexplained tachycardia.
Baseline height measurement, waist and hip circumference measurements, and bioelectrical impedance analysis were done for each woman. The forms were distributed to the target women and leading women following the baseline measurements. Each woman was given a standard diet list and asked to report her nutrition practices daily through the forms.
The three main tasks of the leading women were: 1) individual and group interviews (at least once a month) with the women under their guidance, 2) organizing regular neighborhood exercise sessions, and 3) keeping records.
At the 3rd month of intervention, we measured the waist and hip circumference and weight of the target women and applied the IPAQ and HADS depression subscales. The intervention ended at the 6th month of the intervention period.

Outcomes of the intervention
The outcomes of the intervention were classi ed into two groups: 5. Body muscle mass (trunk plus extremities) The independent variables of this study were sociodemographic variables; physical activity (by metabolic equivalent of task (MET) score), attitude to obesity, and depressive mood; and health and body image perception, family history of obesity, previous weight loss attempts, previous physical activity practices, fertility history, and having any chronic illness.

Results
The mean age of the intervention group (n = 126) was 42.8 ± 12.2. The sociodemographic characteristics of the target intervention group are shown in Table 1. Of the women, 72.2% perceived themselves to be overweight/obese and 81.7% were not satis ed or hesitant about their body weight before the intervention. About half of the women had attempted to lose weight at least once previously.
Baseline data revealed that the prevalence of overweight plus obesity (BMI ≥25.00) was 68.9% (n = 637), including 43 morbid obese (BMI >40.00). IPAQ results showed that 4.8%, 45.2%, and 50.0% women reported high, moderate, and low physical activity scores, respectively, prior to the intervention. The mean ATOP scale score was 58.55 ± 16.4, and the HADS assessment indicated a 31.7% prevalence of depressive mood.
The median age of the leading women was 44.0 (min 29.0, max 62.0); 50.0% of this group were primary, 33.3% secondary, and 16.7% high-school graduates.
Signi cant improvements were observed in weight and waist circumference of the women at the 3rd and 6th month of the intervention. The mean weight loss was 1.15 ± 2.51 kg, and waist circumference decrease was 0.96 ± 2.53 cm at the 3rd month of the intervention (p <0.05). At the 6th month of the intervention, the mean weight loss was 1.13 ± 4.15 kg and waist circumference decrease was 1.30 ± 2.53 cm. Of the women, 10.9% lost at least 5% of their weight at the 3rd month and 13.1% did so at the 6th month of the intervention, and 8.4% and 11.2% moved to a better BMI category at the 3rd and 6th month of the intervention, respectively ( Table 2). The mean body muscle mass increase was 0.08 ± 4.51 (p = 0.021), and the mean body fat mass decrease was 1.73 ± 4.02 (p <0.001; Table 2).
The physical activity level of the women, in terms of MET score, was signi cantly increased, from 867.2 ± 798.5 to 1445.4 ± 1444.6, in the rst period (baseline-3rd month; p <0.001), but it did not change in the second period (3rd-6th month) of the intervention (p >0.05). The rate of depressive mood was 37.4% at baseline, 18.7% at the 3rd month, and 15.9% at the 6th month of intervention (p <0.001).
The second part of the results presents the causality between anthropometric changes and the independent variables.
Average weight loss between baseline and the 3rd month of intervention was only sensitive to working status, health insurance coverage, residency/migration status, and spouse's education. The remaining independent variables were not sensitive to weight loss in the rst 3 months of the intervention (Table 3). Table 4 presents the independent variables that would explain the effect of the intervention on weight loss of at least 5% and BMI category reduction at the 3rd month of the intervention.
The percentage of women who lost at least 5% weight was higher among secondary school graduates than primary school graduates. There was no signi cant relationship between weight loss of 5% and BMI category reduction and other independent variables (Table 4).
A statistically signi cant dose-response relationship was found for BMI category reduction and ATOP score (p = 0.049). On the other hand, strati ed analysis of ATOP showed that signi cant weight loss was observed among those women who had a positive attitude towards obesity (p = 0.002).
Physical activity in terms of MET score and ATOP were not found to have a signi cant relationship with weight loss (any weight loss, at least 5% weight loss, or BMI category reduction).

Discussion
The overall overweight and obesity prevalence was 68.9% in the community, which is quite high and consistent with the country data (12,13).
Peer interventions have frequently been used internationally to promote health and healthy lifestyle behaviors (5,6), including well-balanced nutrition and physical activity (7,8). Almost all community-based obesity interventions in Turkey have implemented behavioral education, diet, and exercise programs with professional support (14)(15)(16). This study has two distinctive features, namely the intervention is both peer-based and community-based.
Mean weight loss during the course of our intervention was 1.13 ± 4.15 kg. However, it was found to be 2.8 kg in an obesity intervention conducted in an urban primary care area in Turkey (16), 2.1 kg in a health promotion club in Japan (17), and 6.4 lbs among college students in the United States (19), which were all somewhat higher than our gure.
If we assume that sensible weight loss is weight loss of ≥5% body weight and/or BMI category reduction, then only about one in 10 women showed considerable weight loss at the end of the 3rd month of the intervention. In national and global obesity intervention studies, the rate of individuals who lost at least 5% body weight varied between 20% and 30.3% (16, 19,20). Our weight loss gures are lower than the previous studies, which were mostly conducted in urban populations. The higher weight loss gures in these studies compared to ours may be attributed to either the higher level of education of the urban women or their greater willingness to be involved in the interventions. In our study, overweight/obese women were invited to participate in the intervention rather than them coming deliberately to the primary health-care center. The majority of the published eld obesity interventions were institution-based. In these interventions, an overweight/obese person would deliberately go to a clinic to lose weight, while in community-based interventions, a person might not initially be willing to participate. Therefore, the observed success of the intervention-although not as good as an institution-based intervention-is more realistic, and the results can be more generalizable.
The percentage of women whose BMI category reduced in the rst 3 months of the intervention was 8.4%, and it was 11.2% throughout the intervention. BMI category reduction ranged between 20.0% and 27.9% in some other interventions (16, 21), and these were clearly higher than ours. On the other hand, the rate of BMI category reduction was found to be about 6.5% in a recent rural eld obesity intervention in Turkey based on only a public mass education campaign. This might provide good evidence of the usefulness of our peer-based intervention in rural districts (22).
The improvements of the anthropometric measurements in the rst three months of the intervention were signi cantly better than those of the second intervention period (3rd-6th month). The intervention started in the summer months, when people were more physically active with, e.g., vineyard farming activities, and lasted until the end of fall, which is the region's wet season . As a matter of fact, at the end of the 3rd month of the intervention, the mean MET scores of the women were signi cantly higher than the baseline MET scores, but they fell below the baseline value at the 6th month of the intervention. Therefore the decrease in weight loss in the second period of the intervention might be associated with slowing of physical activity during the same period.
The predictors of weight loss in this study were higher education of women, having universal health insurance, being native (not migrated from Bulgaria), being unemployed, and married with a low-educated husband. It was demonstrated in several obesity interventions that higher educated people would bene t from community-based health promotion programs more than others (23,24). Unemployed women whose husbands had a low education level may have bene ted more from the intervention because they were much more overweight at baseline. Migration itself is associated with overweight/obesity risk and nutrition transition. This may be explained by the cultural aspects of food consumption and that families of Bulgarian origin tend to consume more pastry than native families. Some literature evidence has shown that it is always more di cult for migrants to follow health promotion activities than the native population (25). Our intervention program was carried out in a rural district, and cultural issues are very important among rural and migrant populations. Gender norms and expectations that are very important in these populations may also restrict the opportunity to exercise (26).
One of the main determinants of adherence to a weight-loss regime is lack of acceptance of obesity as a health problem. A second negative demotivating factor might be nding a balance between professional/house work, and family routines, since devoting time to health and physical activity appears difficult for adults with obesity (27). Pressure by husbands on women to maintain the family's food regime may be another reason for loss of motivation. Indeed, we did not detect any signi cant weight loss in husbands during the intervention. The positive effects of family solidarity and partner support during obesity treatment have also been reported in several previous studies (28).
Finally, we found that a positive attitude to obese people helps weight reduction. In his review, Gordon Cochrane (29) claimed that self-e cacy correlates positively with success in all realms of personal endeavor, including weight loss.
There were some limitations for this study. We may assume that the means of identi cation of each leading women and selecting the target women under their supervision might be an important reason for the low rate of participation. The selection of leading women was based solely on the subjective assessment of primary health-care professionals, and no objective criteria were used. Therefore, it is questionable whether the leading women were eligible leaders or not.
Another limitation of our study is that we did not set a weight-loss goal for women at the beginning of the intervention, as had been done in some other intervention studies (30). Also, the very short duration of this intervention may mask its long-term residual effects, and further monitoring of the target women may be required.

Conclusion
Community-based obesity interventions are much more promising than institution-based ones, because they gives us the opportunity to reach passive, non-volunteer groups in communities. We suggest a very new leading women intervention approach for rural, community-based obesity interventions. *Mann Whitney-U; **Kruskal Wallis Anova Table 4 The relationship between BMI category reduction and body weight loss of at least 5%