Abstract
Gastric bypass procedures (GBP) are becoming more commonly utilized for the treatment of morbid obesity. However, the range of sustained, long-term weight loss following the procedure varies considerably. An extraordinarily high incidence of comorbid anxiety, mood, and other psychiatric disorders exists among GBP recipients. Relatively little is known about how best to meet the postoperative psychological needs of this population, and standardized treatment protocols are lacking. GBP support groups represent one forum to improve long-term weight loss outcomes during the postoperative psychological recovery period. The purpose of this study was to report descriptive findings on factors associated with attendance of postoperative support groups and to identify variables common among group participants that influence outcomes. Using a combination of self-report measures and validated screening tools (including the Psychiatric Diagnostic Questionnaire (PDSQ) and Short Form-12 v2 (SF-12v2)), we found that approximately two-thirds of participants who underwent GBP screened positive for one or more Axis I diagnoses. Participants with no previous psychiatric treatment were significantly more likely to lose ≥30% of their preoperative weight when compared to those with a psychiatric history (OR=3.333 [CI 1.102, 10.085], p=0.0330). Those participants with a physical component summary score on the SF-12v2 that fell below the median of the group were significantly less likely to lose more than the median proportion of weight lost over time for the group when compared to those with a score greater than or equal to the median (OR=0.351 [CI 0.130, 0.947], p=0.0388).
Introduction
T
The positive physiologic impact of bariatric surgery on type II diabetes mellitus, hypertension, hypertriglyceridemia, low LDL cholesterol, sleep apnea, and obesity hypoventilation syndrome are measurable and have been well documented in the literature.3,7 However, the impact of bariatric surgery on psychosocial, behavioral, and psychological function is inconsistently reported and less clearly understood. Relatively few studies have been conducted to contribute to our understanding of the impact of psychiatric comorbid conditions on bariatric surgery outcomes, and little is known about the prevalence of such conditions in weight loss surgery candidates. 8
Currently, there are no specific, standardized, empirically tested recommendations regarding psychological treatment for patients following bariatric surgery. Postoperative lifestyle changes are necessary in order for gastric bypass procedure (GBP) patients to achieve and maintain permanent weight loss. The need for such changes represents a formidable challenge for those who have historically relied upon food as a means of coping with difficult feelings and life stressors. During the ongoing recovery from obesity and its related conditions, many patients struggle to varying degrees, with the reemergence of former maladaptive eating habits. 9 Although literature is scarce, that which does exist promotes the attendance of postsurgery support groups for bariatric surgery patients.10–12 Hildebrandt evaluated the impact of support group attendance on mood as well as weight loss, and identified a positive association with attendance and weight loss but not with mood. 12
Magro et al. reported on failed surgical outcomes, defined as <50% of excess body weight lost. 13 This study noted that 60% of participants with failed outcomes never attended nutritional follow-up, and 80% never attended psychological follow-up. Given the increasing utilization of bariatric surgery and the variability in clinical results, a better understanding of the factors that impact outcomes, both positively and negatively, is warranted. Postoperative treatment services vary significantly between bariatric surgery programs in the United States. Notable differences are those in which the emphasis is placed on psychotherapeutic support, and the extent to which mental health professionals are represented within treatment protocols. Beliefs espoused by the surgeon, and the awareness of this population's need for psychiatric follow-up among GBP providers, may influence collective perceptions about the responsibility to provide postoperative psychological services, including support groups. This study examines the relationship between postoperative support group attendance for GBP patients on weight loss and psychosocial, behavioral, and psychological functioning.
Methods
Participants
Study participants were recruited between July 5, 2013, and September 7, 2013, using the database of a bariatric surgery practice in a Boston suburb. A total of 400 potential participants meeting the inclusion criteria were randomly selected and mailed a packet that included: an invitation letter, a letter of support from the surgeon encouraging participation, a questionnaire collecting information on demographic and other self-reported variables, the Psychiatric Diagnostic Questionnaire (PDSQ), the Short Form-12 v2 (SF-12v2) to assess current HRQoL, and a self-addressed, stamped return envelope. Inclusion criteria limited the recruitment of participants to men and women who were over 18 years old at the time of data collection and had a GBP performed at the same Bariatric Surgery Center of Excellence (BSCOE) on or before June 1, 2011. Those who had gastric banding, vertical sleeve gastrectomy, and other primary, secondary, or revision procedures were excluded.
Data collection instruments
Demographic data on age, gender, marital status, physiologic and psychiatric comorbid conditions, preoperative and postoperative support group attendance, tobacco and alcohol use history, current use of medications, date of surgical procedure, self-reported preoperative weight and current weight, and psychiatric history were obtained from self-administered information forms. The self-reported PDSQ was used to assess current psychiatric symptoms The PDSQ is a validated measure that contains 111 yes/no questions regarding the presence of symptoms of 13 DSM-IV Axis I disorders. 14 The presence of each disorder was determined based on specific PDSQ subscales for each, all of which have 89% sensitivity and 97% negative predictive value. These subscales have been found to be valid for assessing the 13 DSM-IV disorders included on the screener. An individual screened positive for a disorder if he/she endorsed the minimal number of symptoms for that diagnostic category. A positive screening result indicates that the individual would be more likely to qualify for a diagnosis of that disorder than someone who did not screen positively.
HRQoL was assessed using the SF-12v2. The SF-12v2 measures physical and mental health by means of two summary scores: a physical component summary (PCS) and mental component summary (MCS). The SF-12v2 uses norm based scoring so that results can be generalized or compared to other health-based surveys. Standardized scores are calculated using a mean of 50 and a standard deviation of 10, where higher scores indicate better health.
Completed questionnaires were anonymous and were assigned a randomly generated code number upon data entry. They were stored in a locked cabinet in the researcher's office and destroyed following data analyses. The human rights and privacy of participants were vigilantly upheld throughout the study. IRB approval was granted by the Institutional Review Board of the UML Office of Institutional Compliance on June 11, 2013, for the period June 11, 2013, to June 10, 2014.
Outcome measures
Primary outcome measures included long-term, sustained weight loss and postoperative support group attendance. Long-term, sustained weight loss was examined in three ways: (1) the difference between preoperative weight and current weight; (2) the difference between preoperative weight and current weight over the time since surgery; and (3) the proportion of weight loss (calculated using the preoperative weight and current weight) over the time since surgery. Weight loss as the difference between preoperative and current weight was recorded as a binary variable depending on whether ≥30% of preoperative weight was lost. Weight loss examined over time around the median was also recorded as a binary variable. Postoperative support group attendance was measured using a single question from the demographic information collected from all participants.
Statistical analysis
A descriptive, correlational design was used to investigate and report on quantitative data. A retrospective, nonexperimental approach was used to identify the relationship between demographic and other variables to understand better the complex association between attendance of support groups and other psychological factors after GBP and weight loss.
Several independent variables were examined and included the following: postoperative support group attendance (where sustained, long-term weight loss was the outcome measure); age, race, marital status, education, and occupation reported as categorical variables; and gender as a binary categorical variable. Psychiatric comorbid conditions were recorded as a categorical variable, measured as the number of disorders positively screened for at the time of participation. HRQoL was recorded using both the PCS and MCS as binary variables spilt at the median to increase the power of the study.
The primary outcome measure of long-term, sustained weight loss in both bivariate and multivariate analyses were analyzed. The relationship between potential risk factors and the primary outcome variable was assessed using simple logistic regression models with odds ratios (ORs) and 95% confidence intervals (CIs). Two-tailed significance levels for all analyses were set at p<0.05, and calculations were performed using SAS v9.3. 15
Results
The sample population consisted of 71 participants from the 400 invited to participate who completed and returned all recruitment materials (17.75% response rate); most were between the ages of 45 and 64 years (73.2%; Mage=55 years), and the majority (n=58; 81.6%) were women (nine participants were male; gender was unreported for four participants). Participants were predominantly white (94.4%), married (53.5%), had some college education or a college degree (76.1%), and were employed at the time of study participation (56.3%). Average preoperative weight was 286 pounds, and average current weight was 186 pounds, giving an average difference of 100 pounds. The average amount of time between GBP and study participation was 6 years. Table 1 shows these and other unadjusted demographic characteristics in relation to long-term, sustained weight loss measured three ways.
Note: Two participants did not report their GBP date, so weight loss over time was only calculated for 69 of the 71 participants. Not all participants reported on certain variables, so some participants were excluded from analyses where their information was missing.
GBP, gastric bypass procedure; OR, odds ratio; CI, confidence interval.
When examining long-term, sustained weight loss as the difference between preoperative weight and current weight, 51 participants (71.8%) had lost ≥30% of their preoperative weight. Two participants did not report their GBP date, so weight loss over time was only calculated for 69 of the 71 participants. When examining the outcome as the difference between preoperative weight and current weight over the time since surgery, 30 of the 69 participants (43.5%) had lost more weight than the median (20) of the group, and when examining the outcome as the proportion of weight lost (calculated using the preoperative weight and current weight) over the time since surgery, 34 of the 69 participants (49.3%) had lost more weight than the median (0.07) of the group. Sixty-five participants (91.5%) reported attending a preoperative support group, while 59 (83%) reported attending a postoperative support group.
Comorbid psychiatric conditions
Forty-eight of the 71 participants (67.6%) screened positive for one or more Axis I diagnoses on the PDSQ assessment, with 19 (26.8%) screening positive for four or more. Although no results were statistically significant due to the small sample size, those screening positive for one, two, or four or more Axis I diagnoses on the PDSQ were less likely to demonstrate long-term, sustained weight loss. Individual Axis I diagnoses were not examined for this study. Table 2 shows the unadjusted relationship between the number of Axis I diagnoses positively screened for on the PDSQ and long-term, sustained weight loss measured three ways.
Note: Two participants did not report their GBP date, so weight loss over time was only calculated for 69 of the 71 participants. Not all participants reported on certain variables, so some participants were excluded from analyses where their information was missing.
PDSQ, Psychiatric Diagnostic Questionnaire.
Prior psychiatric treatment was one of only two variables that were found to be statistically significant in this study. Participants who had never received psychiatric treatment were more likely to lose ≥30% of their preoperative weight at the time of recruitment (OR=3.333 [CI 1.102, 10.085], p=0.0330). The same effect was found for both of the other weight loss outcomes, but neither was found to be statistically significant. The other statistically significant variable was the PCS score. Those participants with a score below the median of the group were less likely to lose above the median proportion of weight over time for the group (OR=0.351 [CI 0.130, 0.947], p=0.0388). Again, the same effect was found for both of the other weight loss outcomes, but neither was found to be statistically significant. These variables and others, including having ever received a psychiatric diagnosis, MCS scores, and postoperative support group attendance, in relation to all three weight loss outcomes can be found in Table 3.
Note: Two participants did not report their GBP date, so weight loss over time was only calculated for 69 of the 71 participants. Not all participants reported on certain variables, so some participants were excluded from analyses where their information was missing.
HRQoL, health and quality of life.
HRQoL
Postoperative support group attendance was associated with higher physical HRQoL scores when using postoperative support group attendance as the outcome and physical HRQoL as a binary predictor variable, spilt above or below the median for the group (OR=1.575 [CI 0.427, 5.813], p=0.4950). Postoperative support group attendance was associated with lower mental HRQoL scores when using postoperative support group attendance as the outcome and mental HRQoL as a binary predictor variable, spilt above or below the median for the group (OR=0.639 [CI 0.184, 2.222], p=0.4811).
Multivariable analyses
Similar associations with long-term, sustained weight loss were found in multivariable logistic regression models as were found in bivariate analyses. As seen in bivariate analyses, participants who had received psychiatric treatment, screened positively for four or more Axis I diagnoses on the PDSQ, and were between the ages of 35 and 44 years were less likely to have long-term, sustained weight loss. Those having a PCS score above the median for the group were more likely to have long-term, sustained weight loss.
Discussion
This study documented factors associated with postoperative support group attendance, psychiatric comorbid conditions, and HRQoL in a cohort of individuals >2.5 years after undergoing gastric bypass surgery. A high rate of psychopathology was found. The results of this study are generally consistent with findings from other studies. Kalarchian et al. reported a lifetime prevalence of 42% for depressive disorders and 37.5% for anxiety disorders in patients seeking bariatric surgery. Anxiety disorders were reported as being most prevalent at the time of the preoperative evaluation. 8 Psychiatric disorders in bariatric surgery patients have been associated with a higher body mass index (BMI) and lower functional health status and HRQOL. Those individuals with a history of mood or anxiety disorders compared to those with no psychiatric history have been associated with a smaller decrease in BMI during the first 6 months following surgery. 16
In this study, when examining the relationship between psychiatric treatment and long-term, sustained weight loss, those who never received psychiatric treatment (and therefore presumably without a history of psychiatric comorbid conditions) were more likely to lose weight. Those with higher PCS scores were also more likely to lose weight. The positive association between preoperative and postoperative functional health status, BMI, HRQoL, and psychological status has been documented. 16 While obesity is classified as a medical disorder, there is general consensus that psychological factors contribute to its etiology. Findings about how such factors influence the development of obesity and surgical treatment outcomes are inconsistent and inconclusive. 17
Similarly, our finding that those who screened positively for four or more psychiatric disorders were less likely to lose weight is in line with other research, which has found a graded association between the number of psychiatric diagnoses and weight loss. Rutledge et al. reported that patients were six times more likely to regain weight or lose no more additional weight after one postoperative year when two or more psychiatric conditions were present, as compared to patients without psychiatric conditions. 17
No statistically significant findings as they related to postoperative support group attendance were found in this study. However, previous research has found that disturbed eating patterns and psychological difficulties are seen following surgery, and the causal relationship between such emergent psychological difficulties and support group attendance has not been established. 10 It is unclear whether those with maladaptive postoperative behaviors are more likely to attend support groups as a means of seeking treatment for comorbid psychiatric conditions. The goals of postoperative support groups differ from that of individual psychological care. Generally speaking, postoperative groups for patients who undergo GBP provide support and psychoeducation to facilitate requisite lifestyle changes. It has also been reported that patients who attend group meetings regularly tend to lose more weight. 12
Other research has found that attendance at postoperative support groups was one of the strongest predictors of success following GBP, along with surgeon follow-up. 18 The current study did not address attendance frequency or the size and location of the support groups. However, all groups attended by study participants were exclusive and different from the mandatory informational meetings provided by the study site.
Limitations
Our study had several limitations. First, the small sample size used in this study impacts the power of the analyses, affecting the associations, generalizability, and interpretation of the findings. Second, the data used for this study were self-reported and relied solely on the integrity of the participants' responses. This study did not look at the group content, size, or frequency and duration of attendance, which may affect the outcome and should be considered in future work. Lastly, this study was susceptible to selection bias due to the convenience sample used. The nonrandomized design resulted in an insufficient representation of participants, creating a lack of comparable controls. It was not possible to identify a causal relationship between the variables studied and the outcomes.
Implications for future research
Despite a lack of statistically significant findings in this study, the results speak to trends in the data and suggest areas for further research. Descriptive, correlational studies powered by large, diverse (random) samples that can be replicated are needed. Further exploration of the graded association between the number of comorbid psychiatric conditions and surgical failure will help to determine the significance of the relationship between the two. Qualitative studies to learn more about which aspects and types of support groups are most helpful are needed. Studies designed to identify characteristics associated with highest risk for treatment attrition and weight recidivism are indicated to increase the long-term success rates following GBP procedures.
Conclusions
This study examined trends of long-term, sustained weight loss in a population of GBP recipients >2.5 years post procedure. Significant relationships between the receipt of prior psychiatric treatment and lower PCS scores were found to be related to poorer long-term, sustained weight loss in this group. These findings suggest that there is need for further investigation of the relationship between comorbid psychiatric conditions and successful weight loss outcomes following GBP. A more comprehensive understanding of the impact of one's psychiatric history on surgical outcomes might inform the development of evidence based standardized psychological treatment protocols. Furthermore, patients who have been treated for multiple psychiatric conditions should be strongly encouraged to attend postoperative support groups and to seek other relevant forms of treatment such as individual counseling and psychopharmacological services. Early identification of psychiatric comorbidities could improve weight loss outcomes by introducing and reinforcing the need for ongoing postoperative psychological treatment preemptively during the preoperative phase.
Footnotes
Acknowledgments
The authors would like to acknowledge Louise Pothier, BSN, RNFA, CBN, CNOR, and Karen M. Flanders, MSN, CBN, NP-C, for their generous support during the course of this project.
Author Disclosure Statement
No competing financial interests exist.