Follow-up studies among adults show that weight gain in young women is, on average, relatively large compared with older women(Reference Williamson, Kahn and Remington1–Reference Wendel-Vos, Nooyens and Schuit3). Excessive weight gain during pregnancy, failure to lose weight in an appreciable period of time postpartum and weight gain during the postpartum period may well partly explain this weight gain among young women(Reference Linne, Dye and Barkeling4–Reference Rossner and Ohlin6). New mothers are more likely to gain weight than young adult women without children(Reference Ball, Brown and Crawford7). In addition, longitudinal studies have shown that postpartum weight retention predicts overweight in the long term(Reference Linne, Dye and Barkeling4, Reference Rooney, Schauberger and Mathiason5). Therefore, promoting weight control among new mothers is valuable in obesity prevention.
A decrease in physical activity during the transition into motherhood may explain the higher increase in weight gain among new mothers compared with non-mothers. A meta-analysis showed that mothers were less likely to be physically active than fathers and non-parents(Reference Bellows-Riecken and Rhodes8). In particular, mothers of young children (under the age of 5 years) seem to be at risk for physical inactivity(Reference Sternfeld, Ainsworth and Quesenberry9, Reference Nomaguchi and Bianchi10). Changes in nutritional behaviour before and after pregnancy are found inconsistently(Reference Cuco, Fernandez-Ballart and Sala11–Reference Olson13). However, this does not mean that there is no room left to improve nutritional behaviour among mothers. An unhealthy diet and a decrease in physical activity are both related to weight gain among mothers with young children(Reference Ball, Brown and Crawford7, Reference Harris, Ellison and Holliday14, Reference Olson, Strawderman and Hinton15).
Preliminary findings of a recently published Cochrane review(Reference Amorim, Linne and Lourenco16) extracted from six studies showed that among women recruited in the postpartum period (up to 12 months postpartum) a prescribed diet combined with structured exercise, or diet alone, compared with usual care, seemed to enhance weight loss(Reference Amorim, Linne and Lourenco16). However, this Cochrane review focused on the effect of a change in diet or exercise on change in weight without examining the factors that lead to behaviour change. In order to develop weight-loss programmes, it is useful to have insight into how these behaviours can be promoted.
In addition, it seems worthwhile to look at how mothers can be encouraged to attend such programmes. The years it takes to establish a young family involve a major life transition for women in terms of their social, occupational and biological lives(Reference Lewis and Ridge17). Most mothers of young children experience some constraints, for example lack of time and energy, and the responsibilities of child care. These factors may influence their interest in and ability to attend interventions(Reference Brown, Brown and Miller18).
Targeting nutrition and physical activity interventions to these mother-specific factors may enhance both their effectiveness and attendance rate. Therefore, the purpose of the present review is to gain insight into targeted intervention components which may contribute to the attendance and effectiveness of interventions promoting physical activity and/or healthy eating to mothers with young children (0–5 years).
Materials and methods
Search strategy and data sources
Intervention studies were identified through a structured electronic database search in OVID Embase and MEDLINE. The search strategy, shown in Table 1, included searching on words in the title or abstract and MeSH terms. The search was limited to articles with MeSH term ‘human’ (or ‘humans’ in MEDLINE) and articles published from 1997 to 2009.
*WIC, Special Supplemental Nutrition Program for Women, Infants, and Children. The WIC programme provides food, nutrition counselling and access to health resources for low-income pregnant and postpartum women with children up to the age of 5 years in the USA (http://www.fns.usda.gov/wic/).
Selection criteria
Studies eligible for inclusion were: intervention studies aimed at promotion of physical activity and/or healthy eating; the target group was mothers (aged 18+ years) with young children (aged 0–5 years); the study design included a pre- and post-measurement among an intervention and a control group. Studies were excluded if the intervention was a prescribed diet or exercise programme with no attention paid to lifestyle change (e.g. sessions with prescribed exercise to achieve a certain percentage of the heart rate reserve for a certain amount of time). Also, studies were excluded if the study population consisted of mothers with mental health problems because this was considered to be a different subpopulation. Finally, studies were excluded if no information was available about the age of the children. An exception was made for participants of the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), since this programme provides food, nutrition counselling and access to health resources specifically for low-income pregnant and postpartum women with children up to the age of 5 years in the USA (http://www.fns.usda.gov/wic/).
Procedure
The titles and abstracts were assessed by the first author (M.A.H.). In cases of doubt the second author was consulted (K.H.). If a reference was considered to be relevant, the full paper was retrieved. Full papers were assessed against the review selection criteria. Reference lists of relevant articles were scanned. Extracted data included study characteristics, intervention characteristics, intervention components targeted at mothers, and reported intervention attendance and effectiveness. By intervention components targeted at mothers, we mean the intervention components adapted to the situation of mothers with young children. By attendance, we mean not only attendance at intervention classes but also exposure to the intervention in other ways (i.e. dose received).
Information on the characteristics of the studies was collected so that the quality and generalizability of the studies could be assessed. The following information about study characteristics was extracted: the country where the study was conducted, design (true experiment or quasi-experiment), description of the intervention and comparison group, response and attrition rate, and primary outcome measures.
Additionally, data were extracted separately for intervention components presumed to be related to attendance and effectiveness. Information on both kinds of results is important, since a low attendance or a low effectiveness will result in a low impact of the intervention. The results were analysed in a descriptive manner.
For attendance, intervention components targeted at mothers were summarized. Subsequently, the intervention components in studies with a low attendance were compared with those in studies with a high attendance. In addition, information was collected about possible explanations for non-attendance.
For effectiveness on physical activity and healthy eating, intervention components specifically targeted at mothers were summarized, related to effectiveness (statistically significant effects or not) and additional information was collected on the possible explanations for their effectiveness. Furthermore, if key elements of a systematic development of interventions were used (use of formative research, theory, behavioural change strategies, evidence and targeting other characteristics than being a mother), these elements were then also summarized. Such elements should be taken into account when interpreting the results of the relationship between targeted components and the effectiveness of interventions because a systematic intervention development increases the likelihood of obtaining the intended positive effects(Reference Bartholomew, Parcel and Kok19).
Results
Study selection
The initial search identified 1556 publications. After eliminating duplicates and reviewing the titles and abstracts of these publications, the total was reduced to forty-two. These forty-two articles were completely reviewed, after which thirty-one publications were excluded because they did not meet one or more of the inclusion criteria. The main reasons for exclusion were lack of a control group; a different outcome measurement than physical activity or nutritional behaviour; or because mothers of children above 5 years of age were included in the study. Moreover, four review articles were found, all describing studies which mainly investigated the relationship between weight-related behaviours and weight only using prescribed diet or exercise with no attention for lifestyle change(Reference Amorim, Linne and Lourenco16, Reference Dewey20–Reference Kuhlmann, Dietz and Galavotti22). Equally, two intervention studies were excluded for this reason. Finally, some articles seemed to report the results of the same study because they described the same intervention and the same number of participants. In one case, the article was chosen which described the effect evaluation(Reference Kinnunen, Pasanen and Aittasalo23) instead of the feasibility study(Reference Kinnunen, Aittasalo and Koponen24). In the other two cases, articles were selected which evaluated the effect on actual health-related behaviour(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26) instead of on the stages of change(Reference Feldman, Damron and Anliker27) or purchasing of fresh fruit and vegetables(Reference Herman, Harrison and Jenks28). After checking references no other relevant articles were found. Therefore, a total of eleven articles were eligible, describing twelve interventions (Table 2).
WIC, Special Supplemental Nutrition Program for Women, Infants, and Children; F&V, fruit and vegetables; PA, physical activity.
Study characteristics
The eligible studies differed on many characteristics (Table 3). Most of the studies were conducted in the USA(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26, Reference Campbell, Carbone and Honess-Morreale29, Reference Havas, Anliker and Greenberg30, Reference Leermakers, Anglin and Wing31, Reference Fahrenwald, Atwood and Walker32, Reference Østbye, Krause and Lovelady33), two in Australia(Reference Miller, Trost and Brown34, Reference Watson, Milat and Thomas35), one in Canada(Reference Cramp and Brawley36) and one in Finland(Reference Kinnunen, Pasanen and Aittasalo23). Four interventions were implemented in a clinical setting(Reference Kinnunen, Pasanen and Aittasalo23, Reference Leermakers, Anglin and Wing31, Reference Østbye, Krause and Lovelady33, Reference Watson, Milat and Thomas35), three in a community setting(Reference Miller, Trost and Brown34, Reference Cramp and Brawley36) and five interventions were implemented within or alongside the WIC(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26, Reference Campbell, Carbone and Honess-Morreale29, Reference Havas, Anliker and Greenberg30, Reference Fahrenwald, Atwood and Walker32). Three studies were quasi-experiments(Reference Kinnunen, Pasanen and Aittasalo23, Reference Herman, Harrison and Afifi26, Reference Watson, Milat and Thomas35) and eight true experiments(Reference Havas, Anliker and Damron25, Reference Campbell, Carbone and Honess-Morreale29–Reference Miller, Trost and Brown34, Reference Cramp and Brawley36).
PA, physical activity; F&V, fruit and vegetables; IPAQ, International Physical Activity Questionnaire; PAR, 7 d Physical Activity Recall.
*Passively recruited participants (via a local community newspaper article on postnatal fitness).
†798 reached by phone, 136 not eligible, sixty-four did not complete baseline (=598 eligible), 148 refused; response calculated as 1−(148/598)×100 = 75·3 %.
Six studies included only postpartum mothers with children up to the age of 1 year(Reference Kinnunen, Pasanen and Aittasalo23, Reference Herman, Harrison and Afifi26, Reference Leermakers, Anglin and Wing31, Reference Østbye, Krause and Lovelady33, Reference Watson, Milat and Thomas35, Reference Cramp and Brawley36), in contrast with the five other studies which included mothers of children up to 5 years(Reference Havas, Anliker and Damron25, Reference Campbell, Carbone and Honess-Morreale29, Reference Havas, Anliker and Greenberg30, Reference Fahrenwald, Atwood and Walker32, Reference Miller, Trost and Brown34). Moreover, the study populations differed on socio-economic and ethnic composition. The American studies directed at WIC mothers mainly included a relatively high percentage of low-income women and women from minority groups(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26, Reference Campbell, Carbone and Honess-Morreale29, Reference Havas, Anliker and Greenberg30, Reference Fahrenwald, Atwood and Walker32). Control groups consisted mainly of usual care(Reference Kinnunen, Pasanen and Aittasalo23, Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30) or no intervention(Reference Campbell, Carbone and Honess-Morreale29, Reference Miller, Trost and Brown34, Reference Watson, Milat and Thomas35). The study population size varied from forty-four to 3122 participants.
The variation in response rates was large, 30·8 % to 91·2 %, but almost half of the studies gave no information about response rates. No information could be obtained about response rate for one study since respondents were recruited passively through a local community newspaper article on postnatal fitness(Reference Cramp and Brawley36). In the other studies, respondents were recruited actively, i.e. were personally approached to participate in the study. Short-term attrition rates varied from 7·6 % to 31·0 %.
Three interventions were directed at promoting healthy eating, physical activity as well as weight change(Reference Kinnunen, Pasanen and Aittasalo23, Reference Leermakers, Anglin and Wing31, Reference Østbye, Krause and Lovelady33); four interventions were aimed at promoting healthy eating(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26, Reference Campbell, Carbone and Honess-Morreale29, Reference Havas, Anliker and Greenberg30); and four at physical activity(Reference Fahrenwald, Atwood and Walker32, Reference Miller, Trost and Brown34, Reference Watson, Milat and Thomas35). In all studies, nutritional behaviour was measured by self-reports; physical activity by self-reports or objective measurement (pedometer/accelerometer); and weight loss by self-reports or anthropometric measurements.
Attendance
Within the six studies that reported attendance, full attendance ranged from 9 % to 92 % (Table 4). In two studies(Reference Havas, Anliker and Greenberg30, Reference Østbye, Krause and Lovelady33) subgroup analyses were performed. Older (>30 years) and higher educated mothers were more likely to attend in both studies. In one study participants were more likely to be white(Reference Østbye, Krause and Lovelady33), while in the other study there was no difference regarding attendance between ethnic groups(Reference Havas, Anliker and Greenberg30). Moreover, married(Reference Østbye, Krause and Lovelady33), higher-income(Reference Østbye, Krause and Lovelady33) and unemployed(Reference Havas, Anliker and Greenberg30) mothers were more likely to attend compared with mothers who did not participate.
PA, physical activity; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children.
Components within intervention studies reaching a low v. a high attendance
Two studies succeeded in reaching a high attendance of mothers(Reference Kinnunen, Pasanen and Aittasalo23, Reference Cramp and Brawley36) (Table 4). The intervention of Kinnunen et al.(Reference Kinnunen, Pasanen and Aittasalo23) was embedded within routine visits to child health clinics, normally attended by 98 % of the target population. This resulted in a 92 % attendance to their intervention. Furthermore, Cramp and Brawley(Reference Cramp and Brawley36) reached an attendance of 75 %. Their attempts involved conducting the intervention in a community-based fitness facility and providing child care for a nominal fee. Other interventions whereby child care was provided or an activity with mothers and children was organized did not result in high attendance(Reference Havas, Anliker and Damron25, Reference Østbye, Krause and Lovelady33, Reference Watson, Milat and Thomas35).
Moreover, several other approaches were used in order to increase attendance among mothers with young children, although with minor results (9–27 % attendance). To accommodate varying schedules of mothers, intervention sessions were repeated several times(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30, Reference Østbye, Krause and Lovelady33) or the intervention was held at the mothers’ preferred time of day(Reference Watson, Milat and Thomas35). Furthermore, intervention setting and time schedules were adapted to WIC voucher pickup to maximize the limited opportunities for reaching mothers(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30). One group activity with mothers – pram walking – was organized to overcome social isolation among other things(Reference Watson, Milat and Thomas35).
Explanations for attendance derived from additional attendance analyses
Four studies had conducted explorative attendance research through a questionnaire and/or focus group discussions(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30, Reference Watson, Milat and Thomas35) (Table 4). Results showed that reasons for non-attendance were, for instance, a lack of interest or withdrawal from the organization that had implemented the intervention, WIC(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30, Reference Damron, Langenberg and Anliker37). Moreover, some reasons mentioned for non-attendance were related to barriers specific to mothers, such as lack of child care or conflicting schedules(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30, Reference Østbye, Krause and Lovelady33, Reference Watson, Milat and Thomas35). Other reasons for non-attendance were more general, for example work or school conflicts, transport difficulties, sickness and mobility of participants(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30, Reference Østbye, Krause and Lovelady33, Reference Watson, Milat and Thomas35, Reference Damron, Langenberg and Anliker37).
Effectiveness
Six out of twelve interventions did not result in positive significant effects (Table 5). Two interventions resulted in inconsistent (i.e. significant and non-significant) results on varying types of physical activity, eating behaviours and weight change(Reference Kinnunen, Pasanen and Aittasalo23, Reference Leermakers, Anglin and Wing31), and four resulted in no significant effects(Reference Campbell, Carbone and Honess-Morreale29, Reference Miller, Trost and Brown34, Reference Watson, Milat and Thomas35). Regarding the remaining six interventions, positive significant changes were reported on physical activity(Reference Fahrenwald, Atwood and Walker32, Reference Miller, Trost and Brown34, Reference Cramp and Brawley36) and healthy eating(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26, Reference Havas, Anliker and Greenberg30) in the short term (1–8 months).
General: PA, physical activity; F&V, fruit and vegetables; %E, percentage of energy; TV, television; PAR, 7 d Physical Activity Recall; SES, socio-economic status; WIC, Special Supplemental Nutrition Program for Women Infants, and Children.
Theoretical frameworks: SoC, Stages of Change model; TTM, Transtheoretical Model; SLT, Social Learning Theory; SCT, Social Cognitive Theory; SET, Self-Efficacy Theory.
Behaviour change strategies: EDU, education; PC, discussing pros and cons; SS, providing/building social support; PE, peer educator; GP, guided practice; GS, goal setting; RF, reinforcement; DS, discussing strategies to overcome barriers; ST, skill training; INC, incentives; ML, modelling; FB, feedback; SM, self-monitoring; CB, capacity building.
Three studies conducted additional subgroup analyses(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30, Reference Østbye, Krause and Lovelady33). No intervention effects were found in subgroup analyses based on race, education, parity or BMI category in one study(Reference Østbye, Krause and Lovelady33). Both studies of Havas et al.(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30) found that mothers younger than 30 years and with at least high-school education showed significantly greater increases in healthy eating. One study also found this pattern for married and non-working women(Reference Havas, Anliker and Damron25). Furthermore, white women showed significantly greater increases in fruit and vegetable consumption in one study(Reference Havas, Anliker and Damron25), while in the other study significant changes in healthy eating were found in both black and white participants(Reference Havas, Anliker and Greenberg30).
Of the three studies measuring long-term results (7 months to 1 year)(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30, Reference Miller, Trost and Brown34), only Havas et al.(Reference Havas, Anliker and Damron25) reported long-term nutritional behaviour change. Henceforth, the description of the results will focus on short-term effects since these are available for all studies.
Components within interventions which resulted in significant v. non-significant effects
More systematically developed interventions were more likely to be effective (Table 5). Four interventions were developed using less than three out of five key components of a systematic development (i.e. formative research, theory-based, behaviour change strategies, evidence-based or targeted to socio-economic status/ethnicity), none resulting in significant effects(Reference Kinnunen, Pasanen and Aittasalo23, Reference Østbye, Krause and Lovelady33–Reference Watson, Milat and Thomas35). Out of the other eight interventions using three key components or more, six interventions appeared to be effective in changing physical activity and healthy eating positively(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26, Reference Havas, Anliker and Greenberg30, Reference Fahrenwald, Atwood and Walker32, Reference Miller, Trost and Brown34, Reference Cramp and Brawley36).
Of the seven interventions directed at promoting healthy eating, three showed positive statistically significant results(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26, Reference Havas, Anliker and Greenberg30), of which only one was targeted at motivations of mothers(Reference Havas, Anliker and Damron25). The motivational appeals regarded ‘set a good example for your children’ and ‘take care of yourself by eating more fruits and vegetables also after pregnancy’(Reference Havas, Damron and Treiman38). In addition, more general behavioural change strategies such as the use of peer educators(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30) and role modelling(Reference Campbell, Carbone and Honess-Morreale29) were used in order to promote healthy eating.
In contrast, all eight interventions trying to promote physical activity included components targeted at mothers; mostly at overcoming barriers specific for mothers (e.g. physical changes, lack of time, social support or energy; Table 3). Five studies with components targeted at mothers did not show significant results on physical activity(Reference Kinnunen, Pasanen and Aittasalo23, Reference Leermakers, Anglin and Wing31, Reference Østbye, Krause and Lovelady33–Reference Watson, Milat and Thomas35). Of these studies, three were aimed at multiple behaviour change(Reference Kinnunen, Pasanen and Aittasalo23, Reference Leermakers, Anglin and Wing31, Reference Østbye, Krause and Lovelady33). The two other interventions consisted of a single intervention; a brochure with tips for how to overcome barriers(Reference Miller, Trost and Brown34) or exercise – pram walking – groups(Reference Watson, Milat and Thomas35). These components were also elements in the three studies that showed significant results on physical activity(Reference Fahrenwald, Atwood and Walker32, Reference Miller, Trost and Brown34, Reference Cramp and Brawley36). For example, exercise classes were organized(Reference Cramp and Brawley36) and brochures were used with tips including how to overcome mother-specific barriers(Reference Fahrenwald, Atwood and Walker32, Reference Miller, Trost and Brown34). However, these studies had additional (targeted) components (i.e. were multi-component) to promote physical activity (as a single behaviour). Both Cramp and Brawley(Reference Cramp and Brawley36) and Fahrenwald et al.(Reference Fahrenwald, Atwood and Walker32) used in addition interactive counselling about perceived barriers and making plans on how to overcome these barriers. Furthermore, in the group III intervention of Miller et al.(Reference Miller, Trost and Brown34) intervention mothers were asked to participate in discussion groups to explore their perceived barriers for physical activity. Outcomes formed the basis for the development of intervention strategies.
Explanations for effectiveness derived from additional analyses
Some studies investigated, and mostly found, evidence for the influence of effective interventions through mediators (mainly social support and self-efficacy)(Reference Miller, Trost and Brown34, Reference Langenberg, Ballesteros and Feldman39, Reference Cramp and Brawley40) and attendance on physical activity or healthy eating(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30, Reference Østbye, Krause and Lovelady33) (Table 5). In addition, Leermakers et al.(Reference Leermakers, Anglin and Wing31) analysed the influence of adherence. The number of self-monitoring records returned was found to be significantly correlated with weight loss. Homework completion or telephone contact was not found to be associated with greater changes. No studies evaluated the targeted intervention components in relationship with effectiveness on physical activity. Regarding targeting, only the perception of pram walking groups by mothers was investigated in one study(Reference Watson, Milat and Thomas35). Mothers emphasized the importance of having walking routes that met the needs of women with babies and young children. The majority of mothers who joined a pram walking group did so for exercise, to get out of the house and to meet other young mothers.
Discussion
The aim of the present review was to gain insight into targeted intervention components which may contribute to the attendance and effectiveness of interventions promoting physical activity and/or healthy eating to mothers with young children. Six out of the twelve studies included here measured attendance, of which two reported a high attendance. Especially the embedding of the intervention in routine visits to child health clinics seems to increase attendance rates. Cited explanations for non-attendance were general as well as mother-specific factors such as lack of interest, withdrawal from intervention setting and conflicting schedules.
Moreover, only six of the interventions reported statistically significant effects. Of these, three studies found positive effects on physical activity and three on healthy eating. Effective interventions directed at physical activity were multi-component and included such elements as counselling on mother-specific barriers or community involvement in intervention development and implementation. Interventions that effectively increased healthy eating did not all have components targeted at mothers. Only one study used mother-specific motivational appeals to promote healthy eating. No studies evaluated the targeted intervention components in relationship with effectiveness of the intervention in particular.
The low attendance reported for most of the interventions shows the difficulty of reaching mothers of young children in programmes for promoting health behaviours – as mentioned by other authors(Reference Leermakers, Anglin and Wing31, Reference Østbye, Krause and Lovelady33, Reference Miller, Trost and Brown34, Reference Gore, Brown and West41, Reference Cody and Lee42). This is also reflected in the large number of attempts to increase attendance with only minor results. For example, in spite of the fact that lack of time is frequently found to be a major barrier for mothers(Reference French, Jeffery and Story43), targeted intervention components such as repeated intervention sessions to overcome scheduling problems did not result in high attendance rates. This may be due to the fact that underlying factors causing time constraints among mothers are neglected in such a strategy (see also Watson et al.(Reference Watson, Milat and Thomas35)). In a survey by Brown et al.(Reference Brown, Brown and Miller18), mothers mostly reported that having no time was due to commitments to children, housework and shopping. This may originate in restrictive role expectations towards mothers and the mother’s perception of ‘being a good mother’(Reference Lewis and Ridge17, Reference Miller and Brown44). Strategies that take these factors into account may provide opportunities for promoting attendance.
Furthermore, child care responsibilities and costs have been identified by mothers as two important barriers for attending weight-loss interventions(Reference French, Jeffery and Story43, Reference Setse, Grogan and Cooper45). However, studies that solved the problem of child care and had no costs did not consistently result in high attendance(Reference Havas, Anliker and Damron25, Reference Watson, Milat and Thomas35, Reference Cramp and Brawley36). This may be due to a lack of interest for lifestyle change on the part of mothers. In a previous qualitative study investigating perceived barriers to attend weight-loss programmes, in addition to costs and child care, disinterested mothers perceived lack of time, family duties and conflicts with working schedule as barriers more compared with interested mothers(Reference French, Jeffery and Story43). In intervention studies with no costs and solutions for child care, indeed a lack of interest was mentioned as one of the reasons for a low attendance by actively recruited participants(Reference Havas, Anliker and Damron25, Reference Damron, Langenberg and Anliker37). While another study with no costs and need for child care achieved a high attendance; in that study participants were recruited passively and thus were more likely to be highly motivated(Reference Cramp and Brawley36). Therefore, motivating mothers, for example through mass media campaigns in order to get them physically active or eat healthier, seems a worthwhile point for reaching more mothers. Since the well-being of children is very important to mothers(Reference Lewis and Ridge17), emphasizing that mothers are role models for their children might increase interest.
Because the well-being of children is important to mothers, it is reasonable that an intervention setting for the good of children will increase attendance. This is supported by the study of Kinnunen et al.(Reference Kinnunen, Pasanen and Aittasalo23), whereby an intervention was embedded within routine visits to child health clinics resulting in a high attendance. Embedding the intervention in the routine of mother and child also gets round the problems of lack of time and child care responsibilities. Two other studies showed, however, that embedding the intervention within mothers’ routine visits to an intervention setting might not contribute to attendance when there is withdrawal from this setting(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30).
With regard to the question on which intervention components contribute to the effectiveness on behaviour change, we found that all reviewed physical activity interventions included components targeted at mothers with young children. For physical activity, interactive counselling or discussion sessions about social and practical barriers perceived by mothers and making plans on how to overcome these barriers within the individual environment(Reference Fahrenwald, Atwood and Walker32, Reference Cramp and Brawley36) or local community settings(Reference Miller, Trost and Brown34) seems to promote physical activity among mothers. Such a strategy can address important constraints frequently cited by mothers, such as lack of time, energy, child care and social support, and obligations to other roles(Reference Bellows-Riecken and Rhodes8).
Only one of the interventions aimed at improving healthy eating had components targeted towards mothers, while three out of four studies aimed at promoting healthy eating as a single behaviour showed statistically significant results. This one study used mother-specific motivational appeals rather than targeting mother-specific barriers. This may be because changing dietary behaviour is easier for mothers when compared with changing their physical activity pattern. In a previous study, young adult women reported that they saw a whole range of healthy eating behaviours as highly feasible but not many physical activity behaviours(Reference Ball, Crawford and Warren46). Mothers did perceive fruit and vegetable consumption to be more feasible compared with non-mothers although they perceived leisure-time physical activity, physical activity for transport purposes and incidental physical activity as less feasible compared with women with no children.
Definite conclusions about which components may contribute to intervention effectiveness cannot be drawn, however, from the studies included in the current review since the actual contribution of components targeted at mothers of young children on the effectiveness has not been evaluated. Only the appreciation of pram walking groups was found by Watson et al.(Reference Watson, Milat and Thomas35). Therefore, the effectiveness on physical activity and on healthy eating may also depend on the multi-component character or the systematic development of most of the effective interventions included in the review. Furthermore, theory-based developed interventions aimed at reaching intervention effects on physical activity or healthy eating through theoretical constructs (i.e. mediators) that are important for mothers, such as self-efficacy and social support, might increase effectiveness(Reference Miller, Trost and Brown34, Reference Langenberg, Ballesteros and Feldman39, Reference Cramp and Brawley40). There is a need for more process evaluations which focus on the question of how intervention effects are obtained.
Furthermore, in order to disseminate the use of interventions, it is necessary to know to which subpopulations the programme effects can be generalized. Notable was that while all studies with mainly low-income or ethnic minority participants in the present review had targeted their interventions to these low-income and/or minority groups(Reference Havas, Anliker and Damron25, Reference Herman, Harrison and Afifi26, Reference Campbell, Carbone and Honess-Morreale29, Reference Havas, Anliker and Greenberg30, Reference Fahrenwald, Atwood and Walker32), only two studies performed subgroup analyses(Reference Havas, Anliker and Damron25, Reference Havas, Anliker and Greenberg30). More of such analyses are needed to gain insight into the effectiveness of interventions for specific target groups. In accordance, future studies that conduct subgroup analysis for mothers v. non-mothers can give more information about the generalizability of interventions directed at women in general towards mothers in particular.
The main limitations of the present review study are the limited amount of eligible studies found, that information-gathering was restricted to printed material, and the low comparability of the studies. For instance, studies differed on the quality of design, sample size and outcome measurements. For this reason, the choice for descriptive analyses was made and no meta-analyses could be done. Nevertheless, these descriptive analyses can be useful for intervention practice since they could provide examples for future intervention development and implementation. For health promotion, it might be worthwhile also to include study designs other than randomized controlled trials (i.e. true experiments) in a review. The use of randomized controlled trials is not always appropriate to evaluate health promotion interventions(Reference Victora, Habicht and Bryce47); consequently reviewing interventions evaluated in other designs such as quasi-experiments can be more applicable for practice.
Conclusions
Although promoting weight control among new mothers is valuable in obesity prevention, the number of experimental intervention studies for promoting physical activity and healthy eating among new mothers is limited. Nevertheless, first recommendations useful for intervention practice can be set. Opportunities for increasing attendance of future interventions directed at mothers are, for example, embedding interventions into the mother and child routine, addressing the restricting role expectations for mothers, or motivating mothers for lifestyle change activities first. In addition, systematic development and multi-component interventions seem important to the success of these interventions. Moreover, chances for promoting physical activity among mothers are targeting the methods used regarding mother-specific barriers. Little evidence is found yet, however, to suggest the need for and recommend ways of targeting at new mothers for stimulating healthy eating. More research is required to substantiate the findings, especially with regard to the promotion of healthy eating. Process and subgroup analyses can contribute more insight into how interventions work and for which subgroups.
Acknowledgements
Source of funding: This study was funded by ZonMW, The Netherlands organization for health research and development. Conflicts of interest: There are no conflicts of interest. Author contributions: M.A.H. performed the literature search, data extraction, data analysis and drafted the manuscript. K.H. helped in determining the eligibility of articles in cases of doubt. K.S. conceived the study. K.H. and K.S. helped to draft the manuscript. All authors contributed to the design of the analysis. All authors read and approved the final manuscript.