Biology 4a assignment

© 2018 American Journal of Preventive Medicine. Pu reserved.

RESEARCH ARTICLE

From the 1Cleve Reserve Univers Research, Medici ter for Behavior Cleveland Clinic,

Address corr Clinic Main Cam 44195. E-mail: ch

0749-3797/$3 https://doi.org

blished by Elsevier Inc. All

Associations Between Unhealthy Weight-Loss

Strategies and Depressive Symptoms

D1X XAlexander Chaitoff, D2X XMPH,1 D3X XCarol Swetlik, D4X XBA,1 D5X XCatherine Ituarte, D6X XBS,1 D7X XElizabeth Pfoh, D8X XPhD, MPH,2

D9X XLing-Ling Lee, D10X XBS,1 D11X XLeslie J. Heinberg, D12X XPhD,3 D13X XMichael B. Rothberg, D14X XMD, MPH2

Introduction: There appears to be a link between weight loss and improved mental health, but less is known about how using unhealthy weight-loss strategies impacts the odds of reporting depression.

Methods: This study includes respondents from the National Health and Nutrition Examination Survey from 2005 to 2014 who attempted to lose weight over the past year. Analysis occurred in 2017. Multivariable logistic regression was used to describe associations between all weight-loss strategies, including those grouped as unhealthy (smoking, vomiting, laxatives, skipping meals, and using diet pills), and the adjusted odds of depression (Patient Health Questionnaire−9 score ≥10). The model was then stratified by BMI, sex, race, and antidepressant use to compare the effect of using at least one unhealthy weight-loss strategy and depression within certain populations.

Results: The sample included 6,765 respondents (weighted n=59.2 million, 95% CI=55.5, 62.9 mil- lion). Of these respondents, 18.0% (n=1,270) reported using at least one unhealthy weight-loss strategy. In unadjusted analysis, unhealthy weight-loss strategies were generally associated with higher incidence and odds of reporting depression. In multivariable analysis, using at least one unhealthy weight-loss strategy was significantly associated with odds of reporting depression (AOR=1.47, 95% CI=1.14, 1.91, p<0.01). When the model was stratified, this effect was statistically significant among respondents with class I or II obsesity (AOR=2.20, 95% CI=1.56, 3.10, p<0.01); female respondents (AOR=1.46, 95% CI=1.06, 2.00, p=0.02); and respondents who did not use an antidepressant (AOR=1.57, 95% CI=1.14, 2.15, p=0.01).

Conclusions: Unhealthy weight-loss strategies are associated with increased odds of depression. This may inform screening practices and public health messaging. Am J Prev Med 2019;56(2):241−250. © 2018 American Journal of Preventive Medicine. Published by Elsevier Inc. All rights reserved.

INTRODUCTION

land Clinic Lerner College of Medicine, Case Western ity, Cleveland, Ohio; 2Center for Value-Based Care ne Institute, Cleveland Clinic, Cleveland, Ohio; and 3Cen- al Health, Department of Psychiatry and Psychology, Cleveland, Ohio espondence to: Alexander Chaitoff, MPH, Cleveland pus, Mail Code NA21, 9500 Euclid Ave., Cleveland OH [email protected]. 6.00 /10.1016/j.amepre.2018.09.017

Half of all Americans, including many within a non-overweight BMI range,1−3 spend billions of dollars annually to lose weight.4−6 To com-

bat the obesity epidemic and improve population health, NIH endorsed guidelines that recommend reducing caloric intake and increasing physical activity for adults with a BMI of 25 or more.7 However, many individuals attempting to lose weight do not follow these guidelines and use weight-loss strategies that are not sustainable or are medically unhealthy.8−10

In addition to their well-described relationships with physical health, there is evidence that weight-loss

strategies are associated with mental health. For exam- ple, medically unhealthy approaches to weight loss, such as laxative use or smoking for meal replacement, have been shown to be associated with depression and other

rights Am J Prev Med 2019;56(2):241−250 241

242 Chaitoff et al / Am J Prev Med 2019;56(2):241−250

dysphonic moods in adolescent and young adult popula- tions.2,11−14 In adults, certain diets, such as the Mediter- ranean diet, have been associated with improving mental health status.15−18 However, studies exploring the more general associations between multiple weight-loss strate- gies on mental health remain largely limited to samples that are composed of college-aged subjects1,12−14,19,20

and usually include only individuals with a BMI of 25 or more.5 These populations are narrow and do not encom- pass many U.S. adults attempting to lose weight. Describing the association between weight-loss strate-

gies and mental health in the general population is war- ranted given the high prevalence of weight-loss attempts among individuals of all ages and BMIs as well as the extent to which depression is underdiagnosed and undertreated in the U.S.21,22 Furthermore, in a national sample, less than 5% of primary care offices were found to routinely screen patients for depression,23 and better quantifying the association between weight-loss strate- gies and depression may lead to clinically useful infor- mation to cue physicians to ask patients about their mental health in addition to further characterizing the complex association between weight and mental health. Thus, the current study utilized data from a nationally representative sample of U.S. adults between 2005 and 2014 to characterize the association between weight-loss strategies and depressive symptoms.

METHODS

Study Population The sample consisted of adults aged ≥18 years who took part in the National Health and Nutrition Examination Survey (NHANES). NHANES is a cross-sectional survey conducted by the Centers for Disease Control and Prevention. NHANES utilizes a complex, mul- tistage, clustered probability method to sample non-institutional- ized U.S. civilians to provide nationally representative estimates.24

Responses of individuals sampled between 2005 and 2014 who completed both the weight-loss history and mental health screener and also reported attempting to lose weight in the past year were eli- gible for inclusion. Respondents were excluded if they reported try- ing to lose weight but had a BMI<18.5, felt that they were currently underweight, or had missing responses. This sample encompassed the 10most recent years of NHANES data.

Measures Primary predictor variables were the weight-loss strategies included in the NHANES survey. NHANES provides a list of 20 predefined weight-loss strategies and asks respondents to self- report if they used any while trying to lose weight in the past year. The strategies are listed in Table 1.25 One strategy, “Had surgery to lose weight,” was not described as a separate category because it was not included as an option until 2013−2014 (0.1%, n=5). One binary composite variable was constructed corresponding to the use of weight-loss strategies that were classified as unhealthy (used/did not use). Specifically, respondents who had reported

using at least one of the following weight-loss strategies were con- sidered to have used an unhealthy method: skipping meals, vomit- ing/using laxatives, smoking to lose weight, taking prescription diet pills, or taking nonprescription diet pills. All other weight- loss strategies were considered healthy or neutral and coded together as not unhealthy, resulting in the final binary variable. This grouping was informed by the coding scheme from Project Eat III and the related literature.11,26

The primary outcome measure was depressive symptoms eval- uated using the Patient Health Questionnaire−9 (PHQ-9), which is a validated measure of depression in adults.27−29 The PHQ-9 consists of nine questions that ask respondents to rank on a scale of 0 (not at all) to 3 (nearly every day) how often they have experi- enced symptoms of depression in the past 2 weeks for a final score that ranges from 0 to 27. In order to maximize sensitivity and specificity, respondents were classified as depressed if they had a PHQ-9 score ≥10 (equivalent to moderate or greater symptoms).30,31

Potential confounder variables were selected based on their previously demonstrated associations with depression. These vari- able groups included demographic,32 general health,33 and meas- ures as well as perceptions of weight status.34 Demographic variables included age (continuous); sex (binary); race (nominal); family income-to-poverty ratio (continuous); education (nomi- nal); and marital status (nominal). Health variables included Lik- ert-type scale measured general health status (binary; excellent/ very good/good versus fair/poor) and antidepressant medication use (yes versus no in the past 30 days). Weight variables included BMI (continuous); body-weight perception (binary; perceive self as normal weight versus overweight); and self-reported weight change over the past year (nominal; weight maintenance versus gained weight versus lost weight35).

Statistical Analysis Chi-square tests were used to compare prevalence of depression among those who did and did not report attempting to lose weight in the past year. Student’s t-tests (for continuous responses) and chi-square tests (for categorical responses) were used to compare the demographic characteristics of respondents across those who did and did not report using an unhealthy weight-loss strategy. Unadjusted counts were used to describe prevalence of each weight-loss strategy as well as the sum of unhealthy weight-loss strategies used by participants. Multivariable logistic regression models were used to assess both (1) the use of each weight-loss strategy and odds of depression, and (2) the composite variable of any unhealthy weight-loss strategy and odds of depression. All multivariable models adjusted for all confounder variables described in the Measures section. To examine the impact of effect modification on the association between the composite variable and depression, the final model was stratified by (1) BMI catego- ries (non-overweight: 18.5−24.9, overweight: 25.0−29.9, class I or II obesity: 30.0−39.9, class III obesity: ≥40.0); (2) sex, (3) race (black, white, Hispanic, other); or (4) antidepressant use (yes, no) based on evidence that different demographic groups may be psy- chologically unique36,37 and that antidepressants are widely used for non-depression diagnoses but may be associated with both reporting depression and weight change.38,39 Finally, a sensitivity analysis was performed that removed prescription diet pills from the composite variable, because such a method does include

www.ajpmonline.org

Table 1. Association Between Depressive Symptoms and Weight-Loss Strategies in Adults Trying to Lose Weight

Weight-loss strategies n (%)a AOR of depressionb,c (95% CI) p-value

Not unhealthy strategies

Ate diet foods or products 719 (12) 1.20 (0.84, 1.72) 0.31

Ate fewer carbohydrates 1,718 (26) 0.87 (0.66, 1.14) 0.31

Ate less candy and sweets 1,236 (18) 0.92 (0.65, 1.29) 0.62

Ate less fat to lose weight 2,381 (34) 1.01 (0.79, 1.29) 0.96

Ate less to lose weight 4,202 (63) 1.30 (1.05, 1.60) 0.02*

Ate more fruits, vegetables, and salads 1,654 (24) 1.00 (0.78, 1.29) 0.99

Changed eating habits 1,192 (18) 0.95 (0.71, 1.27) 0.71

Drank a lot of water 2,662 (40) 0.84 (0.70, 1.04) 0.11

Drank liquid diet formula 343 (5) 1.91 (1.19, 3.07) 0.01*

Exercised to lose weight 4,148 (63) 0.78 (0.60, 1.01) 0.05

Followed a special diet 439 (8) 0.75 (0.45. 1.26) 0.23

Joined a weight-loss program 368 (7) 0.76 (0.40, 1.45) 0.40

Other 71 (1) 3.25 (1.47, 7.19) <0.01** Switched to foods with lower calories 2,182 (34) 1.11 (0.89, 1.38) 0.40

Unhealthy strategies

Skipped meals 1,136 (16) 1.48 (1.11, 1.98) 0.01*

Started to smoke or began smoking again 55 (1) 0.90 (0.31, 2.61) 0.84

Took laxatives or vomited 67 (1) 3.31 (1.47, 7.50) 0.01*

Took non-prescription diet pills/supplements 470 (7) 1.99 (1.32, 2.98) <0.01** Took prescription diet pills 161 (2) 1.69 (0.97, 2.95) 0.06

Note: Boldface indicates statistical significance (*p<0.05, **p<0.01). aWeighted percentage. bAdjusted for age, sex, race, education, marital status, family income-to-poverty ratio, BMI, self-reported weight change in the past year, self-reported health status, body weight perception, and antidepressant use. cDepression defined as PHQ-9 ≥10. PHQ-9, Patient Health Questionnaire−9.

Chaitoff et al / Am J Prev Med 2019;56(2):241−250 243

medical oversight. All confounder variables included in the con- ceptual models used to build multivariable regression models were informed by the literature and interrogated using logistic regression and found to be significantly associated with both depression and use of an unhealthy weight-loss strategy at an a

level of 0.20.40,41

Appropriate weights were applied to take into account the complex, multistage sampling methodology of NHANES, and all statistics except for table counts represent weighted estimates.42

Specifically, the Complex Sample Analysis module within SPSS, version 24 (www.ibm.com/analytics/us/en/technology/spss/), which enables proper weighting of NHANES data, was used for all analysis.43 Data analysis occurred in 2017.

RESULTS

From 2005 to 2014, there were 50,965 respondents to NHANES. Of these, a total of 28,551 and 26,213 respondents were sampled to complete the weight his- tory and mental health screener section of the NHANES survey, respectively, whereas 22,925 completed both sec- tions. Approximately 37% (n=9,498, 95% CI=35.7%, 37.3%) reported attempting to lose weight in the past year, and of these individuals 7,947 also completed the

February 2019

mental health screener module. After excluding respondents with missing data (n=1,118); a BMI <18.5 (n=10); or respondents who considered themselves underweight (n=54), the final sample size was 6,765 respondents (85.1% of respondents meeting inclusion criteria). This represents a weighted estimate of 59.2 mil- lion U.S. adults (95% CI=55.5, 62.9 million). Those who reported attempting to lose weight had a

higher prevalence of depressive symptoms than those who did not report attempting to lose weight (7.7% vs 6.5%, p<0.01). The plurality of respondents had class I or II obesity (37.7%, n=2,781), though a significant per- centage were individuals with a BMI <25 (18.7%, n=1,051). The majority of the sample were female (62.8%, n=4,326); white (71.5%, n=3,259); and not depressed (92.3%, n=6,148) or on an antidepressant (85.6%, n=5,945). Those who used an unhealthy weight- loss strategy tended to be younger, to have a higher BMI, to have been more likely to have lost weight in the past year, to be non-white, to report fair/poor health, to have less education, and to be non-married (Table 2). The majority of participants did not report any

unhealthy weight-loss strategies (69.6%, n=4,702). Of

Table 2. Comparison of Adults That Used an Unhealthy Weight-Loss Method to Adults That Did Not

Characteristics

Did not use an unhealthy weight-loss method (n=5,214)

Used an unhealthy weight-loss methoda

(n=1,551) p-valueb

Age in years, M § SD 47 § 16 42 § 14 <0.01** Female,c n (%) 3,202 (62) 1,004 (64) 0.16

Race,c n (%) <0.01** White 2,559 (73) 643 (65)

Black 938 (9) 421 (15)

Hispanic 1,258 (12) 367 (13)

Other 459 (6) 120 (7)

Marital status,c n (%) <0.01** Married 2,957 (61) 774 (53)

Living with partner 346 (7) 136 (10)

Widowed 294 (4) 57 (3)

Divorced/separated 746 (13) 267 (15)

Never married 871 (15) 317 (19)

Education level,c n (%) <0.01** College graduate 1,570 (37) 325 (25)

Some college 1,633 (32) 561 (36)

Completed high school or equivalent 1,067 (19) 408 (27)

Did not complete high school 944 (12) 257 (12)

Family income-to-poverty ratio,d M § SD 3.3 § 1.6 3.0 § 1.6 <0.01** BMI,c n (%) <0.01** Non-overweight, 18.5−24.9 857 (19) 194 (16)

Overweight, 25.0−29.9 1,749 (35) 446 (30)

Class I or II obesity, 30.0−39.9 2,100 (37) 681 (40)

Class III obesity, ≥40 508 (9) 230 (14)

Self-reported weight change in past year,c n (%) <0.01** Weight maintenance, §1 kg 1,932 (40) 453 (32)

Weight gain, >1 kg 1,031 (18) 257 (16)

Weight loss, >1 kg 2,251 (42) 841 (52)

Self-reported health status,c n (%) <0.01** Good or better 4,117 (85) 1,161 (80)

Fair or poor 1,097 (15) 390 (20)

Body weight perception,c n (%) <0.01** Normal weight 1,032 (21) 236 (15)

Overweight 4,182 (79) 1,315 (85)

Depressive symptoms,c n (%) <0.01** None to mild, PHQ-9 <10 4,796 (93) 1,352 (88)

Moderate to severe, PHQ-9 ≥10 418 (7) 199 (12)

Antidepressant use,c n (%) 0.11

No 4,601 (86) 1,344 (84)

Yes 613 (14) 207 (16)

Note: Boldface indicates statistical significance (*p<0.05, **p<0.01). aUnhealthy weight-loss methods included skipping meals, vomiting/using laxatives, smoking to lose weight, taking prescription diet pills, and taking non-prescription diet pills/supplements. bp-values were calculated with Student’s t-test for continuous variables and Pearson’s x2 test for categorical variables. cWeighted percentage. dFamily income-to-poverty ratio was calculated by dividing the annual family income by the survey-year specific HHS poverty guidelines. PHQ-9, Patient Health Questionnaire−9.

244 Chaitoff et al / Am J Prev Med 2019;56(2):241−250

those that did report an unhealthy weight-loss strategy, the majority (68.5%, n=1,429) reported using only one unhealthy weight-loss strategy, of whom the plurality reported skipping meals (44.3%, n=687). In

multivariable analysis of the association between weight- loss strategies and adjusted odds of depressive symptoms (Table 1), adjusting for demographic characteristics, measures as well as perception of weight, and general

www.ajpmonline.org

Table 3. Multivariable Model of Factors Associated With Depressive Symptoms in Adults Trying to Lose Weight

Variables OR of depressiona (95% CI) p-value

Predictor variable

Used an unhealthy weight-loss strategyb <0.01** No (ref) 1.00 (—)

Yes 1.47 (1.14, 1.91)

Confounder variables

Age in yearsc 0.99 (0.98, 1.00) 0.02*

Sex 0.03*

Male (ref) 1.00 (—)

Female 1.37 (1.03, 1.82)

Race 0.30

White (ref) 1.00 (—)

Black 0.85 (0.63, 1.15)

Hispanic 0.88 (0.65, 1.18)

Other 1.26 (0.82, 1.93)

Marital status <0.01** Married (ref) 1.00 (—)

Living with partner 1.27 (0.92, 1.76)

Widowed 1.24 (0.66, 2.33)

Divorced/separated 2.02 (1.50, 2.72)

Never married 1.21 (0.84, 1.76)

Education level 0.04*

College graduate or above (ref) 1.00 (—)

Some college 1.37 (0.92, 2.10)

Completed high school or equivalent 1.59 (1.04, 2.44)

Did not complete high school 1.85 (1.19, 2.87)

Family income-to-poverty ratio 0.82 (0.74, 0.91) <0.01** Self-reported health status <0.01**

Good or better (ref) 1.00 (—)

Fair or poor 4.80 (3.74, 6.17)

BMI 1.00 (0.99, 1.02) 0.62

Body weight perception <0.01** Normal weight (ref) 1.00 (—)

Overweight 1.87 (1.25, 2.81)

Self-reported weight change in past year 0.26

Weight maintenance, §1 kg (ref) 1.00 (—)

Weight gain, >1 kg 1.31 (0.94, 1.81)

Weight loss, >1 kg 1.16 (0.89, 1.50)

Antidepressant use <0.01** No (ref) 1.00 (—)

Yes 4.02 (3.01, 5.37)

Note: Boldface indicates statistical significance (*p<0.05, **p<0.01). aDepression defined as PHQ-9 ≥10. bUnhealthy weight-loss methods included skipping meals, vomiting/using laxatives, smoking to lose weight, taking prescription diet pills, and taking non-prescription diet pills/supplements. cPer 1-year increase in age. PHQ-9, Patient Health Questionnaire−9.

Chaitoff et al / Am J Prev Med 2019;56(2):241−250 245

health information, a total of six weight-loss strategies were significantly associated with the odds of reporting depression. Specifically, compared with individuals not using the strategy, prevalence of scoring depressed on the PHQ-9 was significantly higher in those who ate less

February 2019

(AOR=1.30, 95% CI=1.05, 1.60, p=0.02); who skipped meals (AOR=1.48, 95% CI=1.11, 1.98, p<0.01); who took laxatives or vomited (AOR=3.31, 95% CI=1.47, 7.50, p=0.01); who took nonprescription supplements for weight loss (AOR=1.99, 95% CI=1.32, 2.98, p<0.01);

Table 4. Multivariable Analyses of the Association Between Use of an Unhealthy Weight Loss Method and Depression Strati- fied by BMI, Sex, Race, or Antidepressant Use

Used an unhealthy weight-loss method (yes versus no)a AOR of depressionb,c (95% CI) p-value

BMId

Non-overweight, 18.5−24.9 1.37 (0.59, 3.14) 0.46

Overweight, 25.0−29.9 0.99 (0.55, 1.78) 0.96

Class I or II obesity, 30.0−39.9 2.20 (1.56, 3.10) <0.01** Class III obesity, ≥40 1.14 (0.63, 2.04) 0.67

Sexe

Male 1.53 (0.94, 2.47) 0.08

Female 1.46 (1.06, 2.00) 0.02*

Racef

White 1.46 (0.97, 2.19) 0.07

Black 1.16 (0.75, 1.81) 0.50

Hispanic 1.25 (0.89, 1.76) 0.20

Other 2.27 (0.90, 5.71) 0.08

Antidepressant useg

No 1.57 (1.14, 2.15) 0.01*

Yes 1.24 (0.77, 2.01) 0.38

Note: Boldface indicates statistical significance (*p<0.05, **p<0.01). aUnhealthy weight-loss methods included skipping meals, vomiting/using laxatives, smoking to lose weight, taking prescription diet pills, and taking non-prescription diet pills/supplements. “No” is the reference value. bAll models included age, education, marital status, family income-to-poverty ratio, self-reported weight change in the past year, self-reported health status and body weight perception as confounder variables. cDepression defined as PHQ-9 ≥10. dModel also includes sex, race, and antidepressant use as confounder variables. eModel also includes BMI, race, and antidepressant use as confounder variables. fModel also includes BMI, sex, and antidepressant use as confounder variables. gModel also includes BMI, sex, and race as confounder variables. PHQ-9, Patient Health Questionnaire−9.

246 Chaitoff et al / Am J Prev Med 2019;56(2):241−250

who used a liquid diet formula (AOR=1.91, 95% CI=1.19, 3.07, p=0.01); and who selected the option of having used other weight-loss methods (AOR=3.25, 95% CI=1.47, 7.19, p<0.01). Alternatively, those who exer- cised to lose weight had reduced odds of reporting depression than those who did not report using this weight-loss strategy, though the association did not reach statistical significance (AOR=0.78, 95% CI=0.60, 1.01, p=0.05). In multivariable analysis that assessed the relationship

between the composite variable (i.e., using at least one unhealthy weight-loss strategy) and the odds of depres- sion (Table 3), using at least one unhealthy weight-loss strategy versus not was significantly associated with odds of reporting depression (AOR=1.47, 95% CI=1.14, 1.91, p<0.01). Additionally, younger age; female sex; being divorced (compared with married); having less education; having a lower income-to-poverty ratio; being in poor health; self-perception of non-normal weight; and using antidepressants were associated with increased adjusted odds of reporting depression (Table 3). In the stratified analysis by BMI category, sex, antidepressant use, or race (Table 4), using at least one unhealthy weight-loss strategy was significantly associated with

increased odds of reporting depression only in respond- ents with class I or II obesity (AOR=2.20, 95% CI=1.56, 3.10, p<0.01); female respondents (AOR=1.46, 95% CI=1.06, 2.00, p=0.02); and respondents not using an antidepressant (AOR=1.57, 95% CI=1.14, 2.15, p=0.01). The sensitivity analysis for respondents who used only

prescription drug use as the unhealthy weight-loss strat- egy, which did not categorize 78 respondents (1.3%) as having used an unhealthy weight-loss strategy, repro- duced the results presented above.

DISCUSSION

This study utilized 10 years of data from NHANES to describe associations between weight-loss strategies and odds of reporting depressive symptoms. Several weight-loss strategies were significantly associated with the odds of reporting depression. Specifically, using at least one unhealthy weight-loss strategy was associated with a 47% increased odds of reporting depression. From the stratified analysis, this association was pres- ent only for respondents who had a BMI between 30.0 and 39.9, who were female, or who did not report taking antidepressants.

www.ajpmonline.org

Chaitoff et al / Am J Prev Med 2019;56(2):241−250 247

Importantly, this study should be applied in the con- text of individuals who are attempting to lose weight. Respondents attempting to lose weight were already sig- nificantly more likely to have depression than those who were not. Previous studies have found that individuals who are attempting to lose weight have a lower health- related quality of life and decreased health, vitality, and physical functioning.44,45 The results presented here sug- gest that choice of weight-loss strategy is associated with depression status, and future work should explore how healthcare professionals can support both healthy weight loss and mental health. This study specifically adds to the large volume of lit-

erature supporting a complicated link between BMI, weight-loss behaviors, patient sex, and psychopathology. Individuals with depression can report decreased appe- tite; thus, individuals with undertreated depression may find skipping meals more achievable than attempting healthier weight-loss methods, such as exercise. Given these findings, weighed with the fact that some antide- pressants increase risk of weight gain and type 2 diabe- tes, providers may counsel patients regarding these topics when considering the link between weight and mental health.46,47

The finding that unhealthy weight-loss behaviors are associated with depression in females mirrors the cur- rent literature, which suggests self-objectification behav- iors reflecting societal pressures with respect to body image can moderate associations between weight-loss behaviors and depression in females.48−50 Alternatively, it should be noted that the effect size was similar in males and females, and therefore the statistical difference may be a function of the larger sample size of females. However, given the totality of the research, females using unhealthy weight-loss strategies should be assessed for depression. Finally, the finding that use of an unhealthy weight-

loss strategy is associated with depression only among individuals within certain BMI groupings is concordant with other studies that suggest individuals within differ- ent weight groups experience different sociocultural and psychological stresses.36,51 However, it is notable that in the studied population, the link between unhealthy weight-loss behaviors and increased risk for depression was seen only in individuals with class I or II obesity but not in those with class III obesity. Additional research, perhaps qualitative in nature, is needed to more deeply describe this association, but it might be that respond- ents with class III obesity more often face significant health problems and challenges with activities of daily living because of their weight that influence risk of depression,52 and the results presented here may suggest that such factors could be more closely associated with

February 2019

depressive status than weight-loss modality. These find- ings highlight the importance of recognizing the psycho- logical uniqueness of different groups of individuals with different weights and the sometimes non-linear associations between weight and mental health. This study may also have a role in assisting public

health and clinical workers identify those at increased risk of mental health pathology. In addition to well- documented high-risk groups, such as patients with chronic diseases (e.g., cancer), individuals who are trying to lose weight, and especially those trying to do so by unhealthy tactic, may also be at higher risk for depres- sion and may deserve special consideration of mental health needs during health care appointments.53−55

Given discussions of BMI and weight have been reported to occur in as many as 70% of visits in a primary care setting, with the number of minutes devoted during the visit proportional to BMI,56 these discussions may pro- vide fruitful opportunities to also discuss mental health status, particularly if an effort is made to understand the approach a patient takes to weight loss. Regardless of depressive status, nearly 20% of

respondents endorsed use of at least one unhealthy weight-loss behavior, with individuals using these modalities significantly likely to be younger, black, with less education, and of lower income. Motivational inter- viewing and providing education about unhealthy weight-loss behaviors and associated risks to the cardio- vascular, metabolic, skeletal, and intestinal systems, in addition to the associations with poor mental health reported here, are feasible in the primary care setting and may be opportunities to connect patients to pro- grams that support healthy weight loss.56−59 Given that many patients with obesity may be at higher risk of hav- ing multiple comorbid conditions, such as hypertension, diabetes, and osteoarthritis,60 and that these issues may be compounded by unhealthy weight-loss use, providing evidence-based recommendations about healthy weight loss may help in the management of multiple problems, including mental health concerns. Finally, this study parallels other research suggesting endorsement of cer- tain weight-loss behaviors may indicate lower risk of depressive status.61,62 Endorsement of exercise reduced the odds of reporting depression by half, a borderline statistically significant finding, supporting previous evi- dence purporting its protective effect on mental health.61

Limitations This study is not without limitations. First, this study is cross-sectional and cannot comment on causality between weight-loss strategies and depression. Further, theweight- loss strategies are self-reported over the entire previous year, which may be subject to recall bias. It is also possible

248 Chaitoff et al / Am J Prev Med 2019;56(2):241−250

a respondent was no longer using a reported weight-loss method when the PHQ-9 was administered, whichmeans it is possible that the effects of the weight-loss strategy on depressionmay occur after use of the weight-loss strategy. Additionally, the NHANES question concerning diet pill use asks respondents to report if they took diet pills pre- scribed by a doctor, which is vague and could include appropriately taken Food and Drug Administration −approvedweight-lossmedications thatwould not neces- sary qualify as an unhealthy weight-loss method.Weight- loss strategies were categorized based on accepted norms of the literature and expert input, the classification is lim- ited by the NHANES questionnaire, and what is deemed sustainable or healthy may change with medical advance- ments and research. Finally, additional weight-loss strate- gies and the intensity with which one uses a given weight- loss method might also be associated with increased odds of depression. Thus, these results might serve as a guide, but not final recommendations, about how to intertwine weight loss anddiscussions regardingdepression.

CONCLUSIONS

Despite these limitations, this study shows that weight- loss strategies, especially unhealthy ones, are associated with depression in certain groups of adults. Conversa- tions about weight loss may therefore have an opportu- nity to impact both physical as well as mental health. Future research might further characterize this associa- tion and investigate how best to use discussions of weight loss to identify patients at risk of depression while limiting burden in an already over-extended provider population.63

ACKNOWLEDGMENTS No financial disclosures were reported by the authors of this paper. No conflicts of interest were reported by the authors of this paper.

REFERENCES 1. Elgin J, Pritchard M. Gender differences in disordered eating and its

correlates. Eat Weight Disord. 2006;11(3):e96–e101. https://doi.org/ 10.1007/BF03327565.

2. Harring HA, Montgomery K, Hardin J. Perceptions of body weight, weight management strategies, and depressive symptoms among U.S. college students. J Am Coll Health. 2010;59(1):43–50. https://doi.org/ 10.1080/07448481.2010.483705.

3. Weiss EC, Galuska DA, Khan LK, Serdula MK. Weight-control practi- ces among U.S. adults, 2001−2002. Am J Prev Med. 2006;31(1):18–24. https://doi.org/10.1016/j.amepre.2006.03.016.

4. Apovian CM. The obesity epidemic—understanding the disease and the treatment. N Engl J Med. 2016;374(2):177–179. https://doi.org/ 10.1056/NEJMe1514957.

5. Snook KR, Hansen AR, Duke CH, Finch KC, Hackney AA, Zhang J. Change in percentages of adults with overweight or obesity trying to

lose weight, 1988−2014. JAMA. 2017;317(9):971–973. https://doi.org/ 10.1001/jama.2016.20036.

6. U.S. weight loss market worth $66 billion. WebWire. www.webwire. com/ViewPressRel.asp?aId=209054. Published May 4, 2017. Accessed September 12, 2018.

7. Jensen MD, Ryan DH, Apovian CM, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Associ- ation Task Force on Practice Guidelines and the Obesity Society. ACCF/AHA TASK FORCE MEMBERS Subcommittee on Prevention Guidelines. Circulation. 2014;129(suppl 2):S102–S138. https://doi.org/ 10.1161/01.cir.0000437739.71477.ee.

8. Bish CL, Blanck HM, Serdula MK, Marcus M, Kohl HW, Khan LK. Diet and physical activity behaviors among Americans trying to lose weight: 2000 Behavioral Risk Factor Surveillance System. Obes Res. 2005;13(3):596–607. https://doi.org/10.1038/oby.2005.64.

9. Kruger J, Galuska DA, Serdula MK, Jones DA. Attempting to lose weight. Am J Prev Med. 2004;26(5):402–406. https://doi.org/10.1016/j. amepre.2004.02.001.

10. Tsai AG,Wadden TA. Systematic review: an evaluation ofmajor commer- cial weight loss programs in the United States. Ann Intern Med. 2005;142 (1):56–66. https://doi.org/10.7326/0003-4819-142-1-200501040-00012.

11. Davila EP, Kolodziejczyk JK, Norman GJ, et al. Relationships between depression, gender, and unhealthy weight loss practices among over- weight or obese college students. Eat Behav. 2014;15(2):271–274. https://doi.org/10.1016/j.eatbeh.2014.03.010.

12. Kelly-Weeder S, Phillips K, Leonard K, Veroneau M. Binge eating and weight loss behaviors of overweight and obese college students. J Am Assoc Nurse Pract. 2014;26(8):445–451. https://doi.org/10.1002/2327- 6924.12070.

13. Malinauskas BM, Raedeke TD, Aeby VG, Smith JL, Dallas MB. Diet- ing practices, weight perceptions, and body composition: a compari- son of normal weight, overweight, and obese college females. Nutr J. 2006;5(1):11. https://doi.org/10.1186/1475-2891-5-11.

14. Eisenberg D, Nicklett EJ, Roeder K, Kirz NE. Eating disorder symp- toms among college students: prevalence, persistence, correlates, and treatment-seeking. J Am Coll Health. 2011;59(8):700–707. https://doi. org/10.1080/07448481.2010.546461.

15. Lasikiewicz N, Myrissa K, Hoyland A, Lawton CL. Psychological bene- fits of weight loss following behavioural and/or dietary weight loss interventions. a systematic research review. Appetite. 2014;72:123– 137. https://doi.org/10.1016/j.appet.2013.09.017.

16. Fabricatore AN, Wadden TA, Higginbotham AJ, et al. Intentional weight loss and changes in symptoms of depression: a systematic review and meta-analysis. Int J Obes (Lond). 2011;35(11):1363–1376. https://doi.org/10.1038/ijo.2011.2.

17. Kim WK, Shin D, Song WO. Are dietary patterns associated with depression in U.S. adults? J Med Food. 2016;19(11):1074–1084. https://doi.org/10.1089/jmf.2016.0043.

18. Skarupski KA, Tangney CC, Li H, Evans DA, Morris MC. Mediterra- nean diet and depressive symptoms among older adults over time. J Nutr Health Aging. 2013;17(5):441–445. https://doi.org/10.1007/ s12603-012-0437-x.

19. Presnell K, Stice E, Seidel A, Madeley MC. Depression and eating pathology: prospective reciprocal relations in adolescents. Clin Psychol Psychother. 2009;16(4):357–365. https://doi.org/10.1002/cpp.630.

20. Loth KA, MacLehose R, Bucchianeri M, Crow S, Neumark-Sztainer D. Predictors of dieting and disordered eating behaviors from adoles- cence to young adulthood. J Adolesc Health. 2014;55(5):705–712. https://doi.org/10.1016/j.jadohealth.2014.04.016.

21. Lecrubier Y. Widespread underrecognition and undertreatment of anxiety and mood disorders: results from 3 European studies. J Clin Psychiatry. 2007;68(suppl 2):36–41.

22. Kroenke K, Spitzer RL, Williams JBW, Monahan PO, L€owe B. Anxiety disorders in primary care: prevalence, impairment, comorbidity, and

www.ajpmonline.org

Chaitoff et al / Am J Prev Med 2019;56(2):241−250 249

detection. Ann Intern Med. 2007;146(5):317–325. https://doi.org/ 10.7326/0003-4819-146-5-200703060-00004.

23. Akincigil A, Matthews EB. National rates and patterns of depression screening in primary care: results from 2012 and 2013. Psychiatr Serv. 2017;68(7):660–666. https://doi.org/10.1176/appi.ps.201600096.

24. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics. The National Health and Nutrition Examination Survey. www.cdc.gov/nchs/nhanes/index.htm. Updated July 31, 2018. Accessed September 12, 2018.

25. Centers for Disease Control and Prevention (CDC). 2013−2014 Data Documentation, Codebook, and Frequencies: Weight History (WHQ_H). wwwn.cdc.gov/Nchs/Nhanes/2013-2014/WHQ_H.htm. Published March 2016. Accessed September 12, 2018.

26. Neumark-Sztainer D, Wall M, Larson NI, Eisenberg ME, Loth K. Diet- ing and disordered eating behaviors from adolescence to young adult- hood: findings from a 10-year longitudinal study. J Am Diet Assoc. 2011;111(7):1004–1011. https://doi.org/10.1016/j.jada.2011.04.012.

27. Arroll B, Goodyear-Smith F, Crengle S, et al. Validation of PHQ-2 and PHQ-9 to screen for major depression in the primary care population. Ann Fam Med. 2010;8(4):348–353. https://doi.org/ 10.1370/afm.1139.

28. Kroenke K, Spitzer RL. The PHQ-9: a new depression diagnostic and severity measure. Psychiatr Ann. 2002;32(9):509–515. https://doi.org/ 10.3928/0048-5713-20020901-06.

29. Spitzer RL, Kroenke K, Williams JBW, Patient Health Questionnaire Primary Care Study Group. Validation and utility of a self-report ver- sion of PRIME-MD: the PHQ primary care study. JAMA. 1999;282 (18):1737–1744. https://doi.org/10.1001/jama.282.18.1737.

30. Kroenke K, Spitzer RL, Williams JBW. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606–613. https://doi.org/10.1046/j.1525-1497.2001.016009606.x.

31. Manea L, Gilbody S, McMillan D. Optimal cut-off score for diagnos- ing depression with the Patient Health Questionnaire (PHQ-9): a meta-analysis. Can Med Assoc J. 2012;184(3):e191–e196. https://doi. org/10.1503/cmaj.110829.

32. Akhtar-Danesh N, Landeen J. Relation between depression and socio- demographic factors. Int J Ment Health Syst. 2007;1:4. https://doi.org/ 10.1186/1752-4458-1-4.

33. Cole MG, Dendukuri N. Risk factors for depression among elderly com- munity subjects: a systematic review and meta-analysis. Am J Psychiatry. 2003;160(6):1147–1156. https://doi.org/10.1176/appi.ajp.160.6.1147.

34. Luppino FS, de Wit LM, Bouvy PF, et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry. 2010;67(3):220–229. https://doi.org/ 10.1001/archgenpsychiatry.2010.2.

35. Paige E, Korda RJ, Banks E, Rodgers B. How weight change is mod- elled in population studies can affect research findings: empirical results from a large-scale cohort study. BMJ Open. 2014;4(6):e004860. https://doi.org/10.1136/bmjopen-2014-004860.

36. Carr D, Jaffe K. The psychological consequences of weight change tra- jectories: evidence from quantitative and qualitative data. Econ Hum Biol. 2012;10(4):419–430. https://doi.org/10.1016/j.ehb.2012.04.007.

37. Wittayanukorn S, Qian J, Hansen RA. Prevalence of depressive symp- toms and predictors of treatment among U.S. adults from 2005 to 2010. Gen Hosp Psychiatry. 2014;36(3):330–336. https://doi.org/ 10.1016/j.genhosppsych.2013.12.009.

38. Wong J, Motulsky A, Eguale T, Buckeridge DL, Abrahamowicz M, Tamblyn R. Treatment indications for antidepressants prescribed in primary care in Quebec, Canada, 2006−2015. JAMA. 2016;315 (20):2230–2232. https://doi.org/10.1001/jama.2016.3445.

39. Mojtabai R, Olfson M. Proportion of antidepressants prescribed with- out a psychiatric diagnosis is growing. Health Aff. 2011;30(8): 1434–1442. https://doi.org/10.1377/hlthaff.2010.1024.

40. Lee PH, Burstyn I. Identification of confounder in epidemiologic data contaminated by measurement error in covariates. BMC Med

February 2019

Res Methodol. 2016;16:54. https://doi.org/10.1186/s12874-016- 0159-6.

41. Bursac Z, Gauss CH, Williams DK, Hosmer DW. Purposeful selection of variables in logistic regression. Source Code Biol Med. 2008;3:17. https://doi.org/10.1186/1751-0473-3-17.

42. Centers for Disease Control and Prevention (CDC). Overview of NHANES Survey Design and Weights. www.cdc.gov/Nchs/tutorials/ environmental/orientation/sample_design/index.htm. Updated May 3, 2013. Accessed August 1, 2017.

43. IBM. IBM SPSS Complex Samples 24. ftp://public.dhe.ibm.com/soft- ware/analytics/spss/documentation/statistics/24.0/en/client/Manuals/ IBM_SPSS_Complex_Samples.pdf. Published 2016. Accessed September 12, 2018.

44. Kolotkin RL, Crosby RD, Williams GR. Health-related quality of life varies among obese subgroups. Obes Res. 2002;10(8):748–756. https:// doi.org/10.1038/oby.2002.102.

45. van Nunen AMA, Wouters EJM, Vingerhoets AJJM, Hox JJ, Geenen R. The health-related quality of life of obese persons seeking or not seeking surgical or non-surgical treatment: a meta-analysis. Obes Surg. 2007;17(10):1357–1366. https://doi.org/10.1007/s11695-007-9241-9.

46. Kivim€aki M, Hamer M, Batty GD, et al. Antidepressant medication use, weight gain, and risk of type 2 diabetes: a population-based study. Dia- betes Care. 2010;33(12):2611–2616. https://doi.org/10.2337/dc10-1187.

47. Domecq JP, Prutsky G, Leppin A, et al. Drugs commonly associated with weight change: a systematic review and meta-analysis. J Clin Endocrinol Metab. 2015;100(2):363–370. https://doi.org/10.1210/ jc.2014-3421.

48. Tiggemann M, Kuring JK. The role of body objectification in disor- dered eating and depressed mood. Br J Clin Psychol. 2004;43(3): 299–311. https://doi.org/10.1348/0144665031752925.

49. Tiggemann M, Williams E. The role of self-objectification in disor- dered eating, depressed mood, and sexual functioning among women. Psychol Women Q. 2012;36(1):66–75. https://doi.org/10.1177/ 0361684311420250.

50. Gillen MM, Markey CN, Markey PM. An examination of dieting behaviors among adults: links with depression. Eat Behav. 2012; 13(2):88–93. https://doi.org/10.1016/j.eatbeh.2011.11.014.

51. Hebl MR, Heatherton TF. The stigma of obesity in women: the differ- ence is black and white. Pers Soc Psychol Bull. 1998;24(4):417–426. https://doi.org/10.1177/0146167298244008.

52. Keddie AM. Associations between severe obesity and depression: results from the National Health and Nutrition Examination Survey, 2005−2006. Prev Chronic Dis. 2011;8(3):A57.

53. Massie MJ. Prevalence of depression in patients with cancer. J Natl Cancer Inst Monogr. 2004;2004(32):57–71. https://doi.org/10.1093/ jncimonographs/lgh014.

54. Katon WJ. Epidemiology and treatment of depression in patients with chronic medical illness. Dialogues Clin Neurosci. 2011;13(1):7–23.

55. Nouwen A, Winkley K, Twisk J, et al. Type 2 diabetes mellitus as a risk factor for the onset of depression: a systematic review and meta-analy- sis. Diabetologia. 2010;53(12):2480–2486. https://doi.org/10.1007/ s00125-010-1874-x.

56. Pollak KI, Coffman CJ, Alexander SC, et al. Predictors of weight loss communication in primary care encounters. Patient Educ Couns. 2011;85(3):175–182. https://doi.org/10.1016/j.pec.2011.03.001.

57. Aspy CB, Mold JW, Thompson DM, et al. Integrating screening and interventions for unhealthy behaviors into primary care practices. Am J Prev Med. 2008;35(5 suppl):373–380. https://doi.org/10.1016/j. amepre.2008.08.015.

58. Ferraro ZM, Patterson S, Chaput JP. Unhealthy weight control practi- ces: culprits and clinical recommendations. Clin Med Insights Endocri- nol Diabetes. 2015;8:7–11. https://doi.org/10.4137/CMED.S23060.

59. Lewis AL, Aveyard P, Jebb SA. Brief interventions for weight loss in primary care. Curr Obes Rep. 2013;2(4):341–347. https://doi.org/ 10.1007/s13679-013-0073-8.

250 Chaitoff et al / Am J Prev Med 2019;56(2):241−250

60. Must A, Spadano J, Coakley EH, Field AE, Colditz G, Dietz WH. The disease burden associated with overweight and obesity. JAMA. 1999;282(16):1523–1529. https://doi.org/10.1001/jama.282.16.1523.

61. Cooney GM, Dwan K, Greig CA, et al. Exercise for depression. Cochrane Database Syst Rev. 2013;(9):CD004366. https://doi.org/ 10.1002/14651858.CD004366.pub6.

62. Li Y, Lv M-R, Wei Y-J, et al. Dietary patterns and depression risk: a meta-analysis. Psychiatry Res. 2017;253:373–382. https://doi.org/ 10.1016/j.psychres.2017.04.020.

63. Linzer M, Poplau S. Building a sustainable primary care workforce: where do we go from here? J Am Board Fam Med. 2017;30(2): 127–129. https://doi.org/10.3122/jabfm.2017.02.170014.

www.ajpmonline.org