DEVELOPMENT AND VALIDATION OF A PSYCHOLOGICAL SCALE FOR BARIATRIC SURGERY: THE BARITEST

ABSTRACT BACKGROUND: It is recommended that bariatric surgery candidates undergo psychological assessment. However, no specific instrument exists to assess the psychological well-being of bariatric patients, before and after surgery, and for which all constructs are valid for both genders. AIMS: This study aimed to develop and validate a new psychometric instrument to be used before and after bariatric surgery in order to assess psychological outcomes of patients. METHODS: This is a cross-sectional study that composed of 660 individuals from the community and bariatric patients. BariTest was developed on a Likert scale consisting of 59 items, distributed in 6 constructs, which assess the psychological well-being that influences bariatric surgery: emotional state, eating behavior, quality of life, relationship with body weight, alcohol consumption, and social support. Validation of BariTest was developed by the confirmatory factor analysis to check the content, criteria, and construct. The R statistical software version 3.5.0 was used in all analyses, and a significance level of 5% was used. RESULTS: Adjusted indices of the confirmatory factor analysis model indicate adequate adjustment. Cronbach’s alpha of BariTest was 0.93, which indicates good internal consistency. The scores of the emotional state, eating behavior, and quality of life constructs were similar between the results obtained in the community and in the postoperative group, being higher than in the preoperative group. Alcohol consumption was similar in the preoperative and postoperative groups and was lower than the community group. CONCLUSIONS: BariTest is a reliable scale measuring the psychological well-being of patients either before or after bariatric surgery.


RESUMO -
BariTest is an instrument that makes it possible to measure and analyze psychological wellbeing and directs the necessary psychological interventions, before and after bariatric surgery, contributing to the psychological assessment.

Central Message
Caution is recommended to indicate bariatric surgery in patients with severe psychiatric disorders without treatment. This is suggested when there is an absence of social support in those who, due to emotional instability, may find it difficult to follow and obey postoperative dietary instructions, and in cases of abuse of illicit drugs and/or alcoholism. significantly higher prevalence of alcohol consumption after bariatric surgery 23 , and some patients have an aggravation of the psychiatric disorder, which may worsen the patient's psychological well-being, despite weight loss 38 . Furthermore, it is important to continue the psychological follow-up after bariatric surgery because some patients do not have a favorable outcome, which can lead to depression, use of alcoholic beverages, and weight regain 8,32 . Between 20 and 30% of patients experience suboptimal weight loss or significant weight regain within the first few postoperative years. The reasons for this involve physiological, behavioral, and psychological characteristics 38 . Nowadays, no psychometric scale has been identified for which all of the instrument assesses both genders, before and after surgery, regardless of the surgical technique, focusing to assess the psychological well-being that can influence the outcome of the operation, such as severe depression, mood disorders, substance use disorder, eating disorders, psychosocial problems, or behavioral noncompliance 38 . Considering the six main psychological aspects that can influence the result of the operation 38 , BariTest was developed to compare the outcomes of psychological well-being that will emerge from bariatric surgery 20,38 .
The BariTest is a patient-reported outcome measures (PROM) psychometric scale which assesses the psychological well-being, before and after the bariatric surgery 9,38,42 , and is composed of six constructs: Of these, 598 were awaiting consultation (preoperative or postoperative) at the bariatric surgery. In addition, for validation purposes, BariTest was applied to 48 nonobese subjects in the community, who had not undergone and did not intend to undergo bariatric surgery ( Table 2). The instrument was also evaluated by a focus group (validity of content), selected as a convenience sample, composed of 10 bariatric patients who analyzed the semantic understanding of the item. Four patients did not respond to the questionnaire and were excluded from the analysis.

Validation of BariTest
The BariTest validation process was carried out through content, construct, and criterion validity. In addition, the instrument's reliability was analyzed, and the instrument's correction and interpretation table was elaborated.
After conducting a literature review and expert discussions, a preliminary version of the BariTest scale was developed.

INTRODUCTION
O besity is a chronic disease of multifactorial causes such as genetic, environmental, socioeconomic, endocrine, metabolic, and psychiatric 17 . When conventional treatments such as diet, medication, and physical exercise do not show any positive results and that obesity causes harm to the individual, bariatric surgery may be recommended 25 .
The candidates for bariatric surgery must have a body mass index (BMI) above 35 associated with a comorbidity (e.g., high blood pressure, diabetes, and hepatic steatosis, among others mentioned in Resolution No. 2,131/15 of the Federal Council of Medicine) 9 or a BMI above 40, considered morbidly obese. The American Society for Metabolic and Bariatric Surgery (ASMBS) 1 recommends that the candidates for bariatric surgery be followed up by a multidisciplinary team. In this team, the psychologist's objective is to assess the candidate's mental aptitude in order to understand the surgical procedure and the psychological aspects that can influence the result of the operation 38,42 .
Wadden and Sarwer 46 suggested that in the psychological evaluation process, 70-90% of patients are unconditionally indicated for surgery, 15-30% are referred for psychological or nutritional treatment as a prerequisite for surgery, and the remaining patients are excluded due to psychiatric reasons such as psychosis, untreated severe depression, mood disorders, eating disorders, substance use disorder, psychosocial problems, or behavioral noncompliance.
Psychological treatment should be started in the preoperative phase because the candidates for bariatric surgery have a higher prevalence of mental disorders than the general population, and psychopathological abnormalities tend to impact both the evolution of obesity and the results of bariatric surgery 29,39 . Caution is recommended to indicate bariatric surgery in patients with severe psychiatric disorders without treatment. This is suggested when there is an absence of social support in those who, due to emotional instability, may find it difficult to follow and obey postoperative dietary instructions, and in cases of abuse of illicit drugs and/or alcoholism 9,41 .
A difficulty that professionals who make psychological assessment for bariatric surgery face is the lack of specific validated instruments for this population 22 . Psychologists vary in their methods of evaluating patients before and after bariatric surgery 10 . They usually apply symptom inventories to screen for depression and eating disorders, and some psychopathology, personality, or cognitive function tests 46 . The most cited assessment instruments in the literature 16 are the Beck Depression Inventory (BDI), the Binge Eating Scale (BES), the Eating Disorder Examination, the Millon Behavioral Medicine Diagnostic (MBMD), and the Minnesota Multiphasic Personality Inventory (MMPI). These instruments were not developed with a focus on the bariatric population and the psychologist should avoid using several instruments because the patient's tiredness may interfere in the accuracy of the answers 20 . Among the instruments intended for bariatric surgery, there is only one psychological instrument validated for the bariatric population, i.e., the PsyBari, developed by David Mahony, PhD, a clinical psychologist at the Lutheran Medical Center, Brooklyn, New York 27 . Despite being practical on a Likert scale and intended to assess bariatric patients before bariatric surgery, not all items of the test were valid for both genders. This is an important characteristic as there are two different test formats for each gender and it is questioned whether it is a single instrument or whether there are two distinct instruments, bringing unnecessary complexity. In addition, PsyBari validation was not performed with post-bariatric patients, and it is known that there is a There are times when I cry a lot. 3 I find myself in a bad mood and/or irritated for no reason.

4
There are days when I wake up extremely excited and others, I hardly want to get out of bed. 5 There are times when I feel like dying. 6 I believe that I do things impulsively. 7 People say that I am anxious.

8
I have difficulty falling asleep because I feel very agitated and/or with rapid thoughts at night. 9 I do and/or say things without thinking. 10 I feel discouraged and hopeless. 11 I have bouts of tachycardia, despair, and the feeling that I am going to die. 12 I have a feeling of regret for the things I do/say. 13 I believe that I am a disappointment for my family and/or friends.
14 There are phases that I work too much and produce a lot, and in other phases, I don't feel like working, and my work doesn't produce. 15 I realize that I talk too much or speak much faster than normal. 16 When I'm eating, I lose control and end up eating too much. 17 When I feel the urge to eat it is difficult to control.

18
When I feel like eating some treats, I cannot put it off.

19
I eat a few times a day, but when I eat, I exaggerate the quantity.

20
When I have emotional problems, I use food to relieve tension or to bring me joy. 21 I have a habit of eating "fast food" (snacks). 22 I eat quickly and chew food sparingly. 23 I think about food most of the day. 24 I am a candy eater. 25 My behavior toward food causes me a lot of suffering.

26
I realize that I eat more at night. 27 I have difficulty in distinguishing between hunger and the desire to eat.

28
I eat sparingly in front of others, but then I make up for it when I'm alone. 29 I eat small amounts of food for several hours in a row (Pinch Habit Initially, the instrument had 99 items. Content validity was performed by assessing seven specialists in bariatric surgery or psychology, and all items were evaluated (Appendix 1). The anonymity of the evaluators was maintained, and each committee member individually determined their agreement on whether each item should remain in BariTest, using a four-point Likert scale: 0= Very Bad, 1= Bad, 2= More or less, 3= Good, 4= Great. At the end of this assessment, the experts carried out a qualitative analysis and offered suggestions for improvements. Items that had a mean of less than 3.5, or that were considered irrelevant to the objective by at least two members of the expert committee, were removed from the instrument (Appendix 3). Thus after this analysis, 40 items were excluded and BariTest completed with 59 items (Table 1). Also a focal group analyzed the understanding of each item, and no items were excluded by this group.
The validity of construct was performed by confirmatory factor analysis (CFA) (Appendix 2). The fitted CFA model was evaluated through the indices 6,7,18,43,47 such as standardized root mean squared residual (SRMR), root mean of the squares of the errors of approximation (RMSEA), comparative fit index (CFI), and Tucker-Lewis index (TLI).
The validity of criterion was performed to ascertain the accuracy of the instrument, by means of stability in equivalent forms of different tests 44 . To determine responsiveness, an analysis of the receiving operating characteristic (ROC) curve was performed, verifying accuracy through sensitivity and specificity (Figure 2). At the time of applying BariTest, 175 patients also received two other questionnaires: the WHODAS 2.0 (Annex 1), which is a self-administered questionnaire that measures functionality and disability related to any disease or health status, avoiding the researcher's bias, and the Obesity -related Problems Scale (OP) (Annex 2), which is a scale of outcomes reported by patients that measures the impact of excess weight on psychosocial functioning. These instruments were chosen because they have been validated 4,5 for the Brazilian population with obesity to measure psychological well-being.

Reliability
Reliability was calculated using the instrument's internal consistency. Cronbach's alpha 3 was calculated for the six dimensions of BariTest, assessed in four situations, i.e., considering the entire sample, only patients in the preoperative period, only in the postoperative period, and separating by gender ( Table 3).

Standardization of BariTest
To correct BariTest, it was necessary to multiply the response of each item by its respective general BariTest coefficient (Appendix 2) and calculate the average. The factorial loads were previously staggered so that each patient achieved a minimum of zero and a maximum of 100 points. It is important to note that some items had the score reversed; thus, items 32, 33,34,35,36,37,39,40,55,56,57,58, and 59 had the inverted correction, whereby 4=0, 3=1, 2=2, 1=3, and 0=4.
To interpret the score obtained, it was necessary to use the reference levels table (Appendix 3), calculated through the standard score (percentile). The characteristics of the patient were considered when they answered the BariTest (preoperative or postoperative phase, age, and gender) to check the percentile corresponding to that score. The purpose of this subdivision was to compare the score obtained with that of another similar subject 35 . The higher the score, the more the unwanted behaviors related to the construct.

Data Analysis
The results were expressed as mean and standard deviation when the scores were normally distributed. Differences between groups were assessed using the t-or F-test when the normality assumption holds, and the Mann-Whitney or Kruskal-Wallis test, otherwise. CFA was performed based on polychoric correlations, since they are indicated 12 instead of the usual Pearson linear correlations when data are expressed on an ordinal scale (Likert). In addition, data imputation based on proportional chance regression models was used to fill the missing values. Patients who did not respond to most questions were excluded from the analysis. All conclusions were based on a significance level of 5%. The statistical software R 36 version 3.5.0 was used in all analyses. The Psych library 37 was used to obtain the Cronbach's alpha, while the Lavaan library 48 was used for the CFA.

Sociodemographic data
This is a cross-sectional study; therefore, the three groups are composed of different people ( Table 2).

Validity of BariTest
For validation of BariTest, CFA (Appendix 2) was performed. The correlation between the items that make up each domain is shown in Figure 1. The factor loadings show how much the item is representative of construct. The more intense color tone shows a strong correlation; in contrast, the lower correlation level shows a weaker tone. The purple color represents a positive correlation, i.e., the answers point in the same direction, and the red represents a negative correlation, in which the answers point to the opposite of what that domain intends to prove. The variation ranges from 1 to −1, and the closer to 1 (purple color) means greater correlation between items. Therefore, the six BariTest factors show for the most part, strong and positive correlation.
The results of quality of the fit model are as follows: RMSEA of 0.064 (0.062; 0.066) and SRMR of 0.073 indicate an adequate fit, while the CFI of 0.926 and TLI of 0.923 indicate an acceptable fit 47 .
BariTest's responsivity (accuracy) was verified in a comparative manner with the WHODAS 2.0 and OP scores (Annexes 1 and 2), by analysis of the areas under the ROC curves. Bariatric surgery causes changes in the psychological well-being of patients undergoing the procedure. The results showed that WHODAS 2.0 has 65% accuracy, OP has 72%, and BariTest has 78% (Figure 2), being. therefore, superior to the others to identify the chances of psychological well-being of the patient with obesity.
Note: This figure reveals how the instrument's items correlate within the construct. Numbers refer to BariTest item numbers. BT: BariTest.

5/13
Reliability BariTest's reliability showed a Cronbach's alpha of 0.93 (95%CI, 0.92-0.94). The reliability of each construct was analyzed, considering the entire bariatric sample, and was separated by gender ( Table 3). The similarity of the results showed 19 that all of the instrument is valid for both genders.

Results of BariTest
The analysis between the constructs and groups (Figure 3) was adjusted for the results by the Bonferroni correction factor, to guarantee the significance level of 5%. The constructs Emotional state, Eating behavior, and Quality of life show a similarity between the results obtained in the community and postoperative groups and better than the preoperative group.
The community in general revealed to have more social support compared with obesity patients (preoperative and postoperative). Relationship with body weight differed in the three groups, possibly because the questions are specific to the bariatric population and the community was unable to answer.
Alcohol consumption was similar in the preoperative and postoperative groups and lower than the community group,  indicating that people in the community consume more than the bariatric population. Five of the six constructs obtained p<0.001, with Social support being p=0.0204.

DISCUSSION
There are numerous advantages for the psychologist to use BariTest, as it is a validated and complementary tool for psychological assessment that measures the psychological well-being of bariatric surgery patients. This instrument is valuable as a systematic procedure to collect, quantify, and evaluate the patient's behavior and compare the psychological outcomes of the surgery 20,44 . The instrument was also applied in the community to nonobese subjects, with the sole purpose of verifying whether bariatric patients are distinct from the general population. Thus, BariTest proved that it is specific for the bariatric population, since the results obtained with candidates or patients who have already undergone bariatric surgery are different from the findings with the nonobese community. The Emotional state construct consists of items that assess mood, anxiety, and impulsivity. Patients with obesity may have some cognitive difficulties, especially in the area of executive function responsible for planning, organizing, and controlling impulses 38,40 . The weight loss after bariatric surgery reduces neuroinflammation to rescue some aspects of defects in cognition and behavior 24 . Anxiety is the most common psychiatric disorder in patients with obesity who are awaiting bariatric surgery 14,21 .
The Emotional state score is similar between the postoperative period 29.7 (SD±16) and the community 27.1 (SD±13.8), but lower than the group that has not yet undergone surgery 37.8 (SD±15.5). This finding corroborates with the literature 11,31,32,34 that shows the prevalence of depressive disorders being lower than in patients who have already undergone bariatric surgery and that patients who are in the preoperative period of bariatric surgery demonstrate more critical levels of depression, higher than those observed in the general population. In addition, worsening depression is associated with weight gain, which in turn leads to worse depression outcomes 2 .
The preoperative patients scored in BariTest's Eating behavior (Figure 3), an average of 51.3 (SD±18.1), which was the highest average of all constructs, demonstrating that the candidate for bariatric surgery does not have a healthy relationship with food. It is important to assist the patient from the preoperative period, since studies have shown that the prevalence of binge eating symptoms in patients who are the candidates for bariatric surgery is 39-50% and is related to a suboptimal weight loss result after bariatric surgery 8,13,28,45 .
Quality of life and Relationship with body weight were constructs of BariTest which revealed a worse score in preoperative than postoperative and community. These data corroborate the prospective cross-sectional study by Moraes et al. 30 who analyzed quality of life before and after bariatric surgery, reporting that 25% of patients considered quality of life and health to be poor or very bad before bariatric surgery, and after the procedure all patients rated it as good or very good.
The BariTest Social support construct revealed that bariatric patients (preoperative and postoperative) have less social support than the community and it is known that social support is associated with greater adherence to treatment and consequently successful outcomes 26 .
BariTest showed that the bariatric sample had an alcohol consumption lower than that of the general population. This finding was different from the study by King et al. 23 and it is known that there is a significantly higher prevalence of alcohol consumption after bariatric surgery. It is believed that patients who are undergoing evaluation for bariatric surgery report a lower consumption of alcohol, since it is a contraindication for surgery. Furthermore, to have a low alcohol consumption in the postoperative period is important due to preventing alcoholism and weight regain 33 .
The results of the present study suggest that BariTest is a psychometric instrument capable of evaluating the psychological well-being of patients of both genders, before and after bariatric surgery (Table 3).
Even though BariTest has been validated with a significant number of patients, this study was cross sectional, because the aim of this study was to elaborate and validate this psychometric scale. Therefore after this stage, a longitudinal study would be very interesting to understand the changes that the surgery provides and perhaps predict the most suitable psychological profile for bariatric surgery. Sarwer et al. 38 emphasize the importance of these studies to improve patient selection, improve psychoeducation and preoperative interventions, in addition to developing intervention strategies for patients who are unable to achieve the expected result after the procedure.

CONCLUSION
BariTest is an instrument that makes it possible to measure and analyze psychological well-being and directs the necessary psychological interventions, before and after bariatric surgery, contributing to the psychological assessment. BariTest was developed as recommended in the scientific literature and proved all of the instrument was valid and reliable (α=0.93), measuring the psychological well-being of bariatric patients, regardless of gender, before and after bariatric surgery.

Annex 2 -Complementary Scales: Brazilian version of the Obesity-related Problems Scale (OP)
How do you feel about your weight or your body shape in the following situations? OP1.
Receiving friends at home OP2.
Visiting the home of relatives or friends OP3.
Going to restaurants OP4.
Doing activities in the community (courses etc.) OP5.
Holidaying away from home OP6.
Trying on and buying clothes OP7.
Intimate relationships (kiss, sex, etc.) OP items are represented by the acronym "OP" followed by their ordering number.
All of them must be answered on a Likert scale as follows: (1) "It bothers me a lot." (2) "It bothers me more or less." (3) "It bothers me a little." (4) "It doesn't bother me."

H1
Overall, in the past 30 days, how many days were these difficulties present?

H2
In the past 30 days, for how many days were you totally unable to carry out your usual activities or work because of any health condition?

H3
In the past 30 days, not counting the days that you were totally unable, for how many days did you cut back or reduce your usual activities or work because of any health condition?

Record number of days ______
This completes the questionnaire. Thank you.

Annex 1 -Continuation
Appendix 1 -Evaluation of the version of BARITEST by the Committee of Experts. 3.7 6. When I feel sad or anxious or idle I have a habit of compensating with food and overeating.
3.7 7. I chew my food well and eat my meals calmly.
3.6 8. I eat few times a day, but when I eat, I overdo it.
3.6 9. I have difficulty in distinguishing between hunger and the desire to eat.
3.9 10. I have crises of eating a lot until I am full.
3.1 11. I have a habit of eating "fast food" (Snacks).
3.8 12. I am a candy eater.
3.5 13. When I feel the urge to eat, it is difficult to control myself.
3.9 14. I intend to eat just a little, but when I see it, I eat a lot more than I want to.
3.4 15. I notice that I eat more at night.
3.9 16. I wake up in the early hours to eat something. 3 17. When I go on a diet, I manage to stop eating some foods that I love, without any problem.
3.3 18. My behavior towards food causes me a lot of suffering.
3.9 19. When I have emotional problems, I use food to relieve tension or to bring me joy.
3.9 20. When I feel like eating a treat, I eat without delaying and/or depriving myself.
3.7 21. I have difficulty leaving food on the plate at the end of a meal.
3.9 22. The next morning, after drinking, I wake up with a hangover (If you don't drink, mark 0).
3.6 23. I drink alcohol on weekends.
3.1 24. I drink alcohol during the week.
3.4 25. People tell me that I am drinking too much. (If you don't drink, mark "never").
3.5 26. I am in the habit of using alcohol to relax and be happy. (If you don't drink, mark "never").
3.9 27. After drinking alcohol, I missed or was late for an appointment the next day. (If you don't drink, mark "never").
3.6 28. I don't like going to social events that don't have alcohol. 3