- Email: [email protected]
Validation of the intuitive Eating Scale for pregnant women
Appetite 112 (2017) 201e209
Contents lists available at ScienceDirect
Appetite journal homepage: www.elsevier.com/locate/appet
Validation of the intuitive Eating Scale for pregnant women Sajeevika Saumali Daundasekara a, *, Anitra Danielle Beasley b, Daniel Patrick O'Connor a, McClain Sampson c, Daphne Hernandez a, Tracey Ledoux a a
Department of Health and Human Performance, University of Houston, 3875 Holman Street, Garrison Gym, Room 104, Houston, TX 77204-6015, USA Obstetrics and Gynecology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA c Graduate College of Social Work, The University of Houston, 110HA Social Work Building, Houston, TX 77204-4013, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 August 2016 Received in revised form 30 January 2017 Accepted 2 February 2017 Available online 3 February 2017
Pre-pregnancy maladaptive eating behaviors have predicted inadequate or excess gestational weight gain and poor dietary intake during pregnancy, but little is known about effects of pre-pregnancy adaptive eating behaviors on pregnancy outcomes. The purpose of this study was to produce a valid and reliable measure of adaptive pre-pregnancy eating behaviors for pregnant women using the Intuitive Eating Scale. Data were collected from 266 pregnant women, aged 18 and older who were attending a private prenatal clinic at Texas Children's Hospital Pavilion for Women in Houston, TX using selfadministered questionnaires. Confirmatory factor analysis was performed to validate the factor structure of the Intuitive Easting Scale (IES). Concurrent validity was determined using correlations between the three subscale scores [unconditional permission to eat (UPE), eating for physical not emotional reasons (EPR), and relying on hunger/satiety cues (RIH)], perinatal depression status (Edinburgh Postnatal Depression Scale), and pre-pregnancy body mass index (BMI) calculated from self-reported height and weight. After discarding 6 items, the second order model did not fit the data, however, the first order model with three latent factors had reasonable fit (RMSEA ¼ 0.097, CFI ¼ 0.961, TLI ¼ 0.951 and WRMR ¼ 1.21). The internal consistency of the scale was confirmed by Cronbach's alphas (UPE ¼ 0.781, EPR ¼ 0.878 and RIH ¼ 0.786). All subscale scores were inversely related to perinatal depression status. EPR and RIH subscale scores were inversely related to pre-pregnancy BMI, supporting the measure's validity. Among pregnant women, the revised 15 item pre-pregnancy IES (IES-PreP) should be used to evaluate pre-pregnancy adaptive eating behaviors. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Intuitive eating Pregnant women Intuitive eating scale Validation Confirmatory factor analysis
1. Introduction The Institute of Medicine, Gestational weight gain (GWG) recommendations for pregnancy is based on the pre-pregnancy body mass index (BMI). Women who are normal weight prior to pregnancy are recommended to gain 11.5e16 kg, the recommendation for overweight women is 7e11.5 kg and for obese women it is 5e9 kg. Inadequate and Excessive GWG are used to describe weight gain below and above this recommendation respectively. Excess GWG and poor dietary intake are related to adverse health outcomes including postpartum weight retention, gestational
* Corresponding author. E-mail addresses: [email protected] (S.S. Daundasekara), Anitra.Beasley@ bcm.edu (A.D. Beasley), [email protected] (D.P. O'Connor), mmsampson@ uh.edu (M. Sampson), [email protected] (D. Hernandez), TALedoux@uh. edu (T. Ledoux). http://dx.doi.org/10.1016/j.appet.2017.02.001 0195-6663/© 2017 Elsevier Ltd. All rights reserved.
diabetes, pregnancy induced hypertension, pre term deliveries, fetal growth restriction, fetal macrosomia, large for gestational age infants, neonatal hypoglycemia, and infant and childhood obesity (Dietrich, Federbusch, Grellmann, Villringer, & Horstmann, 2014; Grieger, Grzeskowiak, & Clifton, 2014; Margerison Zilko, Rehkopf & Abrams, 2010; Thangaratinam et al., 2012). According to 2012e2013 United States data, prevalence of inadequate GWG ranged from 12.6% to 25.5% and prevalence of excessive GWG ranged from 38.2% to 54.7%, while only 32.1% of women had adequate weight gain according to Institute of Medicine recommendations (Deputy, Sharma, & Kim, 2015). Additionally, a recent study showed that on average, pregnant women do not adequately conform to federal dietary guidelines (United States Department of Agriculture (USDA) Center for Nutrition Policy Promotion Promotion, 1995) with 40% of pregnant women not meeting the minimum recommended number of servings of most food groups, dietary fiber, calcium, vitamin D, iron and folate (Pick, Edwards,
202
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
Moreau, & Ryan, 2005). It is important to identify modifiable determinants of poor diet quality, excess GWG, and negative pregnancy outcomes that clinicians can assess early in pregnancy or before so that early interventions to target these factors can be developed. In non-pregnant samples, eating behaviors regarding what, when, and how much to eat influence dietary quality and weight. Common eating behaviors linked to obesity and poor diet quality among non-pregnant adults are emotional eating, external eating and restrained eating (Van Strien, Frijters, Roosen, Knuiman-Hijl, & Defares, 1985; Van Strien, Schippers, & Cox, 1995). Eating behaviors leading to negative health consequences may be considered maladaptive (Tylka, 2006). For example, restrained eating is characterized by rigid restrictions in caloric or food intake with occasional episodes of disinhibited eating (Herman & Polivy, 1980) and is positively related to obesity, weight cycling and binge eating s & Saldan ~ a, 2014; Johnson, Pratt, & Wardle, 2012). Pre(Andre pregnancy maladaptive eating behaviors (e.g. restraint, disordered eating) have been related to excess and inadequate GWG and poor dietary intake during pregnancy (Conway, Reddy, & Davies, 1999; Mumford, Siega-Riz, Herring, & Evenson, 2008; Sollid, Wisborg, Hjort, & Secher, 2004). Women with eating disorders such as anorexia nervosa and bulimia nervosa prior to pregnancy were found to be at higher risk of impaired pregnancy outcomes including pre-term deliveries, small for gestational age babies and low birth weight infants (Micali, Treasure, & Simonoff, 2007; Sollid et al., 2004; Ward, 2008). Pre-pregnancy maladaptive eating behaviors have been linked with negative pregnancy outcomes, which implies these maladaptive eating behaviors should be treated before pregnancy. However, there is little information available as to what types of pre-pregnancy eating behaviors should replace pre-pregnancy maladaptive eating behaviors. Intuitive eating (IE) was originally conceptualized by Evelyn Tribole and Elyse Resch, who are both clinical dietitians. The IE theory posits that strict food rules that ignore ‘body wisdom’ (i.e., hunger/satiety cues), lead to preoccupation with food and loss of control eating prompted by emotions and external cues. Tribole and Resch also suggest that honoring ‘body wisdom’ will normalize eating and reduce preoccupation with food and loss of control eating (Tribole & Resch, 2003). Among non-pregnant adults, IE has been related to lower BMI, stable weight, and fewer dieting behaviors and food anxieties (Schaefer & Magnuson, 2014; Smith & Hawks, 2006; Tylka & Wilcox, 2006; Tylka, 2006; Tylka & Kroon Van Diest, 2013). Interventions that promote IE among nonpregnant overweight and obese populations show weight maintenance and improved body image (Cole & Horacek, 2010; Katzer et al., 2008), and cross sectional studies of IE show higher levels of IE have been positively related to psychological well-being and negatively related to BMI making this an adaptive eating behavior among the general adult population (Augustus-Horvath & Tylka, 2011; Tylka & Wilcox, 2006; Tylka, 2006; Tylka & Kroon Van Diest, 2013). However, to date, there are no studies investigating the relationship between pre-pregnancy IE and pregnancy outcomes. To conduct this type of research, a validated measure of prepregnancy IE is required. If pre-pregnancy IE is protective of negative pregnancy outcomes including excess GWG, then perhaps this measure could also be used to screen for pregnant women at risk of pregnancy complications so that interventions can be initiated as early in pregnancy as possible. The Intuitive Eating Scale (IES) is a 21-item scale with three subscales:(1) unconditional permission to eat (UPE, 9 items), (2) eating for physical rather than emotional reasons (EPR, 6 items), and (3) reliance on internal hunger and satiety cues to determine when and how much to eat (RIH, 6 items). This scale was developed and the 3-factor second order factor structure based on Tribole and
Resch's theory and was validated among predominantly female college students (Tylka, 2006). Some studies have shown that the original measure is valid with other populations but requires minor adjustments in scoring given a different factor structure (e.g. adolescents) (Dockendorff, Petrie, Greenleaf, & Martin, 2012). Before the IES can be used to assess pre-pregnancy intuitive eating among pregnant women, the validity of this measure with this population should be confirmed. The goal of this study was to confirm the validity and reliability of the IES to test pre-pregnancy intuitive eating among pregnant women. Following procedures used by Tylka (2006) the factor structure of the scale was analyzed using a confirmatory factor analysis to determine: (1) the overall fit of data to the scale model, (2) the item loadings, and (3) the relationship between the latent factors. We hypothesized that the IES items would load on their respective latent factors as identified by Tribole and Resch (Tribole & Resch, 2003; Tylka, 2006). In addition, the latent factors would be related, load on the higher order IE factor, and the overall model would provide adequate fit to the data in a sample of pregnant women. According to the previous findings we also hypothesized that the total IES scores would demonstrate concurrent validity and be inversely related to pre-pregnancy BMI and perinatal depression status. 2. Methodology 2.1. Procedure This study was conducted at the Texas Children's Hospital Pavilion for Women in Houston, TX, USA. The participants were pregnant women attending a private prenatal clinic at this location. Data for the study were collected from spring 2013 to summer 2014. The participant inclusion criteria for the study were women 18 years of age or older, singleton pregnancy confirmed by a physician, and willingness and ability to complete the survey in English. Women below 18 years or with current multiple pregnancy were excluded from the study. A research coordinator distributed recruitment flyers to women entering the clinic waiting room, and women who were interested in participating in the survey were given a survey packet including: a screening checklist, a consent form, a resource list, and the questionnaire. The study was described to potential participants as an investigation of health behaviors of pregnant women. A self-screening checklist was used in determining participant's eligibility. Written informed consent was obtained from participants while they were waiting for their appointments in the waiting room. Participants were asked to complete the survey in the same waiting room or at the examination room. No financial compensation was given to participants, however a prenatal care resource list containing contact information of community agencies, or places to get help during pregnancy was given to each participant to help find support for specific needs they have regarding the pregnancy. The survey packet was given to 300 interested and eligible participants. Nineteen women returned the questionnaires without completing any of the items and 15 women answered 90% of the given measures, so they were excluded. The final was 266 participants. This study was approved by the Institutional Review Boards of the University of Houston and Baylor College of Medicine. 2.2. Socio-demographic data and body mass index Socio-demographic data including age, race, marital status, education, employment, gestational age and household income were collected through a questionnaire consisting of items taken
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
from two existing national surveys, the Brief Risk Factor Surveillance System (CDC, 2015a) and the Pregnancy Risk Assessment Management System by Center for Disease Control and Prevention (CDC 2015b). Pre-pregnancy BMI was calculated using the selfreported height and weight according to CDC guidelines (CDC, 2015c). 2.3. Depression Perinatal depression symptoms were evaluated using the 10 item Edinburgh Postnatal Depression Scale (EPDS) (Cox, Holden, & Sagovsky, 1987). This is a self-reported scale originally developed to use with postpartum women and later have been validated for pregnant women, to identify risk for perinatal depression (e.g., “I have been able to laugh and see the funny side of things”; “I have blamed myself unnecessarily when things went wrong”). Participants are asked to select the answer, that best represent their situation within the past 7 days. Each item is scored on a four point scale ranging from 0 to 3 (0 ¼ most of the time/very often to 3 ¼ never/not at all). Negatively stated items were reversed scored, so that high scores indicate higher perinatal depression status. A total score higher than 13 indicate depression levels consistent with depressive disorders. In this analysis we used the total EPDS score as a continuous measure of perinatal depression status within the last week. 2.4. Pre pregnancy eating behavior The 21 item IES (Tylka, 2006) was adapted for the current study with pregnant women to measure pre pregnancy eating behaviors. The original IES was developed to assess three constructs thought to make up intuitive eating including unconditional permission to eat (UPE; e.g., “I tried to avoid certain foods high in fat, carbohydrates, or calories”), eating for physical rather than emotional reasons (EPR; e.g., “I usually stopped eating when I felt full [not overstuffed]”), and reliance on hunger and satiety cues to determine when and how much to eat (RIH; e.g., “I could tell when I was slightly full”). In the current study, the IES was modified by adding the term “Before pregnancy” before each item (e.g., “Before pregnancy, I tried to avoid certain foods high in fat, carbohydrates, or calories”). The format of the response is a five point Likert-type scale (1 ¼ strongly disagree and 5 ¼ strongly agree). Some of the scale items are negatively stated and some are positively stated. In calculating the scale scores, the negative items were reversed scored so that high scores on the total measure and subscales indicate greater intuitive eating. The individual scores of items under each subscale were summed to obtain subscale scores. The total IES score is the sum of three subscale scores. The total subscale scores were divided by the number of items in each subscale to get the mean subscale scores. This resulted in possible mean scores ranges from 1 to 5. 2.5. Statistical analysis Statistical analyses were performed using IBM SPSS statistics for Windows, version 22.0 and Mplus, version 7.4 was used for the confirmatory factor analysis (CFA) using a covariance matrix. The factor structure of the modified IES was analyzed with two CFA models to determine the overall fit of the data to the scale model and to assess whether the items load on their hypothesized latent factors. CFA was used as the IES has a well-established factorial structure based on the concepts of IE (Tylka, 2006) and as our objective was to evaluate the validity of this scale for pregnant population. The weighted least squares with mean and variance
203
adjustment (WLSMV) estimation method, which is recommended for ordinal data was used in the CFA (Bayol, Simbi, Bertrand, & Stickland, 2008; Byrne, 2012). Sample size was adequate to conduct a factor analysis (N ¼ 266) which is about 12:1 participantto-item ratio (Byrne, 2012). The 21 items of the modified scale were the indicators of the three first-order latent factors UPE, EPR, and RIH. We used a first-order CFA to determine whether each item loaded onto their respective first-order factor and to determine whether the first-order factors are correlated to each other as hypothesized. After confirming the first-order model, a second-order CFA was conducted to determine whether the first-order latent factors loaded on the higher-order intuitive eating factor. One item loading for each first order factor was set equal to 1.0 to provide a scale for the first-order factors, and the same strategy was used for the second-order loadings to provide a scale for the second-order factor (Chen, Sousa, & West, 2005). Adequacy of fit was determined by the indices; the Comparative Fit Index (CFI), the TuckereLewis Index (TLI), the root-mean-square error of approximation (RMSEA) and weighted root mean square of residual (WRMR) (Hu & Bentler, 1999). According to Hu and Bentler RMSEA values less than 0.05 indicates a close fit, and Browne and Cudeck suggested that values in the range of 0.05e0.08 indicate fair fit; values above 1.0 suggests poor fit (Browne & Cudeck, 1993). Values above 0.9 for CFI and TLI indicates a reasonably good fit (>0.95 close fit), and according to Yu (2002), values closer to 1.0 for WRMR indicates a reasonable fit (<0.95 good fit). The internal consistency of each subscale was evaluated using the Cronbach's alpha. The concurrent validity was examined through correlations with pre-pregnancy BMI and the perinatal depression status. According to previous studies we expect the IES scores to be negatively related to prepregnancy BMI and the depression status (Dockendorff et al., 2012; Tylka, 2006). 3. Results 3.1. Participants The socio-demographics and pre-pregnancy BMI of the women are given in Table 1. The mean age of the sample was 30.69 ± 4.79 years, mean pre-pregnancy BMI was 25.31 ± 6.31. The mean gestational age of the women was 30.71 ± 6.73 weeks but it ranged from 8 to 41 weeks. 3.2. Confirmatory factor analysis Missing data in the modified IES were evaluated using Little's MCAR test before conducting the CFA and was found to be consistent with being missing completely at random. Therefore, the expectation maximization (EM) method was used to replace the missing values as previously conducted in the validation of IES for adolescents (Dockendorff et al., 2012). Some of the items deviated from a normal distribution, which is expected with ordinal level data therefore the WLSMV for non-normal categorical data was used. The fit statistics of the first order model indicated that the original model (Fig. 1) had a poor fit to the data (RMSEA ¼ 0.135, CFI ¼ 0.848, TLI ¼ 0.828 and WRMR ¼ 2.13). Therefore, the factor loading for the first-order model was evaluated to identify items that did not load strongly onto their hypothesized latent factor. Using methodology from the original IES validation study (Tylka, 2006), the items with a factor loading below 0.45 on their hypothesized latent factor were removed. Items 1, 4, 5, and 19 were discarded from the UPE subscale due to poor factor loadings (0.126, 0.144, 0.433 and 0.120, respectively). The resulting model did not have a significant improvement and further evaluation of item
204
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
Table 1 Sample characteristics. Characteristics Race/Ethnicity (N ¼ 266)
Marital status (N ¼ 266) Education (N ¼ 262)
Employment (N ¼ 264) Household income (N ¼ 249) Pre-pregnancy BMI (N ¼ 261)
Non-Hispanic White Non-Hispanic Black Hispanic Asian American Indian or Alaska Native Others Married Others Some high school High School Graduate Some College College Graduate Employed Unemployed < $50,000 $50,000 or More Underweight (<18.5) Normal (18.5e24.9) Overweight (25e29.9) Obese (30)
N
Percentage
140 24 66 29 1 6 238 28 2 18 47 195 188 76 42 207 10 145 68 38
52.6 9.0 24.8 10.9 0.4 2.3 89.5 10.5 0.8 6.9 17.9 74.4 71.2 28.8 16.9 83.1 3.8 55.6 26.0 14.6
Abbrevations: BMI ¼ Body Mass Index, N¼Number of participants.
loading and modification indices indicated that item 14 had a poor factor loading (0.398), and item 2 loaded strongly on the hypothesized latent factor (EPR) as well as on RIH. Therefore, these two items were also removed from the scale. These adjustments significantly improved the CFI, TLI and WRMR (0.927, 0.912 and 1.66) and resulted in a minor improvement in RMSEA (0.131) even though fit statistics were not adequate. Even after discarding the six items, the model did not fit the data adequately. In the original model we hypothesized that the error correlations among all the item pairs are zero. The modification indices revealed that there are strong error correlations among several items (7 and 8, 13 and 12, 18 and 9). This model with 15 items, 3 correlated latent factors and 3 correlated errors, had 120 data points and 36 parameters to be estimated. Therefore, model is overidentified model with 84 degrees of freedom. Addition of these error correlations to the model significantly improved the model fit indicating that this new first order model fit the data well (RMSEA ¼ 0.097, CFI ¼ 0.961, TLI ¼ 0.951 and WRMR ¼ 1.21). The three first order factors showed significant moderate relationship to one another (estimates 0.3e0.7, P < 0.001). This model proved our hypothesis 1. That overall model fit to data adequately, 2. The remained items significantly and adequately loaded onto their hypothesized latent factors and 3. There are significant relationships between the three latent factors. Next we conducted a second order factor analysis by adding the higher order latent factor, “Intuitive Eating” to the final model containing 15 items and 3 error correlations. The residual variance of the latent factor EPR in this second order model was negative (0.033, p ¼ 0.769) indicating a model misspecification suggesting there is no support for a second order IE factor among pregnant women. Therefore, we had to reject our final hypotheses that the three latent factors load on the higher order IE factor. The results are reported for the final first order model that showed adequate fit to the data. The standardized factor loadings of the final first-order model with error correlations are given in Fig. 2. In the final validated 15 item pre-pregnancy IES which will be referred to as IES-PreP in later text, the UPE subscale consists of four items (mean ¼ 3.36, SD ¼ 1.01). The EPR subscale has five items (mean ¼ 3.80, SD ¼ 0.94), and the RIH subscale has six items (mean ¼ 3.84, SD ¼ 0.62). The Cronbach's alphas of three subscales were 0.781, 0.878 and 0.786, respectively, indicating a satisfactory internal consistency among the items in each subscale (Bland &
Altman, 1997). The items under each subscale of the IES-PreP are given in Table 2. 3.3. Concurrent validity The UPE factor was significantly negatively related to the perinatal depression status among pregnant women, but there was no significant correlation between this factor and the pre-pregnancy BMI. Both the EPR and RIH factors showed significant moderate inverse correlation with both the perinatal depression status and pre-pregnancy BMI (Table 3). 4. Discussion This study was designed to test the validity of the IES (Tylka, 2006) to measure pre-pregnancy intuitive eating behaviors among pregnant women. Intuitive eating is considered an adaptive eating behavior made up of three concepts: unconditional permission to eat, eating for physical rather than emotional reasons, and reliance on hunger/satiety to inform food choices (Tribole & Resch, 2003). Among pregnant women a first order model of prepregnancy intuitive eating behaviors provided an adequate fit to the data while the second order factor model did not fit the data, which is incongruent with the intuitive eating theory and results from IES validity research on college samples (Tylka, 2006). According to Tylka, second order model provided an adequate fit to data from college women (CFI ¼ 0.91, TLI ¼ 0.90, RMSEA ¼ 0.08 and SRMR ¼ 0.07). In other words, among college students, the total IES score represents an overall intuitive eating construct made up of three subscales. However, in this sample of 266 pregnant women, the IES total score has no meaning, and the three subscales represent related but distinct eating behaviors. While these results were not hypothesized based on theory or prior research with college students, the findings reflect mounting evidence that the IES does not represent an overall intuitive eating construct for all individuals. A CFA on IES data from adolescents confirmed a first order factor structure (Dockendorff et al., 2012). The clinical and research implication for this is that the modified IES for pregnant women should be scored as three separate but related subscales not as a total overall score. The theoretical implication of this is that intuitive eating as an overall construct representing unconditional permission to eat, eating for physical
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
205
Fig. 1. The initial first order model of the Intuitive Eating Scale for pregnant women (the standardized coefficients for the confirmatory factor analysis e all the loadings were statistically significant p < 0.05 except for ies 19).
206
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
Fig. 2. The final three factor first order model of the Intuitive Eating Scale for pregnant women (the standardized coefficients for the confirmatory factor analysis e all the loadings were statistically significant p < 0.001).
rather than emotional reasons, and relying on hunger/satiety to inform eating behaviors may not exist among certain segments of the population. Six items had low factor loadings on the hypothesized latent factors, and therefore, were dropped from the respective subscales. In the UPE subscale the items that did not have adequate loadings seemed to be related to having strict food rules (e.g.,
“Before pregnancy, I tried to avoid certain foods high in fat, carbohydrates or calories” and “Before pregnancy, there were forbidden foods that I didn't allow myself to eat”). The items that remained in this scale were related to negative feelings about and lack of control over eating unhealthy foods (e.g., “Before pregnancy, I would get mad at myself for eating something unhealthy” and “Before pregnancy, I felt guilty if I ate a certain food that was
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
207
Table 2 The items in the IES-PreP. Unconditional permission to eat-UPE (4 items)
Eating for physical rather than emotional reasons-EPR (5 items)
Reliance on hunger and satiety cues to determine when and how much to eat -(6 items)
IES 9a- Before pregnancy, I would get mad at myself for eating something unhealthy IES 18- Before pregnancy, I felt guilty if I ate a certain food that was high in calories, fat or carbohydrates IES 20- Before pregnancy, I didn't trust myself around fattening foods IES 21- Before pregnancy, I didn't keep certain foods at home because I thought that I would lose control and eat them
IES 3- Before pregnancy, I found myself eating when I felt emotional, even when I wasn't physically hungry IES 6- Before pregnancy, I found myself eating when I was bored, even when I was not physically hungry
IES 7- Before pregnancy, I could tell when I was slightly full IES 8- Before pregnancy, I could tell when I was slightly hungry
IES 10- Before pregnancy, I found myself eating when I was lonely, even when I was not physically hungry IES 16- Before pregnancy, I used food to help me soothe my negative emotions
IES 11- Before pregnancy, I trusted my body to tell me when to eat IES 12-Before pregnancy, I trusted my body to tell me what to eat
IES 17- Before pregnancy, I found myself eating when I was stressed out, even when I was not physically hungry
IES 13- Before pregnancy, I trusted my body to tell me how much to eat IES 15- Before pregnancy, when I ate, I could tell when I was getting full
Abbreviation; IES, Intuitive eating scale. a The IES numbers is given according to the item number in the original 21 item scale, which is used to refer to these items in the text.
Table 3 Correlations among the measures of the study (N ¼ 266). Measures 1. 2. 3. 4. 5.
a
Unconditional permission to eat Eating for physical reasonsb Reliance on internal hunger/satiety cuesc Pre-pregnancy BMId Perinatal Depressione
Mean
SD
1
2
3
4
5
3.37 3.80 3.85 25.31 6.30
1.01 0.94 0.63 6.31 4.33
0.546** 0.197** 0.090 0.217**
e 0.323** 0.240** 0.320**
e 0.272** 0.220**
e 0.165**
e
Abbrevations: SD, Standard Deviation and BMI, Body Mass Index. Note. ** Correlation is significant at the 0.01 level (2 tailed). a,b,c based on the 15 item IES-PreP, possible scores range 1e5. a 4 item subscale. b 5 item subscale. c 6 item subscale. d calculated using self-reported pre-pregnancy height and weight (BMI ¼ weight (kg)/height2(m)). e perinatal depression status within the past week is measured by 10 item Edinburgh Postnatal Depression Scale, possible scores range 0e30.
high in calories, fat or carbohydrates”). It is possible the participants in this study did not have strict food rules prior to pregnancy, but did feel guilty about eating what they perceived to be unhealthy foods. When using the IES for pre-pregnancy, it may be best to call this subscale “lack of negative feelings about eating unhealthy foods” as the remaining items do not reflect unconditional eating. In order to capture “unconditional eating” among this population, the discarded items from the UPE subscale might need rephrasing and further studying. One item (i.e., “Before pregnancy, I usually stopped eating when I felt full, not overstuffed”) loaded on the EPR factor in the original validation studies but in our analysis it had a similar factor loading on the RIH factor. As explained by Dockendorff and others this item is different from the other items in the EPR factor, which are about eating for reasons other than physical hunger (e.g., “Before pregnancy, I found myself eating when I felt emotional, even when I wasn't physically hungry” and “Before pregnancy, I used food to help me soothe my negative emotions”) and seems more consistence with the items in the RIH factor which are about understanding internal hunger and satiety cues (e.g. “Before pregnancy, I could tell when I was slightly full” and “Before pregnancy, I trusted my body to tell me how much to eat”). Therefore, we decided to remove this item from the scale to avoid overlapping of subscale scores. It was hypothesized that the error terms would not be correlated. However, there was significant correlation among the error terms between items 7 and 8, items 9 and 18, and items 12 and 13. This indicates that each pair of item shares common variance that
is unrelated to the respective constructs represented by those items such as item wording or format. For example, items 7 and 8 both share the following phrase: “I could tell when I was___” (see Table 2). Some of the consistency that respondents had when answering those two items may have been related to that common phrase, rather than the intended construct. The final model fit improved significantly by inclusion of correlated errors because they accounted for additional systematic variation (Byrne, 1998, pp. 163e191), but in our study allowing some errors to correlate did not materially affect the model's individual parameter estimates. We hypothesized that pre-pregnancy BMI and perinatal depression status would be negatively related to the three subscale scores and the subscale scores would be correlated to each other. IE has been considered an adaptive eating behavior partially because it is inversely related to BMI in non-pregnant samples; therefore, it was expected to have the same relationship with pre-pregnancy BMI in this study. IE in non-pregnant populations has been related to better psychological status (Augustus-Horvath & Tylka, 2011; Tylka & Wilcox, 2006; Tylka, 2006; Tylka & Kroon Van Diest, 2013), therefore, pre-pregnancy IE was expected to be inversely related to prenatal depressive symptoms. Our results revealed that UPE factor was not significantly correlated with pre-pregnancy BMI which was not consistent with Tylka's findings (Tylka, 2006). However, the remaining valid items on this subscale sufficiently changed the construct represented from UPE to negative feelings associated with eating unhealthy foods. Therefore, it seems negative feelings associated with eating
208
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
unhealthy foods is not related to BMI. A recent IE study done with female students using the 21 item IES has also reported that UPE subscale scores were not significantly correlated with BMI (Herbert, Blechert, Hautzinger, Matthias, & Herbert, 2013). Women who eat for physical hunger without emotional reasons and women who relayed more on hunger and satiety cues to control their eating had lower BMIs, which was consistent with Tylka's results (Tylka, 2006) suggesting concurrent validity. All three subscale scores showed moderate inverse correlation with perinatal depression scores indicating that women who showed higher pre-pregnancy IE related eating behaviors had lower perinatal depression symptomology. This result is consistent with the eating behavior theory and previous studies where adaptive eating behaviors have been shown to predict better psychological status (Dockendorff et al., 2012; Tylka, 2006). According to Tylka among college women, UPE was moderately related to selfesteem and satisfaction with life, EPR was strongly related to optimism and moderately related to self-esteem and satisfaction with life, and RIH was moderately related to self-esteem, satisfaction with life and optimism. Among adolescents all three subscale scores were inversely related to feeling worthless, and/or sad/ depressed (Dockendorff et al., 2012). Also, the subscale scores were positively correlated with each other as expected. The main limitation of this study is recall bias. Gestation age of participants at the time of the study, ranged from 8 to 41 weeks. Thus, women had to recall their eating behaviors 8e41 weeks prior. In the current measure, no instructions were given to participants to recall their eating behaviors within a certain time period. Researchers, who conducted the few studies of pre-pregnancy eating behaviors (Conway et al., 1999; Laraia, Epel, & Siega-Riz, 2013; Mumford et al., 2008), modified existing eating behavior questionnaires by simply adding “Before pregnancy …” and changing the verb tense in each item to past tense. We followed this model for this study. To overcome recall bias, future researchers would need to test a general sample of adult females with the IES and then follow up with those who become pregnant to determine the relationship between pre-pregnancy intuitive eating behaviors and pregnancy outcomes. But there are many limitations to conducting a prospective study of pregnancy outcomes starting before pregnancy as a large proportion of pregnancies are not planned and many attempts to get pregnant fail. Therefore, the most feasible way of measuring pre-pregnancy IE would be a retrospective survey taken during pregnancy. Also, our sample consisted of predominantly white, married, highly educated women who were between 31 and 40 years of age. About 54% of the participants gained excess weight compared to the recommendations based on the pre-pregnancy BMI. Therefore, to generalize these results to all pregnant women, it is necessary to further study this scale with more diverse samples of pregnant women. Another limitation is that we used self-reported pre pregnancy weight and height data in our analysis. However, studies have shown self-reported weights had high agreement with hospital records (Phelan et al., 2011). This scale could be used in evaluating 1) how eating behaviors change when a woman becomes pregnant, 2) whether prepregnancy eating behavior may be used to identify women at risk for negative pregnancy outcomes, and 3) consequences of intuitive eating. 5. Conclusion The IES-PreP (Intuitive Eating Scale for Pre-pregnancy) is a 15 item measure of three distinct features of adaptive eating behavior that demonstrated adequate factorial and concurrent validity and reliability. The IES-PreP should be used to explore the relationship
between pre-pregnancy adaptive eating behaviors and pregnancy outcomes such as GWG. Further research is needed to validate this measure in more diverse samples of pregnant women. Funding acknowledgement This work was supported by the Grant in aid to start new research from College of Liberal Arts and Social Sciences, University of Houston. References s, A., & Saldan ~ a, C. (2014). Body dissatisfaction and dietary restraint influence Andre binge eating behavior. Nutrition Research, 34(11), 944e950. http://doi.org/10. 1016/j.nutres.2014.09.003. Augustus-Horvath, C. L., & Tylka, T. L. (2011). The acceptance model of intuitive eating: A comparison of women in emerging adulthood, early adulthood, and middle adulthood. Journal of Counseling Psychology, 58(1), 110e125. http://doi. org/10.1037/a0022129. Bayol, S. a, Simbi, B. H., Bertrand, J. a, & Stickland, N. C. (2008). Offspring from mothers fed a “junk food” diet in pregnancy and lactation exhibit exacerbated adiposity that is more pronounced in females. The Journal of Physiology, 586(13), 3219e3230. http://doi.org/10.1113/jphysiol.2008.153817. Bland, J. M., & Altman, D. G. (1997). Statistics Notes: Cronbach ’ s alpha. British Medical Journal, 314(7080), 22e23. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen, & J. S. Long (Eds.), Testing structural equation models (pp. 136e162). Newbury Park, CA: Sage. Byrne, B. M. (1998). Testing the factorial validity of scores from a measuring instrument (second-order CFA model). In structural equation modeling with LISREL, PRELIS, and SIMPLIS: Basic concepts, applications, and programming. Mahway, NJ: Lawrence Erlbaum Associates. Byrne, B. M. (2012). Structural equation Modeling with M plus. Basic concepts, applications, and programming. New York, NY: Taylor & Francis Group. Centers for Disease Control and Prevention. (2015a). Behaviroal risk factor surveilance System, BRFSS questionnaires. http://www.cdc.gov/brfss/questionnaires/ index/ (Accessed 01 August 2016). Centers for Disease Control and Prevention. (2015b). PRAMS, PRAMS questionnaires. https://www.cdc.gov/prams/questionnaire/ (Accessed 01 August 2016). Centers for Disease Control and Prevention. (2015c). Healthy weight. Assessing your weight. http://www.cdc.gov/healthyweight/assessing/ (Accessed 27 July 2016). Chen, F. F., Sousa, K. H., & West, S. G. (2005). Testing measurement invariance of second-order factor models. Structural Equation Modeling, 12(3), 471e492. Cole, R. E., & Horacek, T. (2010). Effectiveness of the “ my body knows when ” intuitive-eating pilot program. American Journal of Health Behavior, 34(3), 286e297. Conway, R., Reddy, S., & Davies, J. (1999). Dietary restraint and weight gain during pregnancy. European Journal of Clinical Nutrition, 53, 849e853. http://doi.org/10. 1038/sj.ejcn.1600864. Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression. Development of the 10-item Edinburgh postnatal depression scale. British Journal of Psychiatry, 150, 782e786. http://doi.org/10.1192/bjp.150.6.782. Deputy, N. P., Sharma, A. J., & Kim, S. Y. (2015). Gestational weight gain d United States, 2012 and 2013. Morbidity and Mortality Weekly Report, 64(43), 1215e1220. Dietrich, A., Federbusch, M., Grellmann, C., Villringer, A., & Horstmann, A. (2014). Body weight status, eating behavior, sensitivity to reward/punishment, and gender: Relationships and interdependencies. Frontiers in Psychology, 5(October), 1e13. http://doi.org/10.3389/fpsyg.2014.01073. Dockendorff, S., Petrie, T., Greenleaf, C., & Martin, S. (2012). Intuitive Eating Scale: An examination among early adolescents. Journal of Counseling Psychology, 59(4), 604e611. http://doi.org/10.1037/a0029962. Grieger, J. a, Grzeskowiak, L. E., & Clifton, V. L. (2014). Preconception dietary patterns in human pregnancies are associated with preterm. The Journal of Nutrition, (8), 1e6. http://doi.org/10.3945/jn.114.190686. Herbert, B. M., Blechert, J., Hautzinger, M., Matthias, E., & Herbert, C. (2013). Intuitive eating is associated with interoceptive sensitivity. Effects on body mass index. Appetite, 70, 22e30. http://doi.org/10.1016/j.appet.2013.06.082. Herman, C. P., & Polivy, J. (1980). Restrained eating. In S.A. (Ed.), Obesity (pp. 208e225). Philadelphia: Saunders. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1e55. http://doi.org/10.1080/ 10705519909540118. Johnson, F., Pratt, M., & Wardle, J. (2012). Dietary restraint and self-regulation in eating behavior. International Journal of Obesity (2005), 36(5), 665e674. http:// doi.org/10.1038/ijo.2011.156. Katzer, L., Bradshaw, A. J., Horwath, C. C., Gray, A. R., O'Brien, S., & Joyce, J. (2008). Evaluation of a “ nondieting stress reduction program for overweight Women: A randomized trial. American Journal Of Health Promotion, 22(4), 264e275. Laraia, B., Epel, E., & Siega-Riz, A. M. (2013). Food insecurity with past experience of
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209 restrained eating is a recipe for increased gestational weight gain. Appetite, 65, 178e184. http://doi.org/10.1016/j.appet.2013.01.018. Margerison Zilko, C. E., Rehkopf, D., & Abrams, B. (2010). Association of maternal gestational weight gain with short- and long-term maternal and child health outcomes. American Journal of Obstetrics and Gynecology, 202(6), 574.e1e574.e8. http://doi.org/10.1016/j.ajog.2009.12.007. Micali, N., Treasure, J., & Simonoff, E. (2007). Eating disorders symptoms in pregnancy: A longitudinal study of women with recent and past eating disorders and obesity. Journal of Psychosomatic Research, 63, 297e303. http://doi.org/10. 1016/j.jpsychores.2007.05.003. Mumford, S., Siega-Riz, A., Herring, A., & Evenson, K. (2008). Dietary restraint and gestational weight gain. Journal of the American Dietetic Association, 108(10), 1646e1653. http://doi.org/10.1016/j.jada.2008.07.016.Dietary. Phelan, S., Hart, C., Phipps, M., Abrams, B., Schaffner, A., Adams, A., et al. (2011). Maternal behaviors during pregnancy impact offspring obesity risk. Experimental Diabetes Research, 2011, 1e9. http://doi.org/10.1155/2011/985139. Pick, M. E., Edwards, M., Moreau, D., & Ryan, E. A. (2005). Assessment of diet quality in pregnant women using the Healthy Eating Index. Journal of the American Dietetic Association, 105(2), 240e246. http://doi.org/10.1016/j.jada.2004.11.028. Schaefer, J. T., & Magnuson, A. B. (2014). A review of interventions that promote eating by internal cues. Journal of the Academy of Nutrition and Dietetics, 114(5), 734e760. http://doi.org/10.1016/j.jand.2013.12.024. Smith, T., & Hawks, S. R. (2006). Intuitive eating, diet composition, and the meaning of food in healthy weight promotion. American Journal of Health Education, 37(October 2014), 130e136. http://doi.org/10.1080/19325037.2006.10598892. Sollid, C. P., Wisborg, K., Hjort, J., & Secher, N. J. (2004). Eating disorder that was diagnosed before pregnancy and pregnancy outcome. American Journal of Obstetrics and Gynecology, 190, 206e210. http://doi.org/10.1016/S0002-9378(03) 00900-1.
209
ska, E., Jolly, K., Glinkowski, S., Duda, W., Borowiack, E., Thangaratinam, S., Rogozin et al. (2012). Interventions to reduce or prevent obesity in pregnant women: A systematic review. Health Technology Assessment, 16(31), 1e191. http://doi.org/ 10.3310/hta16310. Tribole, E., & Resch, E. (2003). Intuitive Eating: A revolutionary program that works. New York, NY: St. Martin’s Press. Tylka, T. L. (2006). Development and psychometric evaluation of a measure of intuitive eating. Journal of Counseling Psychology, 53(2), 226e240. http://doi.org/ 10.1037/0022-0167.53.2.226. Tylka, T. L., & Kroon Van Diest, A. M. (2013). The Intuitive Eating Scale-2: Item refinement and psychometric evaluation with college women and men. Journal of Counseling Psychology, 60(1), 137e153. http://doi.org/10.1037/a0030893. Tylka, T. L., & Wilcox, J. A. (2006). Are intuitive eating and eating disorder symptomatology opposite poles of the same construct? Journal of Counseling Psychology, 53(4), 474e485. http://doi.org/10.1037/0022-0167.53.4.474. United States Department of Agriculture (USDA) Center for Nutrition Policy Promotion Promotion. (1995). The healthy eating index. Van Strien, T., Frijters, J. E. R., Roosen, R. F. M., Knuiman-Hijl, W. J. H., & Defares, P. B. (1985). Eating behavior, personality traits and body mass in women. Addictive Behaviors, 10(4), 333e343. http://doi.org/10.1016/0306-4603(85)90029-2. Van Strien, T., Schippers, G. M., & Cox, W. M. (1995). On the relationship between emotional and external eating behavior. Addictive Behaviors, 20(5), 585e594. http://doi.org/10.1016/0306-4603(95)00018-8. Ward, V. B. (2008). Eating disorders in pregnancy. BMJ: British Medical Journal, 336(7635), 93e96. http://doi.org/10.1136/bmj.39393.689595.BE. Yu, C.-Y. (2002). Evaluating cutoff criteria of model fit indices for latent variable models with binary and continuous outcomes. Doctoral dissertation, University of California, Los Angeles. https://www.statmodel.com/download/Yudissertation.pdf Accessed 07, 27, 2016.
Contents lists available at ScienceDirect
Appetite journal homepage: www.elsevier.com/locate/appet
Validation of the intuitive Eating Scale for pregnant women Sajeevika Saumali Daundasekara a, *, Anitra Danielle Beasley b, Daniel Patrick O'Connor a, McClain Sampson c, Daphne Hernandez a, Tracey Ledoux a a
Department of Health and Human Performance, University of Houston, 3875 Holman Street, Garrison Gym, Room 104, Houston, TX 77204-6015, USA Obstetrics and Gynecology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA c Graduate College of Social Work, The University of Houston, 110HA Social Work Building, Houston, TX 77204-4013, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 August 2016 Received in revised form 30 January 2017 Accepted 2 February 2017 Available online 3 February 2017
Pre-pregnancy maladaptive eating behaviors have predicted inadequate or excess gestational weight gain and poor dietary intake during pregnancy, but little is known about effects of pre-pregnancy adaptive eating behaviors on pregnancy outcomes. The purpose of this study was to produce a valid and reliable measure of adaptive pre-pregnancy eating behaviors for pregnant women using the Intuitive Eating Scale. Data were collected from 266 pregnant women, aged 18 and older who were attending a private prenatal clinic at Texas Children's Hospital Pavilion for Women in Houston, TX using selfadministered questionnaires. Confirmatory factor analysis was performed to validate the factor structure of the Intuitive Easting Scale (IES). Concurrent validity was determined using correlations between the three subscale scores [unconditional permission to eat (UPE), eating for physical not emotional reasons (EPR), and relying on hunger/satiety cues (RIH)], perinatal depression status (Edinburgh Postnatal Depression Scale), and pre-pregnancy body mass index (BMI) calculated from self-reported height and weight. After discarding 6 items, the second order model did not fit the data, however, the first order model with three latent factors had reasonable fit (RMSEA ¼ 0.097, CFI ¼ 0.961, TLI ¼ 0.951 and WRMR ¼ 1.21). The internal consistency of the scale was confirmed by Cronbach's alphas (UPE ¼ 0.781, EPR ¼ 0.878 and RIH ¼ 0.786). All subscale scores were inversely related to perinatal depression status. EPR and RIH subscale scores were inversely related to pre-pregnancy BMI, supporting the measure's validity. Among pregnant women, the revised 15 item pre-pregnancy IES (IES-PreP) should be used to evaluate pre-pregnancy adaptive eating behaviors. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Intuitive eating Pregnant women Intuitive eating scale Validation Confirmatory factor analysis
1. Introduction The Institute of Medicine, Gestational weight gain (GWG) recommendations for pregnancy is based on the pre-pregnancy body mass index (BMI). Women who are normal weight prior to pregnancy are recommended to gain 11.5e16 kg, the recommendation for overweight women is 7e11.5 kg and for obese women it is 5e9 kg. Inadequate and Excessive GWG are used to describe weight gain below and above this recommendation respectively. Excess GWG and poor dietary intake are related to adverse health outcomes including postpartum weight retention, gestational
* Corresponding author. E-mail addresses: [email protected] (S.S. Daundasekara), Anitra.Beasley@ bcm.edu (A.D. Beasley), [email protected] (D.P. O'Connor), mmsampson@ uh.edu (M. Sampson), [email protected] (D. Hernandez), TALedoux@uh. edu (T. Ledoux). http://dx.doi.org/10.1016/j.appet.2017.02.001 0195-6663/© 2017 Elsevier Ltd. All rights reserved.
diabetes, pregnancy induced hypertension, pre term deliveries, fetal growth restriction, fetal macrosomia, large for gestational age infants, neonatal hypoglycemia, and infant and childhood obesity (Dietrich, Federbusch, Grellmann, Villringer, & Horstmann, 2014; Grieger, Grzeskowiak, & Clifton, 2014; Margerison Zilko, Rehkopf & Abrams, 2010; Thangaratinam et al., 2012). According to 2012e2013 United States data, prevalence of inadequate GWG ranged from 12.6% to 25.5% and prevalence of excessive GWG ranged from 38.2% to 54.7%, while only 32.1% of women had adequate weight gain according to Institute of Medicine recommendations (Deputy, Sharma, & Kim, 2015). Additionally, a recent study showed that on average, pregnant women do not adequately conform to federal dietary guidelines (United States Department of Agriculture (USDA) Center for Nutrition Policy Promotion Promotion, 1995) with 40% of pregnant women not meeting the minimum recommended number of servings of most food groups, dietary fiber, calcium, vitamin D, iron and folate (Pick, Edwards,
202
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
Moreau, & Ryan, 2005). It is important to identify modifiable determinants of poor diet quality, excess GWG, and negative pregnancy outcomes that clinicians can assess early in pregnancy or before so that early interventions to target these factors can be developed. In non-pregnant samples, eating behaviors regarding what, when, and how much to eat influence dietary quality and weight. Common eating behaviors linked to obesity and poor diet quality among non-pregnant adults are emotional eating, external eating and restrained eating (Van Strien, Frijters, Roosen, Knuiman-Hijl, & Defares, 1985; Van Strien, Schippers, & Cox, 1995). Eating behaviors leading to negative health consequences may be considered maladaptive (Tylka, 2006). For example, restrained eating is characterized by rigid restrictions in caloric or food intake with occasional episodes of disinhibited eating (Herman & Polivy, 1980) and is positively related to obesity, weight cycling and binge eating s & Saldan ~ a, 2014; Johnson, Pratt, & Wardle, 2012). Pre(Andre pregnancy maladaptive eating behaviors (e.g. restraint, disordered eating) have been related to excess and inadequate GWG and poor dietary intake during pregnancy (Conway, Reddy, & Davies, 1999; Mumford, Siega-Riz, Herring, & Evenson, 2008; Sollid, Wisborg, Hjort, & Secher, 2004). Women with eating disorders such as anorexia nervosa and bulimia nervosa prior to pregnancy were found to be at higher risk of impaired pregnancy outcomes including pre-term deliveries, small for gestational age babies and low birth weight infants (Micali, Treasure, & Simonoff, 2007; Sollid et al., 2004; Ward, 2008). Pre-pregnancy maladaptive eating behaviors have been linked with negative pregnancy outcomes, which implies these maladaptive eating behaviors should be treated before pregnancy. However, there is little information available as to what types of pre-pregnancy eating behaviors should replace pre-pregnancy maladaptive eating behaviors. Intuitive eating (IE) was originally conceptualized by Evelyn Tribole and Elyse Resch, who are both clinical dietitians. The IE theory posits that strict food rules that ignore ‘body wisdom’ (i.e., hunger/satiety cues), lead to preoccupation with food and loss of control eating prompted by emotions and external cues. Tribole and Resch also suggest that honoring ‘body wisdom’ will normalize eating and reduce preoccupation with food and loss of control eating (Tribole & Resch, 2003). Among non-pregnant adults, IE has been related to lower BMI, stable weight, and fewer dieting behaviors and food anxieties (Schaefer & Magnuson, 2014; Smith & Hawks, 2006; Tylka & Wilcox, 2006; Tylka, 2006; Tylka & Kroon Van Diest, 2013). Interventions that promote IE among nonpregnant overweight and obese populations show weight maintenance and improved body image (Cole & Horacek, 2010; Katzer et al., 2008), and cross sectional studies of IE show higher levels of IE have been positively related to psychological well-being and negatively related to BMI making this an adaptive eating behavior among the general adult population (Augustus-Horvath & Tylka, 2011; Tylka & Wilcox, 2006; Tylka, 2006; Tylka & Kroon Van Diest, 2013). However, to date, there are no studies investigating the relationship between pre-pregnancy IE and pregnancy outcomes. To conduct this type of research, a validated measure of prepregnancy IE is required. If pre-pregnancy IE is protective of negative pregnancy outcomes including excess GWG, then perhaps this measure could also be used to screen for pregnant women at risk of pregnancy complications so that interventions can be initiated as early in pregnancy as possible. The Intuitive Eating Scale (IES) is a 21-item scale with three subscales:(1) unconditional permission to eat (UPE, 9 items), (2) eating for physical rather than emotional reasons (EPR, 6 items), and (3) reliance on internal hunger and satiety cues to determine when and how much to eat (RIH, 6 items). This scale was developed and the 3-factor second order factor structure based on Tribole and
Resch's theory and was validated among predominantly female college students (Tylka, 2006). Some studies have shown that the original measure is valid with other populations but requires minor adjustments in scoring given a different factor structure (e.g. adolescents) (Dockendorff, Petrie, Greenleaf, & Martin, 2012). Before the IES can be used to assess pre-pregnancy intuitive eating among pregnant women, the validity of this measure with this population should be confirmed. The goal of this study was to confirm the validity and reliability of the IES to test pre-pregnancy intuitive eating among pregnant women. Following procedures used by Tylka (2006) the factor structure of the scale was analyzed using a confirmatory factor analysis to determine: (1) the overall fit of data to the scale model, (2) the item loadings, and (3) the relationship between the latent factors. We hypothesized that the IES items would load on their respective latent factors as identified by Tribole and Resch (Tribole & Resch, 2003; Tylka, 2006). In addition, the latent factors would be related, load on the higher order IE factor, and the overall model would provide adequate fit to the data in a sample of pregnant women. According to the previous findings we also hypothesized that the total IES scores would demonstrate concurrent validity and be inversely related to pre-pregnancy BMI and perinatal depression status. 2. Methodology 2.1. Procedure This study was conducted at the Texas Children's Hospital Pavilion for Women in Houston, TX, USA. The participants were pregnant women attending a private prenatal clinic at this location. Data for the study were collected from spring 2013 to summer 2014. The participant inclusion criteria for the study were women 18 years of age or older, singleton pregnancy confirmed by a physician, and willingness and ability to complete the survey in English. Women below 18 years or with current multiple pregnancy were excluded from the study. A research coordinator distributed recruitment flyers to women entering the clinic waiting room, and women who were interested in participating in the survey were given a survey packet including: a screening checklist, a consent form, a resource list, and the questionnaire. The study was described to potential participants as an investigation of health behaviors of pregnant women. A self-screening checklist was used in determining participant's eligibility. Written informed consent was obtained from participants while they were waiting for their appointments in the waiting room. Participants were asked to complete the survey in the same waiting room or at the examination room. No financial compensation was given to participants, however a prenatal care resource list containing contact information of community agencies, or places to get help during pregnancy was given to each participant to help find support for specific needs they have regarding the pregnancy. The survey packet was given to 300 interested and eligible participants. Nineteen women returned the questionnaires without completing any of the items and 15 women answered 90% of the given measures, so they were excluded. The final was 266 participants. This study was approved by the Institutional Review Boards of the University of Houston and Baylor College of Medicine. 2.2. Socio-demographic data and body mass index Socio-demographic data including age, race, marital status, education, employment, gestational age and household income were collected through a questionnaire consisting of items taken
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
from two existing national surveys, the Brief Risk Factor Surveillance System (CDC, 2015a) and the Pregnancy Risk Assessment Management System by Center for Disease Control and Prevention (CDC 2015b). Pre-pregnancy BMI was calculated using the selfreported height and weight according to CDC guidelines (CDC, 2015c). 2.3. Depression Perinatal depression symptoms were evaluated using the 10 item Edinburgh Postnatal Depression Scale (EPDS) (Cox, Holden, & Sagovsky, 1987). This is a self-reported scale originally developed to use with postpartum women and later have been validated for pregnant women, to identify risk for perinatal depression (e.g., “I have been able to laugh and see the funny side of things”; “I have blamed myself unnecessarily when things went wrong”). Participants are asked to select the answer, that best represent their situation within the past 7 days. Each item is scored on a four point scale ranging from 0 to 3 (0 ¼ most of the time/very often to 3 ¼ never/not at all). Negatively stated items were reversed scored, so that high scores indicate higher perinatal depression status. A total score higher than 13 indicate depression levels consistent with depressive disorders. In this analysis we used the total EPDS score as a continuous measure of perinatal depression status within the last week. 2.4. Pre pregnancy eating behavior The 21 item IES (Tylka, 2006) was adapted for the current study with pregnant women to measure pre pregnancy eating behaviors. The original IES was developed to assess three constructs thought to make up intuitive eating including unconditional permission to eat (UPE; e.g., “I tried to avoid certain foods high in fat, carbohydrates, or calories”), eating for physical rather than emotional reasons (EPR; e.g., “I usually stopped eating when I felt full [not overstuffed]”), and reliance on hunger and satiety cues to determine when and how much to eat (RIH; e.g., “I could tell when I was slightly full”). In the current study, the IES was modified by adding the term “Before pregnancy” before each item (e.g., “Before pregnancy, I tried to avoid certain foods high in fat, carbohydrates, or calories”). The format of the response is a five point Likert-type scale (1 ¼ strongly disagree and 5 ¼ strongly agree). Some of the scale items are negatively stated and some are positively stated. In calculating the scale scores, the negative items were reversed scored so that high scores on the total measure and subscales indicate greater intuitive eating. The individual scores of items under each subscale were summed to obtain subscale scores. The total IES score is the sum of three subscale scores. The total subscale scores were divided by the number of items in each subscale to get the mean subscale scores. This resulted in possible mean scores ranges from 1 to 5. 2.5. Statistical analysis Statistical analyses were performed using IBM SPSS statistics for Windows, version 22.0 and Mplus, version 7.4 was used for the confirmatory factor analysis (CFA) using a covariance matrix. The factor structure of the modified IES was analyzed with two CFA models to determine the overall fit of the data to the scale model and to assess whether the items load on their hypothesized latent factors. CFA was used as the IES has a well-established factorial structure based on the concepts of IE (Tylka, 2006) and as our objective was to evaluate the validity of this scale for pregnant population. The weighted least squares with mean and variance
203
adjustment (WLSMV) estimation method, which is recommended for ordinal data was used in the CFA (Bayol, Simbi, Bertrand, & Stickland, 2008; Byrne, 2012). Sample size was adequate to conduct a factor analysis (N ¼ 266) which is about 12:1 participantto-item ratio (Byrne, 2012). The 21 items of the modified scale were the indicators of the three first-order latent factors UPE, EPR, and RIH. We used a first-order CFA to determine whether each item loaded onto their respective first-order factor and to determine whether the first-order factors are correlated to each other as hypothesized. After confirming the first-order model, a second-order CFA was conducted to determine whether the first-order latent factors loaded on the higher-order intuitive eating factor. One item loading for each first order factor was set equal to 1.0 to provide a scale for the first-order factors, and the same strategy was used for the second-order loadings to provide a scale for the second-order factor (Chen, Sousa, & West, 2005). Adequacy of fit was determined by the indices; the Comparative Fit Index (CFI), the TuckereLewis Index (TLI), the root-mean-square error of approximation (RMSEA) and weighted root mean square of residual (WRMR) (Hu & Bentler, 1999). According to Hu and Bentler RMSEA values less than 0.05 indicates a close fit, and Browne and Cudeck suggested that values in the range of 0.05e0.08 indicate fair fit; values above 1.0 suggests poor fit (Browne & Cudeck, 1993). Values above 0.9 for CFI and TLI indicates a reasonably good fit (>0.95 close fit), and according to Yu (2002), values closer to 1.0 for WRMR indicates a reasonable fit (<0.95 good fit). The internal consistency of each subscale was evaluated using the Cronbach's alpha. The concurrent validity was examined through correlations with pre-pregnancy BMI and the perinatal depression status. According to previous studies we expect the IES scores to be negatively related to prepregnancy BMI and the depression status (Dockendorff et al., 2012; Tylka, 2006). 3. Results 3.1. Participants The socio-demographics and pre-pregnancy BMI of the women are given in Table 1. The mean age of the sample was 30.69 ± 4.79 years, mean pre-pregnancy BMI was 25.31 ± 6.31. The mean gestational age of the women was 30.71 ± 6.73 weeks but it ranged from 8 to 41 weeks. 3.2. Confirmatory factor analysis Missing data in the modified IES were evaluated using Little's MCAR test before conducting the CFA and was found to be consistent with being missing completely at random. Therefore, the expectation maximization (EM) method was used to replace the missing values as previously conducted in the validation of IES for adolescents (Dockendorff et al., 2012). Some of the items deviated from a normal distribution, which is expected with ordinal level data therefore the WLSMV for non-normal categorical data was used. The fit statistics of the first order model indicated that the original model (Fig. 1) had a poor fit to the data (RMSEA ¼ 0.135, CFI ¼ 0.848, TLI ¼ 0.828 and WRMR ¼ 2.13). Therefore, the factor loading for the first-order model was evaluated to identify items that did not load strongly onto their hypothesized latent factor. Using methodology from the original IES validation study (Tylka, 2006), the items with a factor loading below 0.45 on their hypothesized latent factor were removed. Items 1, 4, 5, and 19 were discarded from the UPE subscale due to poor factor loadings (0.126, 0.144, 0.433 and 0.120, respectively). The resulting model did not have a significant improvement and further evaluation of item
204
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
Table 1 Sample characteristics. Characteristics Race/Ethnicity (N ¼ 266)
Marital status (N ¼ 266) Education (N ¼ 262)
Employment (N ¼ 264) Household income (N ¼ 249) Pre-pregnancy BMI (N ¼ 261)
Non-Hispanic White Non-Hispanic Black Hispanic Asian American Indian or Alaska Native Others Married Others Some high school High School Graduate Some College College Graduate Employed Unemployed < $50,000 $50,000 or More Underweight (<18.5) Normal (18.5e24.9) Overweight (25e29.9) Obese (30)
N
Percentage
140 24 66 29 1 6 238 28 2 18 47 195 188 76 42 207 10 145 68 38
52.6 9.0 24.8 10.9 0.4 2.3 89.5 10.5 0.8 6.9 17.9 74.4 71.2 28.8 16.9 83.1 3.8 55.6 26.0 14.6
Abbrevations: BMI ¼ Body Mass Index, N¼Number of participants.
loading and modification indices indicated that item 14 had a poor factor loading (0.398), and item 2 loaded strongly on the hypothesized latent factor (EPR) as well as on RIH. Therefore, these two items were also removed from the scale. These adjustments significantly improved the CFI, TLI and WRMR (0.927, 0.912 and 1.66) and resulted in a minor improvement in RMSEA (0.131) even though fit statistics were not adequate. Even after discarding the six items, the model did not fit the data adequately. In the original model we hypothesized that the error correlations among all the item pairs are zero. The modification indices revealed that there are strong error correlations among several items (7 and 8, 13 and 12, 18 and 9). This model with 15 items, 3 correlated latent factors and 3 correlated errors, had 120 data points and 36 parameters to be estimated. Therefore, model is overidentified model with 84 degrees of freedom. Addition of these error correlations to the model significantly improved the model fit indicating that this new first order model fit the data well (RMSEA ¼ 0.097, CFI ¼ 0.961, TLI ¼ 0.951 and WRMR ¼ 1.21). The three first order factors showed significant moderate relationship to one another (estimates 0.3e0.7, P < 0.001). This model proved our hypothesis 1. That overall model fit to data adequately, 2. The remained items significantly and adequately loaded onto their hypothesized latent factors and 3. There are significant relationships between the three latent factors. Next we conducted a second order factor analysis by adding the higher order latent factor, “Intuitive Eating” to the final model containing 15 items and 3 error correlations. The residual variance of the latent factor EPR in this second order model was negative (0.033, p ¼ 0.769) indicating a model misspecification suggesting there is no support for a second order IE factor among pregnant women. Therefore, we had to reject our final hypotheses that the three latent factors load on the higher order IE factor. The results are reported for the final first order model that showed adequate fit to the data. The standardized factor loadings of the final first-order model with error correlations are given in Fig. 2. In the final validated 15 item pre-pregnancy IES which will be referred to as IES-PreP in later text, the UPE subscale consists of four items (mean ¼ 3.36, SD ¼ 1.01). The EPR subscale has five items (mean ¼ 3.80, SD ¼ 0.94), and the RIH subscale has six items (mean ¼ 3.84, SD ¼ 0.62). The Cronbach's alphas of three subscales were 0.781, 0.878 and 0.786, respectively, indicating a satisfactory internal consistency among the items in each subscale (Bland &
Altman, 1997). The items under each subscale of the IES-PreP are given in Table 2. 3.3. Concurrent validity The UPE factor was significantly negatively related to the perinatal depression status among pregnant women, but there was no significant correlation between this factor and the pre-pregnancy BMI. Both the EPR and RIH factors showed significant moderate inverse correlation with both the perinatal depression status and pre-pregnancy BMI (Table 3). 4. Discussion This study was designed to test the validity of the IES (Tylka, 2006) to measure pre-pregnancy intuitive eating behaviors among pregnant women. Intuitive eating is considered an adaptive eating behavior made up of three concepts: unconditional permission to eat, eating for physical rather than emotional reasons, and reliance on hunger/satiety to inform food choices (Tribole & Resch, 2003). Among pregnant women a first order model of prepregnancy intuitive eating behaviors provided an adequate fit to the data while the second order factor model did not fit the data, which is incongruent with the intuitive eating theory and results from IES validity research on college samples (Tylka, 2006). According to Tylka, second order model provided an adequate fit to data from college women (CFI ¼ 0.91, TLI ¼ 0.90, RMSEA ¼ 0.08 and SRMR ¼ 0.07). In other words, among college students, the total IES score represents an overall intuitive eating construct made up of three subscales. However, in this sample of 266 pregnant women, the IES total score has no meaning, and the three subscales represent related but distinct eating behaviors. While these results were not hypothesized based on theory or prior research with college students, the findings reflect mounting evidence that the IES does not represent an overall intuitive eating construct for all individuals. A CFA on IES data from adolescents confirmed a first order factor structure (Dockendorff et al., 2012). The clinical and research implication for this is that the modified IES for pregnant women should be scored as three separate but related subscales not as a total overall score. The theoretical implication of this is that intuitive eating as an overall construct representing unconditional permission to eat, eating for physical
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
205
Fig. 1. The initial first order model of the Intuitive Eating Scale for pregnant women (the standardized coefficients for the confirmatory factor analysis e all the loadings were statistically significant p < 0.05 except for ies 19).
206
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
Fig. 2. The final three factor first order model of the Intuitive Eating Scale for pregnant women (the standardized coefficients for the confirmatory factor analysis e all the loadings were statistically significant p < 0.001).
rather than emotional reasons, and relying on hunger/satiety to inform eating behaviors may not exist among certain segments of the population. Six items had low factor loadings on the hypothesized latent factors, and therefore, were dropped from the respective subscales. In the UPE subscale the items that did not have adequate loadings seemed to be related to having strict food rules (e.g.,
“Before pregnancy, I tried to avoid certain foods high in fat, carbohydrates or calories” and “Before pregnancy, there were forbidden foods that I didn't allow myself to eat”). The items that remained in this scale were related to negative feelings about and lack of control over eating unhealthy foods (e.g., “Before pregnancy, I would get mad at myself for eating something unhealthy” and “Before pregnancy, I felt guilty if I ate a certain food that was
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
207
Table 2 The items in the IES-PreP. Unconditional permission to eat-UPE (4 items)
Eating for physical rather than emotional reasons-EPR (5 items)
Reliance on hunger and satiety cues to determine when and how much to eat -(6 items)
IES 9a- Before pregnancy, I would get mad at myself for eating something unhealthy IES 18- Before pregnancy, I felt guilty if I ate a certain food that was high in calories, fat or carbohydrates IES 20- Before pregnancy, I didn't trust myself around fattening foods IES 21- Before pregnancy, I didn't keep certain foods at home because I thought that I would lose control and eat them
IES 3- Before pregnancy, I found myself eating when I felt emotional, even when I wasn't physically hungry IES 6- Before pregnancy, I found myself eating when I was bored, even when I was not physically hungry
IES 7- Before pregnancy, I could tell when I was slightly full IES 8- Before pregnancy, I could tell when I was slightly hungry
IES 10- Before pregnancy, I found myself eating when I was lonely, even when I was not physically hungry IES 16- Before pregnancy, I used food to help me soothe my negative emotions
IES 11- Before pregnancy, I trusted my body to tell me when to eat IES 12-Before pregnancy, I trusted my body to tell me what to eat
IES 17- Before pregnancy, I found myself eating when I was stressed out, even when I was not physically hungry
IES 13- Before pregnancy, I trusted my body to tell me how much to eat IES 15- Before pregnancy, when I ate, I could tell when I was getting full
Abbreviation; IES, Intuitive eating scale. a The IES numbers is given according to the item number in the original 21 item scale, which is used to refer to these items in the text.
Table 3 Correlations among the measures of the study (N ¼ 266). Measures 1. 2. 3. 4. 5.
a
Unconditional permission to eat Eating for physical reasonsb Reliance on internal hunger/satiety cuesc Pre-pregnancy BMId Perinatal Depressione
Mean
SD
1
2
3
4
5
3.37 3.80 3.85 25.31 6.30
1.01 0.94 0.63 6.31 4.33
0.546** 0.197** 0.090 0.217**
e 0.323** 0.240** 0.320**
e 0.272** 0.220**
e 0.165**
e
Abbrevations: SD, Standard Deviation and BMI, Body Mass Index. Note. ** Correlation is significant at the 0.01 level (2 tailed). a,b,c based on the 15 item IES-PreP, possible scores range 1e5. a 4 item subscale. b 5 item subscale. c 6 item subscale. d calculated using self-reported pre-pregnancy height and weight (BMI ¼ weight (kg)/height2(m)). e perinatal depression status within the past week is measured by 10 item Edinburgh Postnatal Depression Scale, possible scores range 0e30.
high in calories, fat or carbohydrates”). It is possible the participants in this study did not have strict food rules prior to pregnancy, but did feel guilty about eating what they perceived to be unhealthy foods. When using the IES for pre-pregnancy, it may be best to call this subscale “lack of negative feelings about eating unhealthy foods” as the remaining items do not reflect unconditional eating. In order to capture “unconditional eating” among this population, the discarded items from the UPE subscale might need rephrasing and further studying. One item (i.e., “Before pregnancy, I usually stopped eating when I felt full, not overstuffed”) loaded on the EPR factor in the original validation studies but in our analysis it had a similar factor loading on the RIH factor. As explained by Dockendorff and others this item is different from the other items in the EPR factor, which are about eating for reasons other than physical hunger (e.g., “Before pregnancy, I found myself eating when I felt emotional, even when I wasn't physically hungry” and “Before pregnancy, I used food to help me soothe my negative emotions”) and seems more consistence with the items in the RIH factor which are about understanding internal hunger and satiety cues (e.g. “Before pregnancy, I could tell when I was slightly full” and “Before pregnancy, I trusted my body to tell me how much to eat”). Therefore, we decided to remove this item from the scale to avoid overlapping of subscale scores. It was hypothesized that the error terms would not be correlated. However, there was significant correlation among the error terms between items 7 and 8, items 9 and 18, and items 12 and 13. This indicates that each pair of item shares common variance that
is unrelated to the respective constructs represented by those items such as item wording or format. For example, items 7 and 8 both share the following phrase: “I could tell when I was___” (see Table 2). Some of the consistency that respondents had when answering those two items may have been related to that common phrase, rather than the intended construct. The final model fit improved significantly by inclusion of correlated errors because they accounted for additional systematic variation (Byrne, 1998, pp. 163e191), but in our study allowing some errors to correlate did not materially affect the model's individual parameter estimates. We hypothesized that pre-pregnancy BMI and perinatal depression status would be negatively related to the three subscale scores and the subscale scores would be correlated to each other. IE has been considered an adaptive eating behavior partially because it is inversely related to BMI in non-pregnant samples; therefore, it was expected to have the same relationship with pre-pregnancy BMI in this study. IE in non-pregnant populations has been related to better psychological status (Augustus-Horvath & Tylka, 2011; Tylka & Wilcox, 2006; Tylka, 2006; Tylka & Kroon Van Diest, 2013), therefore, pre-pregnancy IE was expected to be inversely related to prenatal depressive symptoms. Our results revealed that UPE factor was not significantly correlated with pre-pregnancy BMI which was not consistent with Tylka's findings (Tylka, 2006). However, the remaining valid items on this subscale sufficiently changed the construct represented from UPE to negative feelings associated with eating unhealthy foods. Therefore, it seems negative feelings associated with eating
208
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209
unhealthy foods is not related to BMI. A recent IE study done with female students using the 21 item IES has also reported that UPE subscale scores were not significantly correlated with BMI (Herbert, Blechert, Hautzinger, Matthias, & Herbert, 2013). Women who eat for physical hunger without emotional reasons and women who relayed more on hunger and satiety cues to control their eating had lower BMIs, which was consistent with Tylka's results (Tylka, 2006) suggesting concurrent validity. All three subscale scores showed moderate inverse correlation with perinatal depression scores indicating that women who showed higher pre-pregnancy IE related eating behaviors had lower perinatal depression symptomology. This result is consistent with the eating behavior theory and previous studies where adaptive eating behaviors have been shown to predict better psychological status (Dockendorff et al., 2012; Tylka, 2006). According to Tylka among college women, UPE was moderately related to selfesteem and satisfaction with life, EPR was strongly related to optimism and moderately related to self-esteem and satisfaction with life, and RIH was moderately related to self-esteem, satisfaction with life and optimism. Among adolescents all three subscale scores were inversely related to feeling worthless, and/or sad/ depressed (Dockendorff et al., 2012). Also, the subscale scores were positively correlated with each other as expected. The main limitation of this study is recall bias. Gestation age of participants at the time of the study, ranged from 8 to 41 weeks. Thus, women had to recall their eating behaviors 8e41 weeks prior. In the current measure, no instructions were given to participants to recall their eating behaviors within a certain time period. Researchers, who conducted the few studies of pre-pregnancy eating behaviors (Conway et al., 1999; Laraia, Epel, & Siega-Riz, 2013; Mumford et al., 2008), modified existing eating behavior questionnaires by simply adding “Before pregnancy …” and changing the verb tense in each item to past tense. We followed this model for this study. To overcome recall bias, future researchers would need to test a general sample of adult females with the IES and then follow up with those who become pregnant to determine the relationship between pre-pregnancy intuitive eating behaviors and pregnancy outcomes. But there are many limitations to conducting a prospective study of pregnancy outcomes starting before pregnancy as a large proportion of pregnancies are not planned and many attempts to get pregnant fail. Therefore, the most feasible way of measuring pre-pregnancy IE would be a retrospective survey taken during pregnancy. Also, our sample consisted of predominantly white, married, highly educated women who were between 31 and 40 years of age. About 54% of the participants gained excess weight compared to the recommendations based on the pre-pregnancy BMI. Therefore, to generalize these results to all pregnant women, it is necessary to further study this scale with more diverse samples of pregnant women. Another limitation is that we used self-reported pre pregnancy weight and height data in our analysis. However, studies have shown self-reported weights had high agreement with hospital records (Phelan et al., 2011). This scale could be used in evaluating 1) how eating behaviors change when a woman becomes pregnant, 2) whether prepregnancy eating behavior may be used to identify women at risk for negative pregnancy outcomes, and 3) consequences of intuitive eating. 5. Conclusion The IES-PreP (Intuitive Eating Scale for Pre-pregnancy) is a 15 item measure of three distinct features of adaptive eating behavior that demonstrated adequate factorial and concurrent validity and reliability. The IES-PreP should be used to explore the relationship
between pre-pregnancy adaptive eating behaviors and pregnancy outcomes such as GWG. Further research is needed to validate this measure in more diverse samples of pregnant women. Funding acknowledgement This work was supported by the Grant in aid to start new research from College of Liberal Arts and Social Sciences, University of Houston. References s, A., & Saldan ~ a, C. (2014). Body dissatisfaction and dietary restraint influence Andre binge eating behavior. Nutrition Research, 34(11), 944e950. http://doi.org/10. 1016/j.nutres.2014.09.003. Augustus-Horvath, C. L., & Tylka, T. L. (2011). The acceptance model of intuitive eating: A comparison of women in emerging adulthood, early adulthood, and middle adulthood. Journal of Counseling Psychology, 58(1), 110e125. http://doi. org/10.1037/a0022129. Bayol, S. a, Simbi, B. H., Bertrand, J. a, & Stickland, N. C. (2008). Offspring from mothers fed a “junk food” diet in pregnancy and lactation exhibit exacerbated adiposity that is more pronounced in females. The Journal of Physiology, 586(13), 3219e3230. http://doi.org/10.1113/jphysiol.2008.153817. Bland, J. M., & Altman, D. G. (1997). Statistics Notes: Cronbach ’ s alpha. British Medical Journal, 314(7080), 22e23. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen, & J. S. Long (Eds.), Testing structural equation models (pp. 136e162). Newbury Park, CA: Sage. Byrne, B. M. (1998). Testing the factorial validity of scores from a measuring instrument (second-order CFA model). In structural equation modeling with LISREL, PRELIS, and SIMPLIS: Basic concepts, applications, and programming. Mahway, NJ: Lawrence Erlbaum Associates. Byrne, B. M. (2012). Structural equation Modeling with M plus. Basic concepts, applications, and programming. New York, NY: Taylor & Francis Group. Centers for Disease Control and Prevention. (2015a). Behaviroal risk factor surveilance System, BRFSS questionnaires. http://www.cdc.gov/brfss/questionnaires/ index/ (Accessed 01 August 2016). Centers for Disease Control and Prevention. (2015b). PRAMS, PRAMS questionnaires. https://www.cdc.gov/prams/questionnaire/ (Accessed 01 August 2016). Centers for Disease Control and Prevention. (2015c). Healthy weight. Assessing your weight. http://www.cdc.gov/healthyweight/assessing/ (Accessed 27 July 2016). Chen, F. F., Sousa, K. H., & West, S. G. (2005). Testing measurement invariance of second-order factor models. Structural Equation Modeling, 12(3), 471e492. Cole, R. E., & Horacek, T. (2010). Effectiveness of the “ my body knows when ” intuitive-eating pilot program. American Journal of Health Behavior, 34(3), 286e297. Conway, R., Reddy, S., & Davies, J. (1999). Dietary restraint and weight gain during pregnancy. European Journal of Clinical Nutrition, 53, 849e853. http://doi.org/10. 1038/sj.ejcn.1600864. Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression. Development of the 10-item Edinburgh postnatal depression scale. British Journal of Psychiatry, 150, 782e786. http://doi.org/10.1192/bjp.150.6.782. Deputy, N. P., Sharma, A. J., & Kim, S. Y. (2015). Gestational weight gain d United States, 2012 and 2013. Morbidity and Mortality Weekly Report, 64(43), 1215e1220. Dietrich, A., Federbusch, M., Grellmann, C., Villringer, A., & Horstmann, A. (2014). Body weight status, eating behavior, sensitivity to reward/punishment, and gender: Relationships and interdependencies. Frontiers in Psychology, 5(October), 1e13. http://doi.org/10.3389/fpsyg.2014.01073. Dockendorff, S., Petrie, T., Greenleaf, C., & Martin, S. (2012). Intuitive Eating Scale: An examination among early adolescents. Journal of Counseling Psychology, 59(4), 604e611. http://doi.org/10.1037/a0029962. Grieger, J. a, Grzeskowiak, L. E., & Clifton, V. L. (2014). Preconception dietary patterns in human pregnancies are associated with preterm. The Journal of Nutrition, (8), 1e6. http://doi.org/10.3945/jn.114.190686. Herbert, B. M., Blechert, J., Hautzinger, M., Matthias, E., & Herbert, C. (2013). Intuitive eating is associated with interoceptive sensitivity. Effects on body mass index. Appetite, 70, 22e30. http://doi.org/10.1016/j.appet.2013.06.082. Herman, C. P., & Polivy, J. (1980). Restrained eating. In S.A. (Ed.), Obesity (pp. 208e225). Philadelphia: Saunders. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1e55. http://doi.org/10.1080/ 10705519909540118. Johnson, F., Pratt, M., & Wardle, J. (2012). Dietary restraint and self-regulation in eating behavior. International Journal of Obesity (2005), 36(5), 665e674. http:// doi.org/10.1038/ijo.2011.156. Katzer, L., Bradshaw, A. J., Horwath, C. C., Gray, A. R., O'Brien, S., & Joyce, J. (2008). Evaluation of a “ nondieting stress reduction program for overweight Women: A randomized trial. American Journal Of Health Promotion, 22(4), 264e275. Laraia, B., Epel, E., & Siega-Riz, A. M. (2013). Food insecurity with past experience of
S.S. Daundasekara et al. / Appetite 112 (2017) 201e209 restrained eating is a recipe for increased gestational weight gain. Appetite, 65, 178e184. http://doi.org/10.1016/j.appet.2013.01.018. Margerison Zilko, C. E., Rehkopf, D., & Abrams, B. (2010). Association of maternal gestational weight gain with short- and long-term maternal and child health outcomes. American Journal of Obstetrics and Gynecology, 202(6), 574.e1e574.e8. http://doi.org/10.1016/j.ajog.2009.12.007. Micali, N., Treasure, J., & Simonoff, E. (2007). Eating disorders symptoms in pregnancy: A longitudinal study of women with recent and past eating disorders and obesity. Journal of Psychosomatic Research, 63, 297e303. http://doi.org/10. 1016/j.jpsychores.2007.05.003. Mumford, S., Siega-Riz, A., Herring, A., & Evenson, K. (2008). Dietary restraint and gestational weight gain. Journal of the American Dietetic Association, 108(10), 1646e1653. http://doi.org/10.1016/j.jada.2008.07.016.Dietary. Phelan, S., Hart, C., Phipps, M., Abrams, B., Schaffner, A., Adams, A., et al. (2011). Maternal behaviors during pregnancy impact offspring obesity risk. Experimental Diabetes Research, 2011, 1e9. http://doi.org/10.1155/2011/985139. Pick, M. E., Edwards, M., Moreau, D., & Ryan, E. A. (2005). Assessment of diet quality in pregnant women using the Healthy Eating Index. Journal of the American Dietetic Association, 105(2), 240e246. http://doi.org/10.1016/j.jada.2004.11.028. Schaefer, J. T., & Magnuson, A. B. (2014). A review of interventions that promote eating by internal cues. Journal of the Academy of Nutrition and Dietetics, 114(5), 734e760. http://doi.org/10.1016/j.jand.2013.12.024. Smith, T., & Hawks, S. R. (2006). Intuitive eating, diet composition, and the meaning of food in healthy weight promotion. American Journal of Health Education, 37(October 2014), 130e136. http://doi.org/10.1080/19325037.2006.10598892. Sollid, C. P., Wisborg, K., Hjort, J., & Secher, N. J. (2004). Eating disorder that was diagnosed before pregnancy and pregnancy outcome. American Journal of Obstetrics and Gynecology, 190, 206e210. http://doi.org/10.1016/S0002-9378(03) 00900-1.
209
ska, E., Jolly, K., Glinkowski, S., Duda, W., Borowiack, E., Thangaratinam, S., Rogozin et al. (2012). Interventions to reduce or prevent obesity in pregnant women: A systematic review. Health Technology Assessment, 16(31), 1e191. http://doi.org/ 10.3310/hta16310. Tribole, E., & Resch, E. (2003). Intuitive Eating: A revolutionary program that works. New York, NY: St. Martin’s Press. Tylka, T. L. (2006). Development and psychometric evaluation of a measure of intuitive eating. Journal of Counseling Psychology, 53(2), 226e240. http://doi.org/ 10.1037/0022-0167.53.2.226. Tylka, T. L., & Kroon Van Diest, A. M. (2013). The Intuitive Eating Scale-2: Item refinement and psychometric evaluation with college women and men. Journal of Counseling Psychology, 60(1), 137e153. http://doi.org/10.1037/a0030893. Tylka, T. L., & Wilcox, J. A. (2006). Are intuitive eating and eating disorder symptomatology opposite poles of the same construct? Journal of Counseling Psychology, 53(4), 474e485. http://doi.org/10.1037/0022-0167.53.4.474. United States Department of Agriculture (USDA) Center for Nutrition Policy Promotion Promotion. (1995). The healthy eating index. Van Strien, T., Frijters, J. E. R., Roosen, R. F. M., Knuiman-Hijl, W. J. H., & Defares, P. B. (1985). Eating behavior, personality traits and body mass in women. Addictive Behaviors, 10(4), 333e343. http://doi.org/10.1016/0306-4603(85)90029-2. Van Strien, T., Schippers, G. M., & Cox, W. M. (1995). On the relationship between emotional and external eating behavior. Addictive Behaviors, 20(5), 585e594. http://doi.org/10.1016/0306-4603(95)00018-8. Ward, V. B. (2008). Eating disorders in pregnancy. BMJ: British Medical Journal, 336(7635), 93e96. http://doi.org/10.1136/bmj.39393.689595.BE. Yu, C.-Y. (2002). Evaluating cutoff criteria of model fit indices for latent variable models with binary and continuous outcomes. Doctoral dissertation, University of California, Los Angeles. https://www.statmodel.com/download/Yudissertation.pdf Accessed 07, 27, 2016.