Changes in health-related quality of life in people with morbid obesity attending a learning and mastery course. A longitudinal study with 12-months follow-up

This was a study where patients with morbid obesity on the waiting list for bariatric surgery were given the option of continuing on the waiting list by enrolling in a mandatory educational lifestyle course. Participants were recruited at three different geographical sites on the first day of 10 courses held in the spring of 2009 in the eastern part of Norway. Data were collected at baseline and at 12-months after course completion. At baseline, the participants answered the questionnaire on the first or second day of the education course in a secluded room on-site and returned it in a sealed envelope. The project representative collected the envelopes. Data collection at the follow-up point was done by mailed questionnaires, using stamped and addressed return envelopes. One reminder was sent out to non responders. All 185 persons attending the course were given verbal and written information about the study and invited to participate;142 consented (77%). Course attendance was the only defined criterion. The mean age among those who consented (42.5 years, SD = 10.4) was not significantly different from the age of those who did not (n = 43, M = 44.2, SD = 9.1, t = −.98, p = .33). The proportion of women among the participants (70.4%) did not differ from the proportion in rest of the population (60.5%, χ² 1.51, p = .22). At the 12-month assessment, 71 questionnaires were returned, and of these had 69 (48.6%) valid responses at baseline and 12-months follow-up. Missing scores in the indexes were tolerated up to 20% and were replaced by the individual’s mean response on the index.

The Regional Medical Research Ethics Committee of Norway (REK S-08662c 2008/17575) and the Ombudsman of Oslo University Hospital approved the study. Informed written consent was received from all participants.

Data were analyzed using PASW for Windows (version 18.0; SPSS Inc., Chicago, IL). Independent and paired-samples t- tests were used to analyze continuous variables. Ordinal and categorical data were analyzed using chi-square (χ ² ) and Fisher’s exact test. Cronbach’s α was performed on baseline data to assess the internal consistency of the scales.

To determine the predictive values of the independent variables for HRQoL, we used GLM (generalized linear models) for repeated measures with EQ-VAS scores at baseline and at 12-month follow-up as the outcome. Predictors were entered into the equation according to the theoretical model of the study, which represented the four components of the ICF [25], and the following predictor variables were included in the analyses: Body functions and structures, sex; Activities and participation, work status; Environmental factors, social support; Personal factors, self-efficacy. The general rule recommends that about 15 participants per predictor variable are needed for a reliable model. Because of the high attrition rate and therefore low sample size (n = 69) in the study, the greatest number of predictors in the GLM was restricted to four. This calculation of predictors indicated that a minimum sample size of 60 participants was necessary. The level of significance was set at p< .05. All tests were two-tailed. Cohen’s d was used to calculate effect size [31].

The mean age of the participants with complete data at 12-month follow-up (43.4 years, SD = 10.2) did not differ from that of participants from the baseline sample who consented but were excluded because of missing scores (n = 73, M = 41.6 years, SD = 10.6, t = −1.03, p = .31). The proportion of women in the study sample (n = 51, 73.9%) did not differ from those with missing scores (n = 49, 67.1%, χ ² = .78, p = .46). Significant differences were revealed with employment status, where 63.8% of the participants in the study group were in paid work, compared with 40.3% (χ ² = 7.79, p = .01) in those who were excluded.

When we compared these groups with regard to their respective scores at baseline, we found no statistical differences in self-efficacy or EQ-VAS scores (data not shown).

Baseline characteristics of sociodemographic variables and personal factors of the sample are presented in Table 1. The average age of the participants was 43.7 years and 73.9% were women. There were no sex differences on any baseline characteristics of the participants. At 12-month follow-up, 72.5% (n = 50) had undergone bariatric surgery, and 7.5% (n = 5) conservative treatment. Fourteen participants did not report on the health condition question.

The strengths of our study are the longitudinal design and the inclusion of social and personal variables that have rarely been studied. However, several limitations must be addressed. The absence of objective measures such as weight and BMI prevented us from controlling for this variable. However, previously published studies report no or only a weak relationship between BMI and mental components of HRQoL in obese people [40, 41]. Further, a body weight procedure for the participants was considered to put emphasis on dieting above lifestyle choices, and thus interfere with the intended psychosocial learning processes.

A further limitation might be that social support was assessed by the exclusive reliance on participants’ self-rating on a single item (i.e., evaluation of support from close relationships). Subsequent studies could include more comprehensive measures of social support.

To motivate patients for follow-up was difficult, despite one remainder, and consequently, we lack complete data on 51% of the participants at 12-month follow-up. Low response rates are common in most longitudinal studies with self-reported questionnaires [28], and high rates of attrition are also reported in longitudinal obesity studies [42]. Attrition analyses found that our participants reported higher participation in paid work than those with missing scores, but no other differences were, however, found between the two groups. Thus, it seems reasonable to assume that findings in the present study are representative of the healthiest of those who seek medical treatment for severe obesity. It has been suggested that the effect of obesity on HRQoL might be a main reason to seek treatment [43]. Consequently, morbidly obese who are on the waiting list for treatment may be a self-selected sample with reduced HRQoL, and participants in the present study might not be representative of all obese persons. Despite these possible shortcomings, our results contribute to the limited body of research and knowledge of factors predicting HRQoL in morbidly obese people who seek treatment.