Overweight and obesity in all stages of life have become a public health problem worldwide. In 2016, more than 40 million children aged 0–5 years suffered from overweight or obesity worldwide; if this tendency continues, the number could increase to 70 million by 2025(1). UNICEF ranks Mexico first on the incidence of childhood obesity(2). In Mexico, the prevalence increased from 7·8 to 9·7 % between 1988 and 2012. In the Northern region the prevalence has reached 13 %(Reference Gutiérrez, Rivera-Dommarco and Shamah-Levy3,4) . The origin of childhood obesity is multifactorial and the role of parents in pre-school feeding is critical for promoting positive habits(Reference Savage, Fisher and Birch5). Feeding practices are specific behaviours that characterize mealtime interactions and parental control is a key factor for limiting or encouraging eating. Parents can decide to use direct control methods, for example pressuring to eat, restricting, punishing and rewarding; or indirect control methods, for example promoting healthy home environments. What parents tend to choose is meaningful because feeding control produces either a positive or negative outcome. For example, restrictive feeding practices are likely to be associated with overeating(Reference Faith, Scanlon and Birch6) and modelling or monitoring to be associated with healthy eating(Reference Palfreyman, Haycraft and Meyer7,Reference Arredondo, Elder and Ayala8) . Further, concerns and perceptions of the child’s weight play a key role in prompting the use of certain feeding practices. For example, parents with greater concerns or perceptions of overweight might use more restrictive feeding practices(Reference Swyden, Sisson and Morris9–Reference Min, Wang and Xue11), even though this approach does not necessarily contribute to weight management. Restrictions can modify the child’s food preferences as well as his/her natural response to signals of hunger–appetite–satiety; higher intake of restricted foods in the absence of hunger might occur(Reference Jansen, Mulkens and Emond12–Reference Birch, Fisher and Davison14).
In children of pre-school age, almost all studies related to feeding practices have been conducted in first-world countries, such as the UK(Reference Carnell and Wardle15), Australia(Reference Corsini, Danthiir and Kettler16), Sweden(Reference Nowicka, Sorjonen and Pietrobelli17) or the USA and France(Reference de Lauzon-Guillain, Musher-Eizenman and Leporc18). Two studies have focused on Hispanic families within the USA(Reference Anderson, Hughes and Fisher19,Reference Seth, Evans and Harris20) , whereas only one has been conducted in a Latin American country (Chile)(Reference Santos, Kain and Dominguez-Vásquez21). Feeding practices vary by country, as suggested by a study which revealed that, while monitoring and restriction were more prevalent in France, using food as reward was more prevalent in the USA(Reference de Lauzon-Guillain, Musher-Eizenman and Leporc18). Moreover, feeding practices might differ by sociocultural background, as shown by Somaraki et al.(Reference Somaraki, Eli and Ek22) in Sweden. They documented that non-European-born mothers were more concerned about their child’s weight than European-born mothers; and maternal concerns explained 52 % of the difference in restriction between Swedish-born and non-European-born mothers. Hispanic and Mexican-origin mothers are reportedly prone to misinterpreting overweight as a synonym for good health(Reference Gauthier and Gance-Cleveland23,Reference Guendelman, Fernald and Neufeld24) . Latina mothers worry more about underweight than overweight and prefer their child to be heavier(Reference Guendelman, Fernald and Neufeld24–Reference Rosas, Harley and Guendelman26). In addition, Mexican mothers show affection to their family through food and the Mexican culture has been generally described as one emphasizing respect for authority, where parents predominantly employ an authoritarian feeding style(Reference Flores-Peña, Acuña-Blanco and Cárdenas-Villarreal27).
Little is known about which feeding practices prevail and what factors differentiate mothers who use certain feeding practices focused on parental control in Mexico. Therefore, the present study aimed to expand current research by estimating the prevalence of six feeding practices and exploring the associations between mothers’ concern about their child’s excess weight, mothers’ perception of their child’s overweight/obesity, sociodemographic characteristics and those feeding practices of Mexican mothers with children aged 2–6 years. Such estimations are essential to shedding light on modifiable family factors and developing effective interventions to promote healthy child eating habits and preventing obesity in pre-school children.
Methods
The present cross-sectional study was conducted between August and December 2016 in north-eastern Mexico.A consecutive selection was made of mothers aged at least 18 years who had singleton children aged 2–6 years with no endocrine disease, medical restriction on certain foods (e.g. lactose or gluten intolerance, allergy) or visible genetic malformations (e.g. Down syndrome; n 507). The mother had to be the primary person caring for the child feeding, without full-time help from a grandmother or nursery daycare. Mothers of young children were recruited from primary-care clinics selected to ensure participants from every urban municipality of the metropolitan area of Monterrey (seven districts). In the primary-care clinic waiting rooms, potential participants were approached by a member of the research team who invited them to participate. The sample size was large enough for a CI of 95 % and a margin error of 4 %, given prevalence results were between 30 % (regulation and not monitoring) and 60 % (not modelling)(28). The protocol was submitted for approval and registration to the Committees of Research, Ethics, and Biosecurity. All participants were verbally informed about the aims and procedures of the study and they provided oral informed consent before enrolment; their children also provided verbal assent. Research team members were available to answer any questions.
Maternal feeding practices
We found several instruments for measuring feeding practices in the literature, among them the Child Feeding Questionnaire (CFQ) and the Preschooler Feeding Questionnaire (PFQ)(Reference Corsini, Danthiir and Kettler16,Reference Anderson, Hughes and Fisher19,Reference Seth, Evans and Harris20,Reference Birch, Fisher and Grimm-Thomas29–Reference Baughcum, Powers and Johnson31) , whose Spanish versions have been administered to Hispanic mothers within the USA(Reference Seth, Evans and Harris20,Reference Kong, Vijayasiri and Fitzgibbon30) . The CFQ has been used in Chile(Reference Santos, Kain and Dominguez-Vásquez21). In particular, the Comprehensive Feeding Practices Questionnaire (CFPQ) includes twelve feeding practices and it has been validated in English with a sample of highly educated White participants(Reference Musher-Eizenman and Holub32). We assembled a short but wide-ranging questionnaire that was subject to content, construct and convergent validity analysis prior to estimating the feeding practices’ prevalence and before exploring associations. Participants responded using a 5-point Likert scale (1 = ‘never’, 5 = ‘always’).
The content validity stage consisted of a compilation of items selected per category of interest, as follows:
1. Pressure to eat, with items identified in the CFQ and CFPQ. These were mixed because we were interested in some items of the CFPQ not available in the CFQ and vice versa, although some coincided.
2. Restriction, with items identified in the CFQ and CFPQ. These were also mixed because we were interested in some CFPQ items not available in the CFQ and vice versa; some coincided.
3. Child control, with items identified in the CFPQ and PFQ (they were unavailable in the CFQ); most of them coincided.
4. Regulation, with items identified in the CFPQ, PFQ and Parental Feeding Style Questionnaire(Reference Carnell and Wardle15); most of them coincided.
5. Monitoring, with items identified in the CFQ and CFPQ (they were unavailable in the PFQ); most of them coincided.
6. Modelling, with items identified in the CFPQ (they were unavailable in the CFQ and PFQ). We also considered energy-dense food discouraging and nutrient-dense food encouraging items identified in an instrument developed by Murashima et al. (Reference Murashima, Hoerr and Hughes33); only one item coincided.
A panel of public health experts then eliminated duplicate items and selected three to five items per category after a thorough examination of the contents. The proposed items were translated back to English to verify their equivalence to the content in the original language.
The construct validity consisted of an exploratory factor analysis, which revealed twenty-two items with factor loadings of ≥0·30 constituted in six dimensions. Proposed items for Pressuring to Eat, Regulation, Child Control and Monitoring subscales were maintained, but some Modelling and Restriction items experienced changes. The Modelling subscale initially consisted of five items but ended with three; two from the Murashima et al.(Reference Murashima, Hoerr and Hughes33) instrument and one originally planned for the Restriction subscale. The Restriction subscale retained only two items that both belonged to the CFPQ. These concerned access to unhealthy foods and the subscale’s name was adjusted to ‘Food Restriction’. Convergent validity was assessed with the Caregiver’s Feeding Styles Questionnaire. As expected, Pressuring to Eat (always) correlated positively and Monitoring (never) correlated negatively with how strongly parents encouraged eating (demandingness dimension(Reference Hughes, Power and Orlet Fisher34–Reference Pai and Contento36); ρ = 0·60 and ρ = −0·20, respectively; P < 0·01). In addition, both subscales differentiated authoritarian and uninvolved mothers(Reference Hughes, Power and Orlet Fisher34); pressure to eat was greater in mothers exhibiting the authoritarian style than in those exhibiting the authoritative, indulgent or uninvolved style (mean (sd): 3·7 (0·9), 3·3 (0·8), 2·6 (0·8) and 2·5 (0·9), respectively; P < 0·0001). Monitoring was lower in mothers exhibiting the uninvolved style than in those exhibiting an authoritative style (mean (sd): 2·5 (1·2) and 2·0 (1·0), respectively; P < 0·01). In summary, the subscales were established as follows: Monitoring (extent to which the mother keeps track of her child’s eating; three items, α = 0·7); Modelling (extent to which the mother eats unhealthy foods eating in front of the child; three items, α = 0·4); Pressuring to Eat (mother’s attempts to increase her child’s food intake by insisting on eating more, especially during mealtimes; five items, α = 0·6); Regulation (use of food for rewarding or modifying the child’s emotional status; four items, α = 0·4); Food Restriction (extent to which the mother allows the child easy access to unhealthy snack foods; two items, α = 0·8); and Child Control (extent to which the mother allows the child take control of his/her eating behaviour; five items, α = 0·6).
Maternal perceptions and demographics
The mother’s concern for the child’s weight was measured by asking what she was most concerned about: (i) her child’s excess weight; (ii) her child’s low weight; or (iii) neither one. The perception of the child’s weight was measured by showing the mother sketches designed for parents of children aged 2–5 years, matched with the child by sex(37). There were seven figures arranged in order from least to most body weight (1 = extreme thinness, 7 = obesity). The mother was asked to indicate which figure looked more like her son’s/daughter’s weight. Then, the perceived child’s weight was classified as underweight, normal weight, overweight or obese. We did not try to measure misperception; therefore, the weight perception was not necessarily aligned with the child’s actual weight. The mothers also provided information on their own age, education, occupation and marital status; and their child’s age, sex and birth order.
Other study variables
The eating style of the child was evaluated using the Food Responsiveness subscale of the Children’s Eating Behaviour Questionnaire, based on the child’s response to signals of desire to eat(Reference Wardle, Guthrie and Sanderson38). This variable was included for statistical control purposes, given its positive correlation with feeding practices, in particular food restriction(Reference Silva Garcia, Power and Fisher39,Reference Webber, Hill and Cooke40) . Participants responded using a 5-point Likert scale with higher scores indicating greater food responsiveness (α = 0·8).
Anthropometric data
The mother’s and child’s height (in centimetres) and weight (in kilograms) were measured using a Taylor® (USA) portable digital scale calibrated daily and a wall stadiometer. Measurements were taken without shoes and with light clothing, with feet together, and with the heels, back and hips touching the wall. The mothers’ BMI was calculated as weight/height2 (kg/m2) and was classified as follows: underweight or normal weight, <25 kg/m2; overweight, 25–29 kg/m2; and obesity, ≥30 kg/m2. The children’s age- and sex-specific BMI Z-scores were calculated using the WHO 2006 Child Growth Standards as a reference and the Anthro plus Software v1.0.4 (nutritional survey module)(41,42) . The children’s weight status classification was based on BMI percentiles for age and sex. BMI < 5th percentile was considered underweight, BMI = 5th–84th percentile as normal weight, BMI = 85th–94th percentile as overweight and BMI ≥ 95th percentile as obesity.
Procedures
Mothers were interviewed by trained personnel using a structured interview protocol. Interviews were conducted in a private room in the clinic before or after the clinic visit, and ranged in duration from 15 to 20 min. At the end of the interview, the mother’s and child’s weight and height were measured, following standardized anthropometric techniques. The trained personnel consisted of one registered dietitian and two medical interns.
Plan for analysis
The analysis consisted of descriptive statistics and t tests for comparing quantitative variables. When the distribution of a variable was not normal, the Mann–Whitney test was applied. Mean scores were obtained for every feeding practice (possible range 1–5). Then, the scores were dichotomized: utilizing the code of 1 for scores of ≥3 to indicate moderate-to-high use; and the code 0 for scores of <3 to indicate low-to-non-use. Before dichotomization, the monitoring scores were reversed so that the moderate-to-high use category for all practices denoted a less-than-optimal practice (not monitoring, not modelling, pressuring to eat, regulation, easy access to unhealthy foods, child control). Subsequently, the point prevalences and 95 % CI were estimated and a multivariate logistic regression analysis was performed; the feeding practice being examined was the dependent variable, whereas the independent variables were concern and perception of the child’s weight, age, education and obesity status, for mothers; and age, sex and birth order, for children. Six separate multivariate logistic regression models were run; all were adjusted for food responsiveness. Concern about the child’s weight was the only variable with 8 % missing values, which corresponded to ambiguous answers that were excluded (mothers expressed they were concerned about their child’s low weight and their child’s excess weight. After a few interviews, the question was rephrased to ‘most concerned about’).
Results
Children’s profile
The children’s mean age, sex, birth order and weight status results are presented in Table 1. The mean (sd) food responsiveness score was 2·5 (1·2) (possible range 1–5); 15·4 % registered the highest score (child was always/almost always asking for food or if allowed to, would be eating most of the time).
Table 1 Children’s and mothers’ sociodemographic characteristics; mothers aged ≥18 years who were in charge of feeding a singleton child aged 2–6 years (n 507), north-eastern Mexico, August–December 2016
Mothers’ profile
The mothers’ mean age, education, occupation, marital and weight status results are also presented in Table 1. More than half of the mothers (58 %) were concerned about their child’s low weight and the rest were concerned about their child’s excess weight; 0 % answered neither one. Furthermore, 26·0 % perceived their child to be underweight, 47·9 % perceived their child as normal weight and 26·0 % perceived their child as overweight or obese.
Maternal feeding practices prevalence
The item response distribution is presented in Table 2. The individual maternal feeding practice with the highest use was ‘My child should always/almost always eat all of the food on his/her plate’ (Pressuring to Eat subscale), followed by ‘I always/almost always drink sweetened beverages in front of my child’ (Modelling subscale), independently of the mother’s concern about her child’s excess weight or the mother’s perception of her child as overweight/obese (Fig. 1).
Table 2 Item response distribution, according to maternal feeding practices, among mothers aged ≥18 years who were in charge of feeding a singleton child aged 2–6 years (n 507), north-eastern Mexico, August–December 2016
Fig. 1 Individual maternal feeding practices† (
, PE4; 
, M2; 
, M1; 
, PE5; 
, PE2; 
, CHC5; 
, CHC1), overall and according to mother’s concern about excess weight and mother’s overweight/obesityperception, among mothers aged ≥18 years who were in charge of feeding a singleton child aged 2–6 years (n 507), north-eastern Mexico, August–December 2016. †PE, Pressuring to Eat; M, Modelling; CHC, Child Control. PE4, My child should always/almost always eat all of the food on his/her plate; M2, I always/almost always drink sweetened beverages in front of my child; M1, I always/almost always eat sweets, candy or salty snacks in front of my child; PE5, I always/almost always try to get my child to eat more even if he/she says he/she is finished eating; PE2, If my child eats only a small helping, I always/almost always try to get him/her to eat more; CHC5, I always/almost always let my child eat between meals whenever he/she wants; CHC1, I always/almost always let my child choose the foods from what was served
The mean scores for maternal feeding practices are presented in Table 3. The mothers’ concern about their child’s excess weight generated differences in pressuring to eat (P < 0·001) and in food restriction (P < 0·05; there was greater pressuring to eat and greater easy access to unhealthy foods if the mother had no concern about her child’s excess weight). The mothers’ perception of the child as overweight/obese generated differences in modelling (P < 0·001) and in pressuring to eat (P < 0·0001; there was less modelling even if the mother perceived her child as overweight/obese; there was greater pressuring to eat if the mother did not perceive her child as overweight/obese). After categorization, the most frequent maternal feeding practices were not modelling and pressuring to eat, followed by food restriction (easy access to unhealthy foods) and child control; these practices prevailed even if the mother was concerned about her child’s excess weight or perceived her child as overweight/obese (Fig. 2).
Table 3 Descriptive statistics, overall and according to mother’s concern about excess weight and mother’s perception of child’s overweight/obesity, among mothers aged ≥18 years who were in charge of feeding a singleton child aged 2–6 years (n 507), north-eastern Mexico, August–December 2016
*P < 0·05, **P < 0·001, ***P < 0·0001.
† 5 = always; the higher the score, the greater use of monitoring, pressuring to eat, regulation, easy access to unhealthy food and child control.
‡ 5 = never; the higher the score, the less use of modelling.
§ Mean value was significantly different by child’s excess weight concern.
‖ Mean value was significantly different by child’s overweight/obesity perception.
Fig. 2 Maternal feeding practices prevalence, with their 95 % CI represented by vertical bars, (a) overall and (b) according to mother’s concern about excess weight and mother’s perception of child’s overweight/obesity (, modelling (absence); , pressure to eat; , food restriction†; , child control; , monitoring (absence); , regulation), among mothers aged ≥18 years who were in charge of feeding a singleton child aged 2–6 years (n 507), north-eastern Mexico, August–December 2016. †Easy access to unhealthy foods
Maternal feeding practice characterization according to potential motivators and sociodemographic profile
The multivariate analysis revealed associations that varied across maternal feeding practices. The odds of easy access to unhealthy foods were >1 even if the mother was concerned about her child’s excess weight (OR = 1·5; 95 % CI 1·1, 2·3). Perception of the child’s obesity was associated with less pressuring to eat (OR = 0·6; 95 % CI 0·4, 0·9), whereas the odds of child control were >1 even if the mother perceived her child as overweight/obese (OR = 1·7; 95 % CI 1·1, 2·7). The sociodemographic factor that was related to the highest number of feeding practices was education: a higher education was associated with more pressuring to eat, less regulation and less easy access to unhealthy foods; or monitoring was less absent (Table 4).
Table 4 Multivariate logistic regression analysis for maternal feeding practices† among mothers aged ≥18 years who were in charge of feeding a singleton child aged 2–6 years (n 507), north-eastern Mexico, August–December 2016
AOR, adjusted odds ratio, considering all variables exposed in the model.
*P < 0·05, **P < 0·01, ***P < 0·001.
† Six separate multivariate logistic regression models were run; the maternal feeding practice being examined was the dependent variable, whereas the independent variables were mother’s and child’s characteristics.
‡ Monitoring scores were reversed previous dichotomization, so that the moderate-to-high use category for all practices denoted a less-than-optimal practice.
Discussion
The present study contributes to knowledge on the magnitude of feeding practices and factors distinguishing utilization in a population predominantly characterized as young; with a partner; with a first or second child of pre-school age; with an elementary, secondary or college education; and with residence in an urban region with high levels of childhood obesity. More than three of every ten pre-schoolers were overweight or obese, a figure far above the Mexican national mean of one in every ten children under 5 years(Reference Gutiérrez, Rivera-Dommarco and Shamah-Levy3). Both statistics are based on equivalent WHO BMI-for-age and sex criteria to define overweight/obesity. Discrepancies between the prevalence rates in north-eastern Mexico and national estimates of childhood obesity have been explained by economic development: there is a higher prevalence of overweight and obesity in areas of greater development, for example northern Mexico and the Mexico City region; there is a lower prevalence in areas of less development, for example the central and southern regions. As early as 1999, the National Health and Nutrition Survey uncovered a growing disparity in food and nutrient intakes between regions of the country. In 2012, that survey revealed greater intakes of added sugars and saturated fat by children aged 1–4 years from the north v. centre and south; urban v. rural; and high v. medium/low socio-economic status (P < 0·05)(Reference López-Olmedo, Carriquiry and Rodríguez-Ramírez43).
The mean scores on feeding practices were estimated. The Pressuring to Eat mean score was 3·0, which was lower than that reported in Chile (≥3·6)(Reference Santos, Kain and Dominguez-Vásquez21), but higher than that reported in Spanish-speaking Hispanics in the USA(Reference Seth, Evans and Harris20), the UK(Reference Carnell and Wardle15) and Australia(Reference Corsini, Danthiir and Kettler16), all of which were ≤2·7. In these countries, a scale very similar in content to ours had been employed. Differences were also observed in monitoring and child control. For Monitoring, the mean Mexican score was lower than that of Australia and the USA (Hispanics); and for Child Control, higher than that of France(Reference Corsini, Danthiir and Kettler16,Reference de Lauzon-Guillain, Musher-Eizenman and Leporc18,Reference Anderson, Hughes and Fisher19) . In these countries, a scale very similar in content to ours had also been employed. These results support previous suggestions that feeding practices vary by sociocultural background(Reference de Lauzon-Guillain, Musher-Eizenman and Leporc18,Reference Somaraki, Eli and Ek22) . In terms of percentages, not modelling ranked first in frequency; more than six of every ten mothers did not use modelling strategies and consumption of sweetened beverages, candy or snacks in front of the child was a common practice. Regrettably, the multivariate analysis showed that neither concern nor perception of overweight/obesity was associated with modelling, like Lauzon-Guillain et al.(Reference de Lauzon-Guillain, Musher-Eizenman and Leporc18), who did not find an association either. Maternal feeding practices provide an opportunity to model good eating habits. A good eating behaviour by mothers relates to children’s higher intake of healthy foods and lower consumption of sweet snacks and fast foods(Reference Quah, Syuhada and Fries44,Reference Yee, Lwin and Ho45) . The importance of teaching by example needs to be reinforced.
Pressuring to eat was a frequent feeding practice; more than five of every ten mothers used it. ‘The child should eat all the food on the plate’ was the most commonly reported individual maternal feeding practice. The mothers’ perception of their child as overweight/obese was associated with lower odds of pressuring to eat, suggesting that the mothers were indeed responding accordingly. However, four to five of every ten mothers pressured their child to eat despite concern about their child’s excess weight and having a perception of the child as overweight/obese. This maternal feeding practice could be explained by two factors: the sociocultural significance of food in Mexico and the lack of distinction between pressuring to eat healthy and unhealthy foods. Future research must consider the quality of the diet in this context while acknowledging that pressuring to eat is problematic as a practice, irrespective of the type of food that the child is being pressured to eat. Regarding food restriction, about four to five of every ten mothers allowed easy access to unhealthy foods. Restriction involves a type of limit to access palatable and usually unhealthy foods. Some parents tend to use restrictive feeding practices such as withholding high-fat/high-sugar foods, especially when concern about child’s excess weight or a perception of overweight/obesity exists(Reference Gray, Janicke and Wistedt46). However, in the present study, the odds of easy access to unhealthy foods were higher despite concern about the child’s excess weight. Controlling the availability of foods is a vital aspect of healthy eating, but caution is needed because when children are exposed to restricted items, they could be more likely to choose these foods and consume them in excess(Reference Jansen, Mulkens and Jansen13). Another practice, child control, was permitted by more than four of every ten participants, even if concern or overweight/obesity perception was present. Moreover, the odds of child control were higher despite overweight/obesity perception. Giving children control over the foods they select and consume has been associated with a greater intake of sweet snacks and fast foods(Reference Quah, Syuhada and Fries44). Hence, it is important to raise awareness on the negative consequences of this feeding practice. Regulation and not monitoring were the least common maternal feeding practices, and they were not associated with concern or overweight/obesity perception.
Some sociodemographic variables characterized maternal feeding practices, but the associations varied across practices. Health programmes should consider the mother’s education, given its association with four of the six feeding practices examined (educated mothers used more pressuring to eat, less regulation and less easy access to unhealthy foods; or monitoring was less absent). Kröller and Warschburger(Reference Kröller and Warschburger47) found that the higher the mother’s education, the greater the monitoring; and Musher-Eizenman et al.(Reference Musher-Eizenman, de Lauzon-Guillain and Holub48) found that the higher the mother’s education, the less the regulation. The finding of higher educational levels associated with greater pressuring to eat differs from other studies that did not report such an association(Reference Somaraki, Eli and Ek22,Reference McPhie, Skouteris and McCabe49) . High parental education has been related to health consciousness in food choices(Reference Northstone and Emmett50,Reference Kant and Graubard51) and this observed association could be connected to pressure to eat healthy foods. According to Hinnig et al.(Reference Hinnig, Monteiro and de Assis52), children with highly educated parents who lived in highly developed countries tended to have a healthier diet, but the association was not clear in medium and less-developed countries. More research is needed to clarify the link between education, pressure to eat and the quality of the diet.
Another associated sociodemographic variable was the mother’s overweight/obesity status that was associated with higher odds of easy access to unhealthy foods and letting the child control her/his eating. Santos et al.(Reference Santos, Kain and Dominguez-Vásquez21)reported greater use of restriction (CFQ subscale) by heavier mothers, but only in girls; and Haycraft et al.(Reference Haycraft, Karasouli and Meyer53) reported greater use of child control by obese mothers. An older maternal age was associated with less absence of modelling, less pressuring to eat and less regulation, whereas an older child age was associated with greater absence of modelling, greater regulation and greater easy access to unhealthy foods. Additionally, children’s male sex was associated with greater child control and less use of monitoring. In other words, there was less child control and greater use of monitoring with daughters. Latino mothers (Mexican-American) engage in more restraining behaviours with their daughters than with their sons. A cultural awareness of standards for female physical attractiveness might influence parents’ feeding practices(Reference Olvera-Ezzell, Power and Cousins54). In other countries, such as Poland, parents use the regulation feeding practice less often in 5-year-old girls than in boys regardless of weight status(Reference Lipowska, Lipowski and Jurek55). Other authors have not found maternal feeding practice differences based on the child’s sex(Reference Somaraki, Eli and Ek22,Reference Gray, Janicke and Wistedt46,Reference Kröller and Warschburger47) . Finally, higher birth order was associated with greater absence of modelling and greater use of pressuring to eat. Parents might struggle to persuade one child to eat enough, whereas for a sibling, they struggle to stop him or her from eating too much(Reference Gibson and Cooke56). Farrow et al.(Reference Farrow, Galloway and Fraser57) provided evidence that the parent–child relationships involving food can vary within families; parents use more pressure to eat with siblings who are slower to eat, are fussier and are less responsive to food. It is also plausible that parental attention towards each child during mealtime situations decreases with an increasing number of children and the presence of siblings might be protective against the development of picky eating(Reference Hafstad, Abebe and Torgersen58). More research is needed to clarify this particular finding.
Limitations of the study
The present study had some limitations. The sample size was large and the participating mothers displayed a wide range of education. However, all lived in an urban location and thus the present results cannot be generalized to rural settings. In the future, there is also a need to study other parental characteristics that were not considered here such as mother’s own eating habits, which have been found to influence feeding practices (e.g. restrained, uncontrolled and emotional eating)(Reference de Lauzon-Guillain, Musher-Eizenman and Leporc18). The present investigation did not consider primary caregivers other than mothers. However, statistics from the 2017 National Survey of Employment and Social Security showed that 72·6 % of Mexican mothers or 72·0 % of mothers with children aged ≤6 years from the north-east of Mexico take care of their children because they do not work; less than 5 % use a daycare centre; and the rest (23 %) receive help from a grandmother, their husband or a friend to care for the child while they work(59). Therefore, the generalizability of our results to a wider population of mothers of pre-school children is plausible.
Conclusions
The present study revealed that the most frequently used practices were pressuring to eat and child control; the most frequently absent practices were modelling and food restriction. All of them prevailed despite concern about the child’s excess weight or a perception of the child as overweight/obese. Therefore, these maternal feeding practices need to be improved and there is need to focus on mother’s concern about the child’s excess weight and mother’s perception of the child as overweight/obese. It is essential to determine what factors make mothers use less healthy feeding practices to optimize intervention planning, principally in this region with high child obesity rates. The sociodemographic factor that differentiated the highest number of feeding practices was education. Other factors were associated with certain maternal feeding practices, but the associations varied across practices. Feeding practices are potentially modifiable in favour of maintaining a healthy weight, and an understanding of feeding practices is fundamental to child health because it is the ideal growth stage for interventions aimed at healthy eating.
Acknowledgements
Acknowledgements: The authors gratefully acknowledge the unconditional collaboration of Dr Laura H. de la Garza Salinas, whose support was essential to the development of this study during the data collection phase. Financial support: Funding for this study was provided by the Health Research Fund of the Mexican Institute of Social Security (grant number FIS/IMSS/PROT/G16/1583), which had no role in the study design, collection, analysis or interpretation of the data, writing of the manuscript, or the decision to submit the paper for publication. Conflict of interest: All authors declare that they have no conflicts of interest. Authorship: A.M.S.M. and D.B.E.d.L. designed the study and wrote the protocol. G.E.M.F. and G.M.N.R. conducted literature searches and provided summaries of previous research studies. F.J.G.d.l.G., D.B.E.d.L. and G.E.M.F. contributed with data acquisition and data entry. A.M.S.M., D.B.E.d.L. and H.F.C.F. conducted the statistical analysis. A.M.S.M. and H.F.C.F. wrote the first draft of the manuscript and all authors contributed to and have approved the final manuscript. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Research and Ethics Committee No. 1944 of the Instituto Mexicano del Seguro Social. Verbal informed consent was obtained from all subjects and it was formally witnessed.
