Health Education Research, Vol. 17, No. 4, 435-449,
August 2002
© 2002 Oxford University Press
Differences in impact between a family- versus an individual-based tailored intervention to reduce fat intake
1 Department of Movement and Sport Sciences, Faculty of Medicine and Health Sciences, Ghent University, 9000 Ghent, Belgium, 2 Department of Health Education and Promotion, Maastricht University, 6211 Maastricht, The Netherlands and 3 Department of Behavioral Therapy and Counseling, Faculty of Psychology and Educational Sciences, Ghent University, 9000 Ghent, Belgium
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
The present study investigates the impact of a tailored nutrition intervention on the reduction of fat intake and on psychosocial determinants of fat intake. Furthermore, differences in impact between a family-based intervention (tailoring two family members, one adult and one child, simultaneously) and an individual-based intervention (tailoring one family member, one adult or one child) were studied. Analyses were conducted among 180 respondents, comparing 44 adolescents in the family condition with 50 adolescents in the individual condition and 44 parents in the family condition with 40 parents in the individual condition. Respondents in both conditions reported positive reactions towards the tailored fat feedback letters. Tailored fat feedback resulted in significantly more positive psychosocial determinants of fat intake and, among respondents with high fat intake at baseline, in a significant decrease in percent energy from fat. Parents in the family-based intervention group reported higher social support scores at post-test. No differences in post-test fat intake were found between the two study conditions. It is concluded that the results further illustrate the potential of tailored fat feedback, but the results do not provide evidence for superiority of family-based tailoring above individual-based tailored interventions for fat. Further research may be aimed at investigating the impact of comprehensive tailored family interventions, in which more than two family members participate.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Epidemiological evidence exists for the relationship between fat intake and the prevention of cardiovascular disease (Willet, 1994; Ascherio et al., 1996
). In addition, recent studies emphasize that the reduction in fat intake has to go together with a simultaneous reduction of energy intake (Hu et al., 1997; Parks, 2001; Raeini-Sarjaz et al., 2001). Despite the general awareness among the population of risks associated with a high fat diet, the prevalence of high fat intake is still very high. In Belgium, 60–70% of the adult population consume more than 35% energy from fat, the upper level of recommended fat intake. Only 11–16% reach the prevailing desirable level of 30% energy from fat (De Bourdeaudhuij and Van Oost, 2000). One major problem in nutrition education aiming at decreasing fat intake is that large proportions of populations have misconceptions about personal dietary intake levels and especially about fat intake (Lloyd et al., 1993; Brug et al., 1997; Kearny et al., 1997; Lechner et al., 1998; De Bourdeaudhuij and Brug, 2000; De Bourdeaudhuij and Van Oost, 2000). Since awareness of personal risk behavior has been identified as an important prerequisite for behavior change motivation (Weinstein, 1988), people who think they already eat a low fat diet will not make efforts to decrease their fat intake. In previous studies it was found that despite high levels of fat intake, respondents did not report an increase in awareness of personal fat intake after reading standardized, general fat feedback letters. They were even more convinced that their fat intake was low already, resulting in lower perceived personal fat intake at post-test (De Bourdeaudhuij and Brug, 2000). Fat reduction interventions may become more effective when these interventions give personal feedback on fat intake, together with personalized information about important fat sources and suggestions on how to cut back on fat. In the past few years, tailored health promotion materials intended to reach one specific person, based on characteristics that are unique to that person (De Vries and Brug, 1999; Kreuter et al., 2000), have been used especially in smoking cessation interventions (Dijkstra et al., 1999; Strecher, 1999), but also for dietary change (Brug et al., 1996, 1999). Brug et al. reviewed eight studies on tailored nutrition interventions for primary prevention of chronic disease and concluded that, especially for fat reduction, tailored nutrition interventions were more effective than general nutrition education (Brug et al., 1999). As many individually different misconceptions exist about fat content of foods and possible low fat alternatives, tailored fat feedback can be considered to be high in potential and may be the most promising approach for intervening effectively.
To date, only one study has investigated tailored nutrition interventions focusing on more than one family member simultaneously [(De Bourdeaudhuij and Brug, 2000), explained in the next paragraph]. In all other fat feedback studies only one individual in the family was involved (Brug et al., 1996, 1999). However, it has been argued that for dietary behavior the family context cannot be ignored (Nader et al., 1989, 1992). Significant family resemblance in eating behavior, food choice, fat and caloric intake has been found (Laskarzewski et al., 1980; Eastwood et al., 1982; Lee and Kolonel, 1982; Stafleu et al., 1994). Moreover, individual changes in eating behavior will almost inevitably affect other family members. If one individual in the family wants to change his/her fat intake, other family members may offer resistance or may have to make dietary changes too. Previous studies showed that one member of the family is seldom strong enough to influence family food choices. Mothers, in particular, were found to prefer family peace and harmony above conflicts about (healthy) food. From this it was expected that nutrition education interventions could be more successful if several family members at the same time decide to change their eating behavior in a more healthy way (De Bourdeaudhuij, 1997a–c; De Bourdeaudhuij and Van Oost, 1997, 1998a, b). De Bourdeaudhuij and Van Oost investigated the contribution of personal and family determinants in explaining variance in dietary behavior in a sample of 104 family dyads consisting of an adolescent and his/her parent (De Bourdeaudhuij and Van Oost, 2000). Results revealed that family determinants explained a maximum of 10% additional variance in dietary behavior over and above personal determinants. However, this study did not provide information about possible additional effects of family-based interventions on dietary behavior change as only actual fat intake and determinants were measured cross-sectionally. The contribution that family interventions could have over and above person-oriented nutrition education effects in dietary behavior change was considered to be an important topic for further research.
In a previous family intervention study (De Bourdeaudhuij and Brug, 2000), the impact of tailored versus standardized nutrition education on fat intake and on psychosocial determinants of fat intake in families was investigated, in a randomized controlled dietary feedback study. Thirty-five family quartets (both parents and two adolescents, all healthy individuals, n =140) were required as units of intervention. Tailored (experimental group, 18 quartets) or standardized (control group, 17 quartets) fat feedback letters were directed at each family member simultaneously at home. The tailored intervention was found to be more effective than the non-tailored intervention in reducing fat intake when all family members were included. However, follow-up analyses revealed that mothers only profited from the tailored intervention. For fathers and adolescents, both interventions resulted in a significant decrease in fat scores. Further, tailored feedback resulted in stronger awareness of personal fat intake and awareness of fat intake of family members. This previous study showed that differences in fat reduction between family members receiving general or tailored nutrition education letters were smaller than expected. This previous study did not answer the question whether family-based tailored interventions are more effective than tailored interventions focusing on a single person. This was investigated in the present study.
In the present study, the differences in impact between a family- and an individual-based tailored nutrition education programme focusing at fat reduction were studied. The first aim was to investigate changes in fat intake and in certain psychosocial determinants of fat intake (attitude, self-efficacy, social support, awareness, family perception, friends perception and intention) over time. In line with previous research, it was hypothesized that tailored fat feedback in the family condition as well as in the individual condition would result in a decrease in percent energy from fat and an increase in psychosocial determinants related to eating a low fat diet.
The second aim was to study whether tailored fat feedback is more effective if two family members (family condition) participate in the tailored intervention compared with a tailored intervention directed to single family members (individual condition). Differences between parents in the individual versus the family condition and differences between adolescents in the individual versus the family condition were studied. It was hypothesized that the decrease in percent energy from fat would be larger in adolescents and parents in the family condition compared with adolescents and parents in the individual condition. It was further expected that the increase in determinants of eating a low fat diet would be larger in the family condition compared to the individual condition. Earlier studies have shown that changes in transtheoretical stages of change (Prochaska et al., 1983) can be used as intermediate impact indicators in evaluation of tailored nutrition education interventions (Brug and Van Assema, 2000). Therefore, differences between the two study conditions in stage transitions were also studied.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Sample and procedure
The subjects participating in this study were recruited from 52 classes, with students between 15 and 18 years old, out of two secondary schools. The classes were randomized in three groups before recruitment. One group was for recruitment of family condition participants, one group for recruitment of adolescents for the individual condition and one group for recruitment of parents for the individual condition. For practical reasons it was not possible to recruit respondents for both conditions from the same class, leading to the use of a quasiexperimental design. We ensured inclusion of more classes for recruitment for the family condition since we expected a lower response for this group. Fifty-five adolescents were willing to participate together with one parent (n = 110), out of 19 classes (`family condition', 15% response). Seventy-one adolescents agreed to take part in the study alone, out of 16 classes in which only adolescents were asked to participate, constituting the `individual adolescent condition' (22% response). Forty-seven parents participated out of 17 classes in which the invitation to take part was only directed to the parents of the adolescents, constituting the `individual parent condition' (14% response). A letter explaining the purpose and procedure of the study was provided together with an informed consent form for participating parents and adolescents. In addition, informed consent from the parents was required for all adolescents participating in the study. Participants were only aware of their own study condition, not of the other conditions.
At baseline each participant received a questionnaire designed to obtain information on the psychosocial determinants of fat intake together with a food-frequency questionnaire to measure energy percentages of fat. After completion, the questionnaires were sent to the laboratory in a prepaid envelope. Six weeks later, all participants were mailed nutrition education letters to their home addresses, tailored to personal fat intake levels, motivation to reduce fat intake, awareness of personal fat intake, and attitudes and self-efficacy expectations related to fat reduction. Feedback letters were not computer-tailored, but tailored manually. Four weeks after the subjects had received their feedback letters, they were asked to complete and return the post-test questionnaires. Eleven families (n = 22, 20%) dropped out from the family condition, mainly because only the adolescent or the parent was willing to complete the questionnaire a second time. Twenty-one subjects (30%) dropped out of the individual adolescent condition and five subjects (11%) dropped out of the individual parent condition. All analyses were conducted among the 180 participants completing both questionnaires, which is 79% of the subjects who completed the baseline screening. More females participated in the study, 60% of the adolescents in the individual condition and 75% in the family condition were females (
2 = 2.4, NS); 74% of the parents in the individual condition and 82% in the family condition were mothers (2 = 0.8, NS).
Questionnaires
The measures for the independent variables were derived from Brug et al. (Brug et al., 1997) and De Bourdeaudhuij and Brug (De Bourdeaudhuij and Brug, 2000), based on Operant and Social Learning Theories (Bandura, 1996), and on socialpsychological theories such as the Theory of Planned Behavior (Ajzen and Madden, 1986) including attitudes, social influences and self-efficacy expectations. Based on research on determinants of motivation to reduce fat consumption (Brug et al., 1994; De Bourdeaudhuij and Brug, 2000), awareness of personal intake, and awareness of other family members' and friends' fat intake were also included as relevant outcome measures.
Attitudes were assessed by using general as well as more specific measures (Meertens et al., 2000). As a general attitude measure, respondents were asked to what degree they thought that eating fat is bad or good, unpleasant or pleasant and nasty or tasty. These items were answered on seven-point scales. In addition, specific components of attitudes or `beliefs' were included, by asking about perceived benefits and barriers related to eating a low fat diet. These beliefs included health consequences, weight loss, time constraints, lack of knowledge or skills, inner urge, etc., measured on five-point scales ranging from 1 (not at all a benefit/barrier) to 5 (very strong benefit/barrier). Specific attitude measures were included because in the tailoring procedure specific beliefs of the respondents were used to construct the personalized feedback letters. An attitude scale was computed, consisting of 10 items, recoded to a 10-point scale ranging from 1 (a very negative attitude toward a low fat diet) to 10 (a very positive attitude towards a low fat diet; Cronbach's 
= 0.75). Perceived social support was measured by asking respondents to what degree they expected social support from family members, relatives and friends if they tried to eat less fat. Items included talking about fat intake, eating a low fat diet together, reminding or encouraging low fat intake and criticizing or making fun of low fat diets, all answered on five-point scales. For parents, significant others were children, relatives and friends; for adolescents, they were parents, brothers/sisters, relatives and friends. The social support scale consisted of 16 items, recoded to a 10-point scale ranging from very low social support (1) to very high social support (10) for eating a low fat diet (Cronbach's = 0.81). Self-efficacy was measured by asking respondents how difficult or easy they thought it was to eat less fat and how confident they were to be able to maintain a low fat diet in various situations (when having a lot to do, when being alone, when having to buy other products, when really wanting to eat less fat). The final self-efficacy scale consisted of nine items, recoded to a 10-point scale ranging from a very low self-efficacy (1) to a very high self-efficacy (10; Cronbach's = 0.85). Three items reflecting awareness of personal fat intake were included: self-assessed fat intake was measured by asking respondents to evaluate their fat intake on a seven-point scale ranging from very low (1) to very high (7); perceptions of the amount of fat that family and friends eat (family perception; friend perception) were assessed by asking respondents to evaluate the fat intake of their family members and their friends on a seven-point scale ranging from very low (1) to very high (7). One question was included to measure the subjects' intentions to eat a low fat diet, ranging from no intention (1) to very strong intention (5). Awareness and intention items were also recoded into 10-point scales to make comparison between variables easier. The algorithm that was used to allocate subjects to the different stages of change was very similar to that used in earlier studies (Glanz et al., 1994; Brug et al., 1997) and was based on the following questionnaire items:
- I do not try to decrease fat intake at the moment and I do not intend to do this in the future (Precontemplation).
- I intend to eat less fat within 6 months (Contemplation).
- I intend to eat less fat within 30 days (Preparation).
- I tried to eat less fat in the past month(s) (Action).
- I do not try to eat less fat because this is not necessary for me as I already eat little fat (Maintenance).
In the post-test questionnaire the same variables were assessed. Further, the participants were also asked about their reactions to the nutrition information letter that they received and whether the nutrition information letter had resulted in changes in opinions about their diets, in their intentions and in their dietary behavior (see Table I for the questionnaire items). These questions, also used in previous studies (Brug et al., 1996, 1998; De Bourdeaudhuij and Brug, 2000), are not considered to be valid change measures, but rather a reflection of a higher awareness in subjects about the necessity of behavior change after receiving tailored feedback letters.
|
To assess eating patterns, registered dieticians adapted a 56-item food-frequency questionnaire validated in the Netherlands (Feunekes et al., 1993
) for our Flemish population. Some minor changes were made, adding some typical Flemish food. Respondents were asked to report the frequency of consumption of each food item per day, per week or per month with a reference period of the past 4 weeks. From this food-frequency questionnaire, intake of total fat was computed in percentages of total energy intake. For this purpose, a computer program was written by Aben et al. (Aben et al., 1993) in which frequency of consumption for each food item was first calculated per week and then multiplied by a weighted fat factor based on the fat content of food items as reported in the Dutch food composition tables (NEVO, 1987). Standardized portion sizes were used in computing fat scores. Validity and reliability of this adapted food-frequency questionnaire was investigated in a small student sample (n = 45 for validity, n = 90 for reliability). Test–re-test reliability using an interval of 2 weeks showed a Cronbach's coefficient of 0.83 on total fat intake. Pearson's product-moment correlation coefficient was used to evaluate if the linear association between the results of the food-frequency questionnaire and a 7-day food diary showed appropriate validity (r = 0.78). Although a correlation of 0.78 means that only 61% of the variance in fat consumption is explained by the questionnaire, we believe that this questionnaire is the most valid and reliable to be used in the Flemish part of Belgium, and is comparable to or better than other food-frequency questionnaires that are often used in food consumption research. Ongoing research tries to further improve food frequency questionnaires. In the mean time, some prudence in interpreting results is warranted.
Intervention
The intervention was targeted at reducing fat intake. All subjects received feedback letters including messages based on their answers to the screening questions. The tailoring procedure and messages developed by Brug et al. (Brug et al., 1998) were used, which was analogous with the previous family study (De Bourdeaudhuij and Brug, 2000). In both family studies, the letters were constructed manually, as we first wanted to establish the potential of family-based tailoring before investing in a computer program. Respondents received feedback about their fat intake as well as about their attitudes, perceived support and self-efficacy toward fat reduction. The feedback messages about fat intake included respondents' actual fat consumption expressed in percent energy from fat, and a comparison of this percentage with the Flemish recommendations and with the mean scores of the other adolescents, mothers or fathers participating in this study. Strictly speaking, the Flemish dietary guidelines recommend a percent energy from fat not higher than 30%. In the present study, the recommendations were translated to the respondents as `preferably 30% and not more than 35% energy from fat'. This was inspired by previous studies in which mean scores of about 40% energy from fat were found. We believed that giving very strict normative feedback would result in discouragement of participants questioning the feasibility of the dietary goals and abandoning the intention of behavior change. The messages also included a comparison between the actual consumption and the way participants rated their own consumption. A graph was included to help visualize personal fat intake levels as compared to the recommendation as well as to the mean fat intake of comparable others (adolescents, mothers or fathers). Messages addressed different important dietary fat sources in the Flemish diet, for which low fat alternatives for high fat choices were suggested.
Respondents with low self-efficacy expectations were told how they could deal with high-risk situations such as the presence of high fat foods in the home (avoiding such situations by making low fat shopping lists), being alone (suggestions on low fat alternatives for single dinners) and lack of time for food preparation (recipes for easy to cook low fat alternatives). Finally, all subjects who reported a positive intention to reduce their fat intake were advised to change these plans into direct action in the week(s) to come, preferably in the next week. No attempts were made to encourage positive family influence, e.g. by including recommendations to share the results with other family members.
Analyses
A 2-test was used to look at condition-specific differences in drop-out (family condition, individual adolescent condition and individual parent condition). t-tests and 2-tests were executed to explore differences in mean fat scores, sex and education, between subjects who dropped out and subjects who completed the baseline and the post-test questionnaires.
For all statistical tests, parents' and adolescents' results were analyzed separately, investigating differences between parents in the individual versus the family condition, and differences between adolescents in the individual versus the family condition. 2-tests and t-tests were conducted to study differences in respondents' reactions to the feedback letters in both conditions.
Mean energy percentages fat and mean determinant scores at baseline and at post-test were computed for adolescents and parents. The mean fat scores revealed that for both adolescents and parents fat intake was relatively low at baseline when compared to our previous family study (means present study: adolescents = 37.8% and parents = 34.6%; means previous study: adolescents = 40.2% and parents = 38.0%). Since only subjects with energy percentages higher than 35% were encouraged to change, while those with energy percentages between 30 and 35% were encouraged to maintain their relatively favorable diets, only suggesting that some further decrease in fat could give additional health benefits, we decided to analyze both the whole research sample and the subsample with fat intake above 35% of total energy intake (higher risk sample) on time and condition effects.
Repeated-measures ANOVAs with baseline and post-test energy percentages of fat as dependent variables were executed to test for significant changes in fat intake over time. Repeated-measures MANOVAs with seven baseline and post-test determinants' scores (attitude, self-efficacy, social support, awareness, family perception, friend perception, intention) as dependent variables, and univariate follow-up tests were used to look at general changes in determinants of fat intake over time. All analyses were executed for the total sample, and the total at-risk sample, as well as for the subsample of adolescents and parents in both samples. Finally, in order to test for differences in intervention effects between the family condition and the individual condition, analyses of covariance were conducted, testing for significant differences at post-test in percent energy from fat, and psychosocial determinants of fat intake, while adjusting for baseline levels.
One-tailed tests were executed to examine the hypotheses. For all analyses, P < 0.05 was considered to be significant.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Drop-out analysis
A trend for significance was found for dropout within the four conditions [
2(3) = 6.25, P = 0.10], revealing a higher drop-out rate in the individual adolescent condition (29%) and less drop-outs in the individual parent condition (12%) compared to drop-out in the family condition (25%).
No significant difference was found between drop-outs (mean = 37.1%) and the remaining research sample (mean = 36.3%) in percent energy from fat at baseline [t(224) = 0.73, NS]. No significant differences were found for psychosocial determinants such as attitude [t(226) = –1.15, NS], self-efficacy [t(226) = –1.59, NS], social support [t(226) = 0.98, NS], awareness [t(226) = 0.46, NS], family perception [t(226) = 1.78, NS], friend perception [t(226) = 1.89, NS] and intention [t(226) = 0.55, NS]. Further, no differences were found between drop-outs and the research sample in sex [(1) = 1.61, NS] or education [(5) = 7.30, NS].
Participants' reactions to the tailored letters
Table I
shows the respondents' reactions to the feedback letters. Adolescents and parents in both conditions rated the information in their feedback letters as interesting, credible and comprehensible. They also considered the letters to be personally relevant for them and contained new information. No significant differences were found between both conditions for adolescents or for parents.
Almost all respondents read the feedback letter completely and a large proportion saved the nutrition information letter. About 60% of the adolescents and over 40% of the parents discussed the letter with others. 2 analyses did not reveal differences between conditions for adolescents or parents in using the letters.
Over 60% of the adolescents who received the tailored intervention reported that they changed their opinion about their diet and intended to change their diet in the future. Percentages for parents were slightly lower, with over 50% reporting a change in opinion about their diet and over 60% a positive intention. About half of the adolescents and the parents also reported they actually changed their diet as a result of the nutrition information letter. Again, no condition differences were found for parents.
Effects of tailored nutrition education on fat intake and determinants of fat intake
Repeated-measures analyses of variance did not reveal a significant reduction in percent energy from fat between baseline and post-test for the entire study population [F(1,177) = 1.87, NS], nor for adolescents [F(1,93) = 2.42, NS] or parents [F(1,83) = 0.13, NS] separately. In contrast, multivariate analyses including all measured determinants of fat intake revealed significantly more positive determinants for the whole sample [F(7,159) =12.98, P < 0.001], as well as for adolescents [F(7, 85) = 6.22, P < 0.001] and parents [F(7,67) = 6.75, P < 0.001] separately. Univariate tests showed that in all three groups a significant positive evolution in determinants was only found for social support [F(1,165) = 7.89, P < 0.005 for whole sample; F(1,91) = 4.36, P < 0.05 for adolescents; F(1,73) = 3.48, P < 0.05 for parents] and for intention [F(1,165) = 83.72, P < 0.001 for whole sample; F(1,91) = 40.70, P < 0.001 for adolescents; F(1,73) = 43.91, P < 0.001 for parents].
Among subjects with fat consumption levels above the Belgian upper level of recommended intake of fat (35% energy from fat) at baseline, analyses did reveal a significant reduction in percent energy from fat [F(1,95) = 29.13, P < 0.001] between baseline and post-test. A similar intervention effect was found in the separate analyses among adolescents [F(1,55) = 15.27, P < 0.001] and parents [F(1,39) = 13.76, P < 0.001] in this high-risk group (means: pre = 42.5%/post = 38.9; pre = 40.0%/post = 36.2% respectively). Similarly, a significant increase in determinants of fat intake was found for the whole at-risk group [F(7, 82) = 5.89, P < 0.001] as well as for adolescents [F(7,48) = 3.07, P < 0.01] and parents [F(7,27) = 3.58, P < 0.01] separately. Univariate tests showed that for the whole at-risk group, a significant positive evolution was only found for social support [F(1,88) = 3.99, P < 0.05] and for intention [F(1,88) = 37.76, P < 0.001]. For the adolescent at-risk group, only the evolution in intention showed significance [F(1,54) = 19.24, P < 0.001]. For the parent at-risk group, intention [F(1,33) = 19.84, P < 0.001] as well as family perception [F(1,33) = 5.59, P < 0.05] yielded significance.
In general the variance of percent energy from fat explained by the psychosocial determinants is rather low, explaining 16% of the variance at baseline and 13% at post-test for the total research group, and 16% at baseline and 19% at post-test for the `high-risk' group.
Differences in impact between individual and family condition
Table II summarizes the mean fat and determinants scores at baseline and post-test for the total sample of adolescents and parents in the individual and family condition. Analyses of variance on post-test scores including baseline level as covariate did not reveal any significant difference between the individual and family condition for adolescents (all Fs < 2.26, NS). For parents, a significant difference between the individual and family condition was found in social support [F(1,81) = 4.61, P < 0.05]. Parents in the family condition reported more social support at post-test compared with parents in the individual condition.
|
Table III
shows the mean fat and determinants scores at baseline and post-test for respondents with total fat above 35% of energy at baseline. For adolescents in the at-risk group, no significant differences were found. For parents in the at-risk group, again social support was found to be significantly different at post-test between both conditions [F(1,37) = 5.06, P < 0.05] with parents in the family condition at-risk group reporting more social support at post-test.
|
Progression through stages of change
Table IV
shows the number of subjects categorizing themselves within the different stages of change at baseline and at post-test. As the frequencies per category are too small to execute reliable analyses, only general trends will be given. After the intervention, the number of people classifying themselves in the action phase is almost doubled, while less people report to be in contemplation, preparation and maintenance. This trend is clear for adolescents and parents in both conditions.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The aims of the present study were to determine whether changes in fat intake and in psychosocial determinants could be found as a result of tailored nutrition education aiming at reducing fat, and whether this intervention is more effective when two family members participate in the intervention study compared with an intervention directed to an individual family member. The first hypothesis was supported for the at-risk subgroup, but only partly for the total research sample. The tailored fat feedback resulted in a decrease in percent energy from fat and an increase in psychosocial determinants related to eating a low fat diet in the subgroup of respondents with percentage energy from fat above 35. In the total research sample, with a mean baseline percent energy from fat of 36.3, only determinants towards eating a low fat diet increased significantly, but fat intake did not change. This may come as no surprise since tailored feedback especially encourages high fat consumers to re-evaluate their opinions about their fat intake and to make dietary changes. Low fat consumers were mainly encouraged to maintain their present intake levels.
As study a no-intervention control condition was not included in the present, it does not provide hard evidence for a positive effect of the tailored fat feedback in the at-risk group attributable to the intervention. However, our results are in line with previous studies documenting the additional impact on fat reduction of tailored feedback as compared to general nutrition information (Brug et al., 1999). A previous study in which tailored nutrition education focusing on fat reduction was used within the family, provided evidence that the impact differed across family members. Mothers especially profited from the tailored interventions, while for fathers and adolescents, the tailored as well as the non-tailored intervention resulted in a significant decrease in fat scores. Differences between the adolescent sample and the parents sample (78% mothers) in the present study were not found.
A clear increase in the intention to eat a low fat diet in the future was the most important evolution in determinants of fat intake found in all subgroups. Although one can question the strength of the relationship between intention and behavior, this can be considered to be a promising result (Godin and Kok, 1996). The major problem in nutrition education aiming at decreasing fat intake is that people think they already eat a low fat diet, which means that they will not make efforts to decrease their fat intake. The increase in intention we found between baseline and post-test after tailored fat feedback can be considered to be a sign of increased awareness. This increase in awareness can also be derived from the shifts in classification along the stages of change from baseline to post-test. A number of subjects classifying themselves as being in maintenance at baseline were probably precontemplators, being unaware of their `risk behavior'. The shifts found in our results show that people become aware of their fat intake and try to change their behavior. However, awareness must be considered to be a necessary, but not sufficient, condition for behavior change.
Apart from intention, very few significant effects were found for the other determinants. This is in line with previous studies in which this was explained by the fact these determinants reflect what people think they eat and not what they actually eat (Brug et al., 1994; Lechner et al., 1998; De Bourdeaudhuij and Brug, 2000). An additional explanation could be that the selection bias in the present study favoring respondents with lower fat intake levels is also responsible for the more positive attitudes, self-efficacy, awareness, etc., of the respondents in the present study.
The program evaluation data showed that the tailored fat feedback letters were read completely, very often saved and discussed with others by approximately half of the respondents. The information letter was perceived to be interesting, credible, comprehensible, personally relevant and containing new information. This is in line with earlier findings on tailored feedback (Brug et al., 1998; De Bourdeaudhuij and Brug, 2000). These positive evaluations suggest that respondents paid considerable attention to the letters leading to subjects' involvement and cognitive processing of the information (Brug et al., 1994, 1998; Skinner et al., 1994; Kreuter et al., 1999, 2000). No differences in appreciation of the feedback letters were found between the individual and the family condition for parents or adolescents.
In contrast to the hypothesis, the decrease in percent energy from fat was not found to be larger in adolescents and parents in the family condition as compared with the individual condition. This argues against the hypothesis that involvement of the family makes it easier to adhere to a low fat diet, both in a sample of moderate fat intake (whole study group) and in the at-risk sample. It was further expected that the increase in determinants towards eating a low fat diet would be larger in the family condition than in the individual condition. For parents, a significant difference in the expected direction between the individual and family condition was only found for social support. Considering all psychosocial determinants of fat intake, social support is the most obvious determinant to change as the result of a family intervention. Nevertheless, we expected that also attitudes, self-efficacy and intention would have benefited more from the family approach, since intervening at a family level may make dietary changes more attractive and easier to accomplish, leading to better motivation. As social support can be considered to be only one aspect of a range of factors influencing fat intake (Connor and Norman, 1996; Godin and Kok, 1996; Bennet and Murphy, 1997), one cannot expect that positive changes in social support alone result in changes in fat intake. The fact that an effect on social support was only found for parents and not for adolescents is in some way congruent with the higher tailoring impact found among mothers compared with adolescents in the previous family study (De Bourdeaudhuij and Brug, 2000) where it was argued that mothers are often more motivated to eat less fat, have more knowledge about the fat content of foods and have more positive attitudes toward low fat diets, but do not feel able to succeed in changing their children's and husband's food habits (Backett, 1992; Andersen et al., 1995; Bourdeaudhuij and Van Oost, 1998b).
A possible explanation for the small differences found between the family and the individual condition could be that, for practical and feasibility reasons, only two members of the same family and not the whole family participated in our study. From earlier experience with family-based research we learned that response rates tend to be extremely low (10%) when all family members are required to participate (De Bourdeaudhuij and Brug, 2000). However, participation of all family members may be a prerequisite for effective family-based interventions. Previous family studies investigating the decision-making power of each family member in food choices indeed showed that involvement of the entire family for the introduction and adoption of healthy eating is desirable (Charles and Kerr, 1988; Mennell et al., 1992; De Bourdeaudhuij and Van Oost, 1998b). Moreover, from respondents' characteristics and fat intake levels, it could be argued that especially highly motivated adolescents and parents (especially mothers) participated. The possible surplus value of targeting families in nutrition education probably lies in including the least motivated family members in the intervention, such as fathers and other children in the home, and not the most motivated members, such as the adolescents who volunteered or mothers. Previous efforts to include all family members in health education interventions experienced similar problems such as high non-attendance rates, drop-out and difficulty with recruitment, resulting in participation of the most motivated and the least unhealthy families (Perry et al., 1987; Baranowski et al., 1990). Indeed, in the present study mean fat intake was relatively low in comparison to earlier studies in Belgium measuring food consumption in the same population (De Bourdeaudhuij, 1997c; De Bourdeaudhuij and Brug, 2000; De Bourdeaudhuij and Van Oost, 2000). This is a clear indication for selection bias in the present study. Although we thought that the use of tailored nutrition education letters directed at each family member simultaneously at home would overcome these difficulties, the necessity to volunteer to complete screening questionnaires at baseline brings us to face the same shortcomings. In addition to the aforementioned argument that there may have been too little `family dimension' in the family condition, it could also be possible that family dimensions were unintentionally present in the individual condition. The finding that 68% of the adolescents in the individual condition reported to have discussed the nutrition information letter with others, compared to 54% of the adolescents in the family condition, suggests that the family and/or friends of the adolescents in the individual condition were also involved in some way.
There are several important limitations to note in interpreting this study. First, as noted before, a selection bias is probable, oversampling mothers and adolescents with lower fat intake levels, more positive attitudes towards eating a low fat diet and higher awareness levels of own fat intake. This is probably due to the quasiexperimental design and the recruitment strategy in which adolescents in classes were asked to volunteer. Generalization of the results to the Belgian population of families as a whole is therefore difficult. Secondly, as mentioned before, a true `family approach' in nutrition education must undoubtedly include as many family members as possible, especially the least-motivated ones, to be effective. Additionally, the present study only investigated the short-term impact of the interventions. It is possible that other results appear in the longer run. A further limitation of the present study is the use of self-reports to assess tailoring effects (Brug et al., 1999). More objective criteria like body mass index, cholesterol levels or other blood parameters would give more verifiable results. Further, due to the fact that the feedback letters were tailored manually and not computerized, a large time interval was present between the completion of the screening questionnaires and the receipt of the feedback letters by mail. This may be one of the reasons for the relatively high drop-out between baseline and post-test. Finally, this study only focused on reduction in fat intake. However, recent studies emphasized the importance of combining fat reduction with simultaneous reduction of energy intake (Hu et al., 1997; Parks, 2001; RaeiniSarjaz et al., 2001). Total energy intake was not measured in the present study and reduction of energy intake was not recommended in the intervention. Further research is needed to develop more comprehensive tailored feedback tools.
Despite these limitations, the present study has several implications for nutrition education practice and research. A tailored intervention aimed at fat reduction was shown to have an impact on fat intake in an at-risk sample. Although in the present study, no no-intervention control condition was included, this is very much in line with findings of previous studies which show that personal feedback on fat intake, together with personalized information about important fat sources, suggestions on how to cut back on fat and personal feedback on psychosocial factors, has considerable potential. Communicating this message, for example, in primary care could increase awareness in people with high levels of fat intake, being the first step in behavior change. As the concept of energy percentage of fat intake and total energy intake is somewhat abstract and not easy to record, computer tailoring will be a useful tool. The implementation of computerized assessment and feedback in nutrition education practice through medical doctors, social services, regional initiatives, etc., may be a valuable tool for the future. The present study generally failed to reveal a substantial surplus effect of tailoring two family members above individual tailoring. Further research may involve more than two family members to investigate the difference in impact between a comprehensive family intervention versus an individual-based intervention. In such studies special attention also has to be paid to the recruitment procedure to avoid a selection bias towards motivated and already healthy people and families.
|
Acknowledgments |
|---|
The authors like to thank Griet Landrie for her contribution in data gathering and tailoring feedback letters. This study was financially supported by the Ghent University and the Flemish Fund for Scientific Research.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Aben, D., Maas, I., van de Waal, M. and Feunekes, G. (1993) De Envetlijst, ontwikkeling en valideringvan een korte voedselfrequentievragenlijst naar energie en vet [The envetquestionnaire: development and validation of a short food frequency questionnaire to assess fat intake]. GGD regio GeldropValkenswaard, GGD Eindhoven en Vakgroep Humane Voeding, Landbouwuniversiteit Wageningen, Wageningen.
Ajzen, I. and Madden, T. J. (1986) Prediction of goal-directed behavior: attitudes, intentions and perceived behavioral control. Journal of Experimental Social Psychology, 22, 453–474.[ISI]
Anderson, A., Milburn, K. and Lean, M. (1995) Food and nutrition: helping the consumer understand. In Marshall, D. (ed.), Food Choice and the Consumer. Chapman & Hall, London, pp. 105–128.
Ascherio, A., Rimm, E., Giovannucci, E., Spiegelman, D., Stampfer, M. and Willett, W. (1996) Dietary fat and risk of coronary heart disease in men: cohort follow-up study in the United States. British Medical Journal, 13, 84–90.
Backett, K. C. (1992) The construction of health knowledge in middle class families. Health Education Research, 7, 497–507.
Bandura, A. (1996) Social Foundations of Thought and Action. A Social Cognitive Theory. Prentice-Hall, Englewood Cliffs, NJ.
Baranowski, T., Simons-Morton, B., Hooks, P. and Henske, J. (1990) A center based program for exercise change among Black-American families. Health Education Quarterly, 17, 179–196.[ISI][Medline]
Bennet, P. and Murphy, S. (1997) Psychology and Health Promotion. Open University Press, Buckingham.
Brug, J. and Van Assema, P. (2000) Differences in use and impact of computer-tailored dietary fat feedback according to stages of change and education. Appetite, 34, 285–293.[ISI][Medline]
Brug, J., Van Assema, P., Kok, G., Lenderink, T. and Glanz, K. (1994) Self-rated dietary fat intake: association with objective assessment of fat, psychosocial factors and intention to change. Journal of Nutrition Education, 26, 218–223.
Brug, J., Steenhuis, I., Van Assema, P. and De Vries, H. (1996) The impact of a computer-tailored nutrition education. Preventive Medicine, 25, 236–242.[ISI][Medline]
Brug, J., Hospers, H. and Kok, G. (1997) Differences in psychosocial factors and fat consumption between stages of change for fat reduction. Psychology and Health, 12, 719–727.[ISI]
Brug, J., Glanz, K., Van Assema, P., Kok, G. and Van Breukelen, G. (1998) The impact of computer-tailored feedback and iterative feedback on fat, fruit, and vegetable intake. Health Education and Behavior, 25, 517–531.[Abstract]
Brug, J., Campbell, M. and Van Assema, P. (1999) The application and impact of computer-generated personalized nutrition education: a review of the literature. Patient Education and Counseling, 36, 145–156.[ISI][Medline]
Campbell, M., DeVellis, B., Strecher, V., Ammerman, A., DeVellis, R. and Sandler, R. (1994) Improving dietary behavior: the effectiveness of tailored messages in primary health settings. American Journal of Public Health, 84, 783–787.
Charles, N. and Kerr, M. (1988) Women, Food and Families. University Press, Manchester.
Conner, M. and Norman, P. (1996) Predicting Health Behavior. Open University Press, Buckingham.
De Bourdeaudhuij, I. (1997a) Family food rules and healthy eating in adolescents. Journal of Health Psychology, 2, 45–56.[Abstract]
De Bourdeaudhuij, I. (1997b) Perceived family members' influence on introducing healthy food into the family. Health Education Research, 12, 77–90.
De Bourdeaudhuij, I. (1997c) Family variables and health behaviours in adolescents, with focus on nutrition. Dissertation, Ghent University, Belgium.
De Bourdeaudhuij, I. and Brug, J. (2000) Tailoring dietary feedback to reduce fat intake: an intervention at the family level. Health Education Research, 15, 449–462.
De Bourdeaudhuij, I. and Van Oost, P. (1997) Gezinsdynamiek en voedingsgedrag: een onderzoek met vragenlijsten en groepsinterviews [Family dynamics and food: a study using questionnaires and focus group interviews]. Gedrag and Gezondheid, 25, 62–73.
De Bourdeaudhuij, I. and Van Oost, P. (1998a) Family characteristics and positive health behaviors in adolescents and families. Psychology and Health, 13, 785–803.
De Bourdeaudhuij, I. and Van Oost, P. (1998b) Family members' influence on decision making about food: differences in perception and relationship with healthy eating. American Journal of Health Promotion, 13, 73–81.[ISI][Medline]
De Bourdeaudhuij, I. and Van Oost, P. (2000) Personal and family determinants of dietary behavior in adolescents and their parents. Psychology and Health, 15, 751–770.[ISI]
De Vries, H. and Brug, J. (1999) Computer-tailored interventions motivating people to adopt health promoting behaviors: introduction to a new approach. Patient Education and Counseling, 36, 99–105.[ISI][Medline]
Dijkstra, A., De Vries, H. and Roijackers, J. (1999) Targeting smokers with low readiness to change with tailored and non-tailored self-help materials. Preventive Medicine, 28, 203–211.[ISI][Medline]
Eastwood, M. A., Brydon, W. G., Smith, D. M. and Smith, J. H. (1982) A study of diet serum lipids, and fecal constituents in spouses. American Journal of Clinical Nutrition, 36, 290–293.
Feunekes, G. I., Van Staveren, W. A., De Vries, J. H., Burema, J. and Hautvast, J. G. (1993) Relative and biomarker-based validity of a food-frequency questionnaire estimating intake of fats and cholesterol. American Journal of Clinical Nutrition, 58, 489–496.
Glanz, K., Patterson, R. E., Kristal, A. R., DiClemente, C. C., Heimendinger, J., Linnan, L. and McLerran, D. F. (1994). Stages of change in adopting healthy diets: fat, fiber, and correlates of nutrient intake. Health Education Quarterly, 21, 499–519.[ISI][Medline]
Godin, G. and Kok, G. (1996) The theory of planned behavior: a review of its applications to health-related behaviors. American Journal of Health Promotion, 11, 87–98.[ISI][Medline]
Hu, F. B., Stampfer, M. J., Manson, J. E., Rimm, E., Colditz, G. A., Rosner, B. A., Hennekens, C. H. and Willett, W. C. (1997) Dietary fat intake and the risk of coronary heart disease in women. New England Journal of Medicine, 337, 1491–1499.
Kearny, M., Gibney, M., Martinez, J., de Almeida, M., Friebe, D., Zunft, H., Widhalm, K. and Kearny, J. (1997) Perceived need to alter eating habits among representative samples of adults from all member states of the European Union. European Journal of Clinical Nutrition, 51, S30–S35.
Kreuter, M., Bull., F., Clark, E. and Oswald, D. (1999) Understanding how people process health information: a comparison of tailored and non-tailored weight-loss materials. Health Psychology, 18, 487–494.[ISI][Medline]
Kreuter, M., Farrell, D., Olevitch, L. and Brennan, L. (2000) Tailoring Health Messages: Customizing Communication with Computer Technology. Lawrence Erlbaum, London.
Laskarzewski, P., Porrison, J., Khoury, P., Kelly, K., Glatfelter, L., Larsen, R. and Glueck, D. (1980) Parent–child nutrient intake interrelationships in school children ages 6 to 19: the Princeton School District Study. American Journal of Clinical Nutrition, 33, 2350–2355.
Lechner, L., Brug, J., De Vries, H., Van Assema, P. and Mudde, A. (1998) Stages of change for fruit, vegetable and fat intake: consequences of misconception. Health Education Research, 13, 113.
Lee, J. and Kolonel, L. N. (1982) Nutrient intakes of husbands and wives: implications for epidemiological research. American Journal of Epidemiology, 115, 515–525.
Lloyd, H., Paisley, C. and Meta, D. (1993) Changing to a low fat diet: attitudes and beliefs of UK consumers. European Journal of Clinical Nutrition, 47, 361–373.[ISI][Medline]
Meertens, R., Schaalma, H., Brug, J. and de Vries, N. (2000) Determinanten van gedrag [Behavioral determinants]. In Brug, J., Schaalma, H., Kok, G., Meertens, R. M. and Van der Molen, H. T. (eds), Gezondheidsvoorlichting en Gedragsverandering [Health Promotion and Behavior Change]. Van Gorcum, Assen, pp. 55–72.
Mennell, S., Murcott, A. and van Otterloo, A. (1992) The Sociology of Food: Eating, Diet and Culture. Sage, London.
Nader, P., Sallis, J., Patterson, T. and Abramson, I. (1989) A family approach to cardiovascular risk reduction: results from the San Diego Family Health Project. Health Education Quarterly, 16, 229244.
Nader, P., Sallis, J., Abramson, I. and Broyles, S. (1992) Familybased cardiovascular risk reduction education among Mexican and AngloAmericans. Family and Community Health, 15, 5774.
NEVO (1987) Nederlands voedingsstoffenbestand 1986/1987. Stichting Nederlands Voedingsbestand, Zeist.
Parks, E. J. (2001) A targeted goal for energy-restricted diets in the management of coronary risk? American Journal of Clinical Nutrition, 73, 147–148.
Perry, C., Crockett, S. and Pirie, P. (1987) Influencing parental health behavior: implications of community assessments. Health Education, 18, 68–77.
Prochaska, J. and DiClemente, J. (1983) Stages and processes of self-change of smoking: toward an integrative model of change. Journal of Consulting and Clinical Psychology, 51, 390–395.[ISI][Medline]
RaeiniSarjaz, M., Vanstone, C. A., Papamandjaris, A. A., Wykes, L. J. and Jones, P. (2001) Comparison of the effect of dietary fat restriction with that of energy restriction on human lipid metabolism. American Journal of Clinical Nutrition, 73, 262–267.
Skinner, C., Strecher, V. and Hospers, H. (1994) Physicians' recommendations for mammography: do tailored messages make a difference? American Journal of Public Health, 84, 43–49.
Stafleu, A. (1994) Family resemblance in fat intake, nutrition attitudes and beliefs: a study among three generations of women. Doctoral thesis, Agricultural and Environmental Science, Landbouwuniversiteit Wageningen, Wageningen..
Strecher, V. (1999) Computer-tailored smoking cessation materials: a review and discussion. Patient Education and Counseling, 36, 107–117.[ISI][Medline]
Weinstein, N. (1988) The precaution adoption process. Health Psychology, 7, 355–386.[ISI][Medline]
Willett, W. (1994) Diet and health: what should we eat? Science, 264, 532–537.
Received on March 12, 2001; accepted on August 31, 2001
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Goulet, B. Lamarche, and S. Lemieux Factors influencing the dietary response to a nutritional intervention promoting the Mediterranean food pattern in healthy women from the Quebec City metropolitan area Health Educ. Res., October 1, 2007; 22(5): 718 - 726. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Vandelanotte, I. De Bourdeaudhuij, and J. Brug Acceptability and feasibility of an interactive computer-tailored fat intake intervention in Belgium Health Promot. Int., December 1, 2004; 19(4): 463 - 470. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Brug, A. Oenema, and M. Campbell Past, present, and future of computer-tailored nutrition education Am. J. Clinical Nutrition, April 1, 2003; 77(4): 1028S - 1034. [Abstract] [Full Text] [PDF]
|
|
| ||||||||