Using psycho‐behavioral phenotyping for overweight and obesity: Confirmation of the 6 factor questionnaire

Abstract Objective Questionnaires that assess dietary habits, eating behaviors, and relevant psychosocial constructs are routinely used in obesity research and clinical practice. The 6 factor questionnaire (6FQ) was previously developed as an assessment tool for psycho‐behavioral phenotyping. The primary purpose of this study was to confirm and validate the original findings in a large diverse adult population. Methods A total of 5399 self‐selected participants (mean age of 48 ± 13 years and body mass index of 32 ± 8 kg/m2) completed the 6FQ online. The association between self‐reported demographic data and 6FQ responses was assessed using linear regression models. Results Mean factor score and odds ratio analyses consistently demonstrated a statistically significant relationship between factors and body weight even after adjusting for age, sex, and race/ethnicity. Conclusions Although the study was correlational in design, the results demonstrate that the 6FQ, an instrument that represents multidimensional unhealthful lifestyle patterns associated with diet, physical activity, cognition, and self‐perception worsen with increasing body weight. Psycho‐behavioral phenotyping may be a useful approach when assessing and treating patients with obesity.


| INTRODUCTION
Since obesity is considered a multifactorial disease due to the interaction of various biological, psychosocial, and cognitive factors experienced throughout life, 1 there is a need to develop a tool that can identify the individuality of these differences for each person and facilitate targeted, evidence-based weight management strategies.
Phenotyping patients based on these characteristics represents a unique method to provide more personalized care. Such a tool could be used by people with overweight or obesity as a guide for selecting a more tailored self-directed approach to weight loss and by clinicians who can enable the initiation of more effective and efficient weight management counseling.
Multiple questionnaires have been developed and validated to assess psychological, behavioral, or dietary factors used primarily for research in individuals with obesity. [2][3][4][5][6][7][8][9] However, further This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. development of a comprehensive, short, and clinically useful assessment questionnaire that identifies unique patterns of behavior and cognition would be useful for people with obesity and clinicians alike.
Toward this end, we previously developed and validated by factor analysis a new 27-item questionnaire using two prospective subject groups (n = 640). 10

| Measures
Body mass index was calculated as weight in kilograms divided by square of height in meters. Body mass index was classified into two categories for analysis: healthy weight (between 18.5 and 24.9 kg/m 2 ) and overweight or obese (≥25 kg/m 2 ). Baseline demographic characteristics were compared between participants in the two BMI categories using generalized linear models for continuous variables and chi-square tests for categorical variables.
For the continuous variables, the numerical responses to each factor item (0 = don't agree at all, 1 = agree a little, 2 = agree, 3 = strongly agree) were added together to create the factor sum score. The sum scores of each of the six factors were then divided by the number of items assigned to each factor score to create the average factor score. The number of items for the six factors was 4 for Fast Pacer, Exercise Struggler, Self-Critic, and All-or-Nothing Doer; 5 for Convenient Diner; and 6 for Easily Enticed Eater. The average factor scores ranged between 0 and 3. As an example, if the summed score for Easily Enticed Eater was 15, it would be divided by 6 (number of items) yielding an average factor score of 2.5.
For the categorical variables, a percentage factor score was calculated as the summed factor score divided by the maximum factor score, multiplied by 100. Using the example above for the Easily Enticed Eater, 15 divided by a maximum score of 18 yields a 186 -KUSHNER AND HAMMOND categorical score of 83%. A cutoff categorical factor score of >66% was empirically used for positive identification of a factor based on the previous validation study. 10

| Statistical analysis
In these analyses, self-reported BMI was the outcome of interest and the six-factor score categories were the exposure of interest. Linear regression models were used to assess the association between each average factor score and self-reported BMI. Logistic regression models were used to calculate the odds of self-reported overweight or obesity, by six-factor score categories, using the healthy weight as the reference group. For both analyses, Model 1 was unadjusted and Model 2 was adjusted for age, sex, and race/ethnicity. Analyses were performed using SAS statistical software (version 9.4, SAS Institute).
Statistical significance was defined as p < 0.05.

| RESULTS
Participant demographics for the study are displayed in Table 1. The relationship between the factor scores and BMI was then assessed by performing two linear regression models ( Table 2). The association between average factor scores and BMI was statistically significant (p < 0.001) for all six factors even after adjusting for age, sex, and race/ethnicity. In the Exercise Struggler factor, for every unit increase in average score, BMI increased by 3.85 kg/m 2 after adjusting for age, sex, and race/ethnicity. The associated BMI increases for Convenient Diner and Easily Enticed Eater factors were 3.77 and 3.09 kg/m 2 , respectively.

| DISCUSSION
This study demonstrates significant correlations between the six factors and body weight in a large diverse adult population who voluntarily chose to take the online questionnaire. The results confirm the correlational relationship that was observed in our previously published development and validation study. 10 We chose to limit our BMI categories into healthy weight and overweight and obese combined in order to avoid misclassifications due to selfreported height and weight. Although the majority of subjects were female (86.9%) and Caucasian (80.0%), it did include a racially, ethnically, and age diverse population of individuals. This is important since the instrument was designed to be used as an assessment tool for overweight and obesity care. Although the study was correlational in design and causality cannot be proved, the results are consistent with our original hypothesis, that is, the six factors that represent multidimensional unhealthful lifestyle patterns associated with diet, physical activity, cognition, and self-perception worsen with increasing body weight. The 6FQ supports two communication approaches intended to facilitate improved health behavior change: patient-centered care and patient treatment tailoring or segmentation. Patient-centered care is defined as care provision that is consistent with the values, needs, and desires of patients and is achieved when clinicians involve patients in health-care discussions and decisions. 13 Core components of patient-centered care include communication (sharing information and sensitivity to patient needs), partnership (relationship building), and health promotion (supporting optimal health and care through reflection on the patient's history). 14 By eliciting the personal habits, attitudes, and emotions of the patient and developing a treatment plan based on these reflections, the 6FQ is intended to create a therapeutic alliance and enable patients to exert control over their health and determinants of obesity.

T A B L E 1 Baseline characteristics of 5399 study participants
The second communication approach the 6FQ supports is patient tailoring or segmentation. Segmentation theory tells us that a "one size fits all" approach does not meet the needs of all patients. 15 Psycho Although this study confirmed that a short 27-item selfadministered questionnaire showed significant correlations between the six factors and body weight in a large heterogeneous population, there are some limitations. All of the data were selfreported. A systematic review showed trends of underestimating weight and BMI and overestimating height by self-report. 22 To overcome this potential error, we simplified our database into only two BMI categories. The majority of the participants were female and Caucasian. Although we were successful in reaching a racially diverse population with varying BMI strata, they may not be representative of all individuals who take the questionnaire. We also do not have data on social determinants of health data that may be related to the factors. As a correlational study, we cannot confirm that the relationship between factors and obesity is causal, although factor scores consistently and significantly increased with increasing body weight.
We also did not establish that targeted counseling using the six All-or-nothing doer 2.64 0.14 <0.0001 Note: Average factor score (from 0 = "Don't agree at all" to 3 = "Strongly agree").
Abbreviation: BMI, body mass index. factors will lead to weight loss. Lastly, we did not demonstrate the utility of the 6FQ in a counseling session.

| CONCLUSION
The study confirms a strong, consistent, and statistically significant association between the six factors and body weight in a diverse selfselected population. The 27-item questionnaire identifies six selfidentified factors assessing behavioral, cognitive, and affective factors. It is designed as a self-help tool or as an intake instrument that may allow clinicians to efficiently and effectively counsel patients on targeted treatment recommendations. Among the multitude of barriers that impact a clinician's decision to engage in obesity care, some of the most important factors are time restraints and lack of clarity on what lifestyle issues to focus on in a particular patient. The 6FQ was developed to directly address these concerns. It is a convenient, short, self-administered instrument that can be completed prior to the patient encounter; it is quickly scored by the patient or clinician; it targets patients' self-identified behavioral, cognitive, and affective lifestyle factors related to weight gain; it should allow clinicians to counsel patients on targeted treatment recommendations. Ongoing and future studies will evaluate its utility and effectiveness as a communication and assessment tool in weight management.