Correlates of meal skipping

A systematic review of the correlates of meal skipping in young adults aged 18-30 years was conducted (by FJP) and has been accepted for publication (157). The full manuscript is provided in Appendix 2A, with an overview of the systematic review presented below.

Following the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) protocol (158), EBSCO host, MEDLINE Complete, Global Health,

17

Scopus, EMBASE, Web of Science and Informit platforms were searched (January 2016) for eligible articles. To be included in the review each article was required to meet the following criteria: 1) original peer-reviewed research published in English, 2) mean age of participants between 18-30 years, or aged 18-30 years at baseline for longitudinal studies, for studies that did not report a mean age, the participants needed to be referred to as university or college students; 3) participants free from disease and community-dwelling, 4) measure of meal skipping, meal omission or meal frequency reported (assessed as any meal skipped throughout the day or according to meal type e.g. breakfast, dinner, and supper, 5) there was at least one correlate of meal skipping reported, 6) study design was one of the following: randomised controlled trail, prospective cohort study, case-control study, nested case-control study, cross-sectional study or longitudinal study.

Three-hundred and thirty-one articles were identified, 141 full-text articles were assessed for eligibility, resulting in 35 included studies. Multiple

methodological and reporting weaknesses were apparent in the reviewed studies with 28 of the 35 studies scoring a negative rating in the risk of bias assessment using the Academy of Nutrition and Dietetics Quality Criteria Checklist (159). Meal skipping (any meal) was reported in 12 studies with prevalence ranging between 5-83%. The remaining 25 studies identified specific meals and their skipping rates, with breakfast the most frequently skipped meal (14-88%) compared to lunch (8-57%) and dinner (4-(8-57%).

The majority of studies (28 of the 35 studies) examined correlates by investigating associations between factors and meal skipping behaviours through Chi-square, One-way analysis of variance (ANOVA), Duncan’s’ multiple range test and regression (linear and logistic) statistical analysis (94, 98, 99, 160-185).

Another approach used to examine correlates of meal skipping (used in 10 studies), was the use of a ranking methodology where participants were asked to rank potential correlates against other meal skipping correlates (100, 163, 166, 168, 175, 186-190). From these ten studies, ten ranked correlates were reported.

18

Correlates reported to be associated with meal skipping were grouped according to the domains of the social-ecological framework outlined previously (63). Associations were reported from the individual, social-environmental and physical-environmental domains, with no associations reported from the

macrosystem domain. Full details of included studies and results of the risk of bias assessment are detailed in the published paper and can be found in Appendix 2A.

Individual influences

Of the 35 studies included in this review, 33 studies assessed correlates from the social-ecological framework that could be considered individual

correlates. These included sex, age, ethnicity, body mass index (BMI), education, menstrual regularity, physical activity, internet use, and a list of perceived cognitive influences.

Sex

Sex was reported as a correlate of meal skipping by 12 studies; three reported total meal skipping (160, 171, 180) and nine reported on specific meal skipping (94, 98, 165, 166, 169-171, 177, 184). Two studies identified no

difference in meal skipping (any meal) in relation to sex (160, 171), while one study reported meal skipping (any meal) to be more likely in males (180). Two studies reported no significant difference in breakfast skipping between sexes (94, 169), three reported breakfast skipping to be more likely in males (98, 177, 184), while two reported breakfast skipping to be more likely in females (166, 171).

However, although Huang et al. (171), showed that females were more likely to skip breakfast in summer months, this associations was not present in winter months. Two studies reported lunch skipping (98, 165), and two dinner skipping (165, 170), both studies found females to be more likely to skip these meals (lunch (98, 165) and dinner (165, 170)) compared to males.

Age

Two studies reported an association between age and breakfast skipping (166, 175). Danquah et al. (166), reported breakfast skipping to be more likely in those aged 15-20 years when compared to those aged 21-30 years. Similarly, Lee and Yoon (175) reported meal skipping (any meal) to be more likely in those aged 18-20 years compared to those aged 24-28 years.

19 Ethnicity

Ethnicity was associated with breakfast skipping in five studies (94, 166, 170, 178, 183). Of the studies that included Caucasian participants (94, 166, 170, 183), three found breakfast skipping to be more likely in those who were

Caucasian compared with other ethnicities (Japanese, Korean, African American) (166, 170, 183), and one found no association (94). Another study found breakfast skipping and meal skipping (any meal) to be more likely in Korean young adults compared with Japanese young adults (178). While, lunch and dinner

consumption was found to be more common in Japanese young adults compared to Caucasian and Korean young adults (183).

Body Mass Index (BMI)

Five studies reported that BMI was associated with meal skipping (161, 164, 165, 176, 180). Meal skipping (any meal) was reported in four studies, two found no association between BMI and meal skipping (any meal) (164, 176), one reported meal skipping (any meal) to be more likely in those with an increased BMI (161), while another found meal skipping (any meal) less likely in those with an increased BMI (180). Breakfast skipping was reported by one study and was more likely with those with higher BMI (165).

Education

Three studies examined education and its association with meal skipping behaviours. Eldisoky (168), reported maternal education status and its

relationship with breakfast skipping (168), although this was not significant.

Kapinos & Yakusheva (172) reported that individuals in their second year of university were more likely to report meal skipping (any meal) compared to those in their first year university. In addition, Danquah et al. (166), reported that students in science courses were more likely to report breakfast skipping compared to those enrolled in humanities courses (166).

Menstrual regularity

Eittah (167), found breakfast skipping to be more likely in young adults with an irregular menstrual cycle compared to those with a regular menstrual cycle.

20 Physical activity

Colić Barić et al. (165), observed that breakfast consumption (6 or 7 times per week) was more likely in those who spent ≥ 3.5 hours exercising per week when compared to those who did 2.6 hours per week.

Internet Use

One article reported meal skipping (any meal) to be more common in individuals who used the internet heavily (> 4 hours per day) (173).

Psychological wellbeing

Three studies documented associations between psychological factors and meal skipping (any meal). Yilmaz et al. (185), identified that meal skipping (any meal) was more likely in young adults with depressive symptoms; Suliga et al.

(181), found meal skipping (any meal) to be more likely in those with a self-perception of being overweight; Bahl et al. (162), found meal skipping (any meal) to be less likely in those who were mindful, and meal skipping (any meal) to be less likely in those who had increased body satisfaction.

Fatigue

Two studies examined the association between fatigue and meal skipping (179, 182). Tanaka et al. (182), found breakfast skipping to be more likely in those experiencing fatigue, while Sato-Mito et al. (179), found meal skipping (any meal) to be more likely in those who’s mid-point in sleep was later (falling asleep after 1.30 AM and the mid-point of sleep falling at 5.31 ± 0.55 AM) (179).

Perceived cognitive influences

The following correlates of meal skipping are perceived cognitive influences, which were rated within their respective studies as subjective attributes.

Time: Time or the lack of time was mentioned in 10 studies and when considered against other correlates, time was ranked as the strongest perceived correlate of meal skipping (any meal) in nine of the 10 studies (100, 163, 168, 175, 186-190).

Hunger: A lack of hunger was reported in four studies and ranged in importance from being the strongest correlate of meal skipping to the third strongest perceived correlate (166, 168, 175, 188).

21

Weight control: Weight control was mentioned in four studies and ranged from being the strongest perceived correlate of dinner consumption to the third strongest perceived correlate of meal skipping (any meal) (100, 166, 168, 187).

Money: Money or the lack of money, was reported in three studies and was ranked as either third or fourth strongest perceived correlate of meal skipping (any meal) (100, 187, 188).

Habit: Dietary habit was reported in two studies, one study ranked habit as the second strongest perceived correlate of breakfast skipping and the other ranked it as the fourth strongest perceived correlate of meal skipping (any meal) (175, 186).

Religion: Fasting/religion was reported in two studies, where it was ranked as being either the third or fourth strongest perceived correlate of meal skipping (any meal) (100, 187).

Taste: Taste was reported as being a correlate of meal skipping (any meal), with one study ranking it as its second strongest perceived correlate (186).

Cooking skills: Lack of cooking skills was reported by one study as the third strongest perceived correlate of meal skipping (any meal) (188).

Social-environmental influences

Of the 35 studies included in this review, one study assessed a correlate that could be considered a social-environmental correlate (186). The variable examined was the notion of being sociable. It is unclear exactly what is meant by this variable, however it is predicted that one would rather socialise then eat a meal.

Desire to be sociable

One study reported a desire to be sociable as a correlate of meal skipping (any meal), ranking it third strongest perceived correlate of meal skipping (any meal) (186).

22

Physical-environment influences

Of the 35 studies included in this review, six studies assessed correlates that could be considered physical-environmental correlates. These included socio-economic status based on rural/urban living environments and housing type.

Rural/urban living environment

In one Polish study, meal skipping (any meal) was more likely in those who resided in a rural area compared to those who lived in an urban area (180).

Housing type

Five studies focused on specific living environments such as housing types.

Kapinos & Yakusheva (172) reported meal skipping (any meal) to be more likely in those living in university/college dormitories. Similarly, Beerman et al. (163), reported meal skipping (any meal) to be more likely in those residing with parents or in university dormitories when compared to those living in Greek university housing (fraternity or sorority housing). Individual meal skipping events were reported in three studies (99, 170, 174). Two studies reported breakfast skipping to be more likely in those who lived away from home (99, 170), while one found breakfast skipping to be more likely in those living by one’s self or with parents compared with living on campus (174). This same article reported the same association for dinner skipping (174).

Overview

In conclusion, meal skipping is influenced by a range of correlates. The majority of these correlates are found within the individual domain of the social-ecological framework (n=10), with less in both the social-environmental domain (n=1) and physical-environment (n=2) domains.