The roots of common obesity lie in an imbalance between energy intake with food and energy expenditure due to physical activity. Although an improvement of both is strongly advocated by the medical profession, the media, and even politicians, the successes in weight reduction are unsatisfactory, because obesity continues to spread in affluent societies [
A basic and widely used approach to weight reduction is calorie restriction. However, experience shows that—in free-living populations—adherence to this measure is difficult, that the resulting weight losses are often suboptimal, and that going off the diet is frequently followed by a rebound weight gain [
We, therefore, proposed to evaluate the effects of alternative weight-reduction strategies, implemented on top of a fat calorie-restriction diet [
The family approach reflects that childhood obesity in primary school-aged children cannot be solved by themselves as they are heavily dependent on their parents [
All study participants were equally advised to follow a fat-calorie restriction diet and to reduce their daily calorie input by 500 kcal. On top of this basic measure, three additional strategies for weight loss were added and compared in this randomized study following a three-factorial design.
The first strategy concerned nutrition and consisted—in addition to the basic fat-calorie restriction—of the preference of carbohydrate which hardly raise insulin (“dual diet”). The second strategy was a financial incentive, with payments for each kilogram of weight loss in the parents and for each reduction in body mass index standard deviation score (BMI-SDS) in the children. The third strategy was the use of telemetric devices consisting of weighing scales and accelerometers issued to the participants. The data of both were transmitted regularly and enabled us to respond with weekly letters for information and motivation. For each of these weight-reduction strategies, there was a control group treated in the same way except for the additional strategy.
The participating families were recruited by means of newspaper advertisements in the area around the German city of Magdeburg. The children had to be older than 7 years to ensure that they were able to read, and younger than 13 to minimize interferences due to puberty. 177 families responded by telephone and received a letter describing the aim and character of the study. 110 families then decided to participate and were invited to the first of four meetings with intervals of one week between successive meetings. At the first meeting, we informed the participants about the project, explained the dietary questionnaires, and randomized them by lot to the various weight-reduction strategies. At the second meeting, anthropometric data were collected, together with dietary questionnaires. Overweight in adults was defined on the basis of their BMI (weight in kilograms divided by height in meters squared) according to the WHO definition [
During the third meeting, all participants were informed about the energy metabolism of the human body, energy contents of a representative variety of foodstuffs, and energy expenditure by means of physical activity. At the end of the meeting, participants were randomized by lot to one of the treatment options, but they were not blinded to the different study treatments. The lottery was made by nonscientific personnel of the institute that was not involved into the study. However, all participants were aware of the financial incentive and the telemonitoring options. To ensure adherence, the information on the different dietary treatment options was given in a very general way, and detailed information was only given to those participants that were randomized to the particular dietary treatment.
In the calorie-restriction group, further information was given using various practical examples. In the group on the combination diet (dual diet) information was given about carbohydrate metabolism and about the glycemic index of carbohydrates, again using many practical examples. Thereafter, the families were no longer contacted until the control day 6 months later. An exception was the group using the telemonitoring devices, in which each participant received a weekly letter.
The study was conducted in different waves, since the department must be able to handle so many participants. We calculated beforehand that we would be able to handle about 35 families at one time and planned to have 3 waves of study each lasting 6 months. In each wave, each treatment was offered and every participant had the opportunity to be randomized to one of the treatments with the exception that the telemonitoring was only offered during the 2nd and the 3rd of three waves, but not in the first wave. The parallel study followed a three-factorial design as illustrated in Figure
Baseline characteristics in adults (means ± standard deviations). “+” measure present, “−” measure absent.
Telemedicine | − | − | − | − | + | + | + | + | Total |
---|---|---|---|---|---|---|---|---|---|
Financial benefit | − | − | + | + | − | − | + | + | |
Dual diet | − | + | − | + | − | + | − | + | |
Calorie restriction | + | + | + | + | + | + | + | + | |
34 | 28 | 24 | 35 | 7 | 5 | 5 | 4 | 142 | |
Age ( | 41 | 38 | 40 | 38 | 40 | 39 | 39 | 38 | 39 |
Male/female ( | 16/18 | 10/18 | 7/17 | 11/24 | 2/5 | 2/3 | 2/3 | 1/3 | 51/91 |
Smoking status, | |||||||||
current | 11 (32.4) | 10 (35.7) | 4 (16.7) | 10 (28.6) | / | / | 1 (20.0) | / | 36 (25.4) |
former | 12 (35.3) | 10 (35.7) | 6 (25.0) | 14 (40.0) | / | 1 (20.0) | 1 (20.0) | 1 (25.0) | 45 (31.7) |
never | 11 (32.4) | 8 (28.6) | 14 (58.3) | 11 (31.4) | 7 (100) | 4 (80.0) | 3 (60.0) | 3 (75.0) | 61 (43.0) |
Hyperglycemia, n (%) | 3 (8.8) | 1 (3.6) | 1 (4.2) | 2 (5.7) | / | / | / | / | 7 (4.9) |
BMI (kg/m²) | 34 | 33 | 33 | 33 | 34 | 33 | 31 | 37 | 33 |
Height (cm) | 175 | 172 | 170 | 170 | 171 | 175 | 171 | 168 | 172 |
Weight (kg) | 104 | 99 | 95 | 95 | 102 | 100 | 90 | 105 | 99 |
Waist circumference (cm) | 110 | 107 | 106 | 106 | 99 | 105 | 102 | 114 | 107 |
Systolic BP (mmHg) | 138 | 132 | 130 | 127 | 121 | 132 | 126 | 119 | 131 |
Diastolic BP (mmHg) | 87 | 84 | 82 | 82 | 76 | 80 | 82 | 78 | 83 |
Glucose (mmol/L) | 5.3 | 5.4 | 5.4 | 5.2 | 5.1 | 5.5 | 5.1 | 5.1 | 5.3 |
Insulin (pmol/L) | 62 | 49 | 51 | 57 | 50 | 95 | 54 | 87 | 57 |
Total Cholesterol (mmol/L) | 5.9 | 5.3 | 5.2 | 5.1 | 5.2 | 5.2 | 5.0 | 4.9 | 5.3 |
LDL cholesterol (mmol/L) | 3.9 | 3.6 | 3.1 | 3.3 | 3.5 | 3.5 | 3.4 | 3.2 | 3.5 |
HDL cholesterol (mmol/L) | 1.2 | 1.2 | 1.4 | 1.2 | 1.2 | 1.2 | 1.2 | 1.3 | 1.2 |
Triacylgylcerol (mmol/L) | 1.8 | 1.3 | 1.8 | 1.5 | 1.1 | 1.5 | 1.1 | 1.3 | 1.5 |
hs-CRP (mg/L) | 5.3 | 3.3 | 4.2 | 4.3 | 2.5 | 3.6 | 21.7 | 2.7 | 4.8 |
Baseline characteristics in children (means ± standard deviations). “+” measure present, “−” measure absent.
Telemedicine | − | − | − | − | + | + | + | + | Total |
---|---|---|---|---|---|---|---|---|---|
Financial benefit | − | − | + | + | − | − | + | + | |
Dual diet | − | + | − | + | − | + | − | + | |
Calorie restriction | + | + | + | + | + | + | + | + | |
25 | 23 | 22 | 29 | 5 | 6 | 5 | 4 | 119 | |
Age ( | 10 | 10 | 10 | 11 | 10 | 10 | 9 | 10 | 10 |
Male/female ( | 12/13 | 8/15 | 7/15 | 17/12 | 4/1 | 4/2 | 0/5 | 2/2 | 54/65 |
BMI (kg/m²) | 26 | 26 | 24 | 26 | 26 | 25 | 23 | 27 | 25 |
BMI-SDS | 2.08 | 2.06 | 1.94 | 2.11 | 1.99 | 1.98 | 1.88 | 1.74 | 2.03 |
Height (cm) | 151 | 150 | 148 | 152 | 148 | 153 | 145 | 146 | 150 |
Weight (kg) | 60 | 60 | 54 | 61 | 57 | 61 | 50 | 51 | 58 |
Systolic BP (mmHg) | 111 | 109 | 111 | 106 | 102 | 97 | 105 | 116 | 108 |
Diastolic BP (mmHg) | 70 | 72 | 73 | 66 | 58 | 58 | 55 | 66 | 68 |
Glucose (mmol/L) | 4.9 | 4.9 | 4.9 | 4.9 | 5.0 | 5.0 | 4.6 | 4.9 | 4.9 |
Insulin (pmol/L) | 56 | 50 | 52 | 56 | 71 | 47 | 64 | 73 | 56 |
Total Cholesterol (mmol/L) | 4.5 | 4.3 | 4.6 | 4.3 | 3.9 | 3.9 | 4.0 | 4.7 | 4.4 |
LDL cholesterol (mmol/L) | 2.9 | 2.6 | 2.9 | 2.7 | 2.5 | 2.4 | 2.6 | 2.7 | 2.7 |
HDL cholesterol (mmol/L) | 1.2 | 1.3 | 1.3 | 1.2 | 1.2 | 1.2 | 1.0 | 1.4 | 1.2 |
Triacylgylcerol (mmol/L) | 1.0 | 1.0 | 1.1 | 1.1 | 0.8 | 1.1 | 1.2 | 0.8 | 1.0 |
hs-CRP (mg/L) | 2.2 | 2.8 | 2.4 | 2.1 | 1.0 | 3.5 | 6.1 | 1.0 | 2.4 |
Distribution of families between the additional weight-reduction strategies. The design permits a comparison of the three strategies and also of different combinations of these strategies.
The telemedical equipment consisted of a weighing scale for each family, an accelerometer for each participant, and a Homebox for each family which received the data from the scale and the accelerometers via Bluetooth and transferred them via a telephone link to a server in Munich. All instruments were bought from Aipermon GmbH, Munich, Germany. The data were transferred to a server in Magdeburg university hospital where weekly reports were generated and sent to each participant of the telemonitoring group. Each report gave the individual’s weight curve from the beginning of the project and a graph showing for each day of the past week the duration of activity as a percentage of 24 hours, bars with four colours representing four different activity levels from active to sporty, the distance covered in kilometres, and the motoric kcal burned. Each letter also contained comments assessing progress over the past week and aiming to motivate the participant.
For the parents, the financial incentive was 5 Euros for every kilogram of weight loss. For children the weight loss was calculated differently, taking into account the individual need of each child to lose weight. Children with a body mass index between the 90th and 97th age-adjusted BMI-percentile were asked to maintain their weight and were paid in dependence on how well they managed to achieve this goal. Children with age-adjusted BMI-percentiles between 97th and 99th, or above the 99th age-adjusted BMI-percentile received 5 Euros per weight losses of, respectively, 500 g or 1 kg.
All participants received a conventional low-fat diet according to recommendations issued by the Deutsche Gesellschaft für Ernährung [
All participants were advised to reduce their daily energy intake by at least 500 kcal. Dietary changes were monitored by means of 3-day food records. These records were completed for the first time before the first dietary training and then again one week before the control visit at 6 months. The data in these records were evaluated by a computerized programme (DGE-PC, Deutsche Gesellschaft für Ernährung, Bonn, version II.2). The program is based on the Bundeslebensmittelschlüssel II.2 which contains the macronutrient and micronutrient content of
The statistical analysis of the data was done using the SAS package, version 9.1 [
Regarding the problem of missing second visits, analyses were performed both on the basis of the last observation carried forward (LOCF) procedure and with the available data. The LOCF procedure was also used for the children, with the exception of BMI and the body weight. In this case, because children continued to grow during the 6 months of the study, a mean growth of 3 cm in height was assumed. The body weight after 6 months was also assessed with the assumption that the children remained on the same weight percentile as at the start of the study. Because of the small proportion of families with more than one adult or more than one child, possible dependencies between members of the same family were ignored.
In a first step, the target variable was analysed by three-factorial ANOVA (PROC GLM) with the factors telemonitoring, financial incentive, and dual diet including all pairwise interaction terms. As it turned out, the interaction terms were not significant; confidence intervals for the three effects were estimated from a model with the main effects only. In order to distinguish more clearly between the effects of different combinations of the factors, a one-factorial ANOVA was then carried out with the one-sided Dunnett test as a post hoc comparison, using the maximally supported group (telemonitoring plus financial incentive plus dual diet) as comparator for the other groups. In order to obtain a higher power, a step-down version of this test [
In the analyses of the children, the step-down steps did not come into effect because of missing significant results in the primary step. Here, an additional paired t-test was applied for the total group (disregarding different strategies) to check the overall effect of the programme.
Tables
Figure
Dropout rates in groups with different combination of weight-reduction strategies. Parents and children evaluated together.
The weight reductions and the changes in nutrient intake after the implementation of the three additional strategies are given for the parents in Tables
Impact of diets, financial incentive, and telemonitoring on weight loss and nutrient intakes in parents.
Impact of diets on weight loss and nutrient intakes in parents (means ± standard deviations).
Completers | Last observation carried forward | |||||
Calorie restriction | Dual Diet | Calorie restriction | Dual Diet | |||
50 | 61 | 70 | 72 | |||
−4.0 | −6.4 | .029 | −2.9 | −6.0 | .001 | |
Nutrients | ||||||
41 | 54 | 57 | 64 | |||
| −305 | −545 | .048 | −235 | −482 | .014 |
| −10.3 | −31.3 | .003 | −8.0 | −27.5 | .001 |
| −10.6 | −14.0 | n.s. | −7.1 | −11.0 | n.s. |
| −38.3 | −47.9 | n.s. | −31.6 | −44.6 | n.s. |
| +2.3 | −2.7 | (.057) | +1.1 | −2.5 | (.08) |
Abbreviations: CH
Impact of a financial incentive on weight loss and nutrient intakes in parents (means ± standard deviations).
Completers | Last observation carried forward | |||||
Without incentive | With incentive | Without incentive | With incentive | |||
49 | 62 | 74 | 68 | |||
−3.4 | −6.9 | .002 | −2.8 | −6.3 | .000 | |
Nutrients | ||||||
42 | 53 | 63 | 58 | |||
| −379 | −491 | n.s. | −299 | −438 | n.s. |
| −22.2 | −22.3 | n.s. | − 16.9 | −19.8 | n.s. |
| −13.0 | −12.2 | n.s. | − 8.9 | −9.4 | n.s. |
| −27.6 | −56.3 | .039 | −24.9 | −53.2 | .015 |
| −0.4 | −0.6 | n.s. | −0.5 | −1.1 | n.s. |
Abbreviations: CH
Impact of telemonitoring on weight loss and nutrient intakes in parents (means
Completers | Last observation carried forward | |||||
Without telemonitoring | With Telemonitoring | Without telemonitoring | With Telemonitoring | |||
93 | 18 | 121 | 21 | |||
Δ-weight (%) | −4.8 | −8.0 | .033 | −4.1 | −6.9 | .041 |
Nutrients | ||||||
83 | 12 | 106 | 15 | |||
| −449 | −389 | n.s. | −374 | −311 | n.s. |
| −21.4 | −27.8 | n.s. | − 17.7 | −22.3 | n.s. |
| −12.6 | −12.1 | n.s. | − 9.1 | −9.6 | n.s. |
| −48.6 | −10.8 | n.s. | −42.7 | −8.6 | n.s. |
| −1.9 | +9.2 | .004 | −2.0 | +7.3 | .003 |
Abbreviations: CH
Impact of diets, financial incentive, and telemonitoring on weight loss and nutrient intakes in children.
Impact of diets on weight loss and nutrient intakes in children (means
Completers | Last observation carried forward | |||||
Calorie restriction | Dual diet | Calorie restriction | Dual Diet | |||
38 | 49 | 57 | 62 | |||
2.7 | 2.3 | n.s. | 2.4 | 2.2 | n.s. | |
−0.18 | −0.19 | n.s. | −0.13 | −0.15 | n.s. | |
Nutrients | ||||||
34 | 42 | 49 | 53 | |||
| −96 | −239 | n.s. | −37 | −205 | .041 |
| −4 | −11 | n.s. | −2.0 | −10 | (.074) |
| −0.3 | −3 | n.s. | 1.5 | −4.1 | n.s. |
| −16 | −28 | n.s. | −6.7 | −23 | n.s. |
| −2.1 | −6.2 | n.s. | −0.9 | −5.3 | (.068) |
Abbreviations: CH
Impact of a financial incentive on weight loss and nutrient intakes in children (means
Completers | Last observation carried forward | |||||
Without incentive | With Incentive | Without incentive | With Incentive | |||
36 | 51 | 59 | 60 | |||
3.3 | 2.0 | n.s. | 2.7 | 1.9 | n.s. | |
−0.16 | −0.21 | n.s. | −0.09 | − 0.19 | .024 | |
Nutrients | ||||||
31 | 45 | 50 | 52 | |||
| −167 | −180 | n.s. | −115 | −132 | n.s. |
| −8.4 | −7.5 | n.s. | −7.2 | −5.5 | n.s. |
| −2.9 | −3.6 | n.s. | −1.8 | −1.0 | n.s. |
| −20 | −24 | n.s. | −11 | −19 | n.s. |
| −4.5 | −4.2 | n.s. | −3.2 | −3.2 | n.s. |
Abbreviations: CH
Impact of telemonitoring on weight loss and nutrient intakes in children (means
Completers | Last observation carried forward | |||||
Without telemonitoring | With telemonitoring | Without telemonitoring | With telemonitoring | |||
72 | 15 | 99 | 20 | |||
2.5 | 2.5 | n.s. | 2.4 | 1.9 | n.s. | |
−0.18 | −0.20 | n.s. | −0.14 | −0.15 | n.s. | |
Nutrients | ||||||
65 | 11 | 86 | 16 | |||
| −154 | −296 | n.s. | −109 | −203 | n.s. |
| −6.6 | −16 | n.s. | −5.5 | −10.8 | n.s. |
| −2.2 | −10 | n.s. | −0.4 | −6.8 | n.s. |
| −21 | −29 | n.s. | −14 | −20 | n.s. |
| −3.1 | −11.8 | .043 | −2.3 | −8.1 | (.075) |
Abbreviations: CH
Each additional strategy proved to be effective in the parents. The additional relative weight loss in the “completeters” with 95% confidence interval (estimated from the ANOVA model, thus corrected for influences of the other factors) was 2.2% (0.2%, 4.3%) in the dual diet group, 3.3% (1.2%, 5.3%) in the financial incentive group, and 3.7% (0.9%, 6.5%) in the telemonitoring group.
The children also derived benefit from participation in this study. Although in absolute terms they gained between 1 and 2 kg weight, their standard deviation score (SDS)—taking into account the weight gain due to body growth in the course of the study—decreased. This decrease between baseline and 6 months was 0.18
As for the interaction within families, there was no correlation found between the weight losses in parents on the one hand and the weight losses in their children on the other hand. However, in the 28% of all families in which two parents participated, there was a close correlation between the weight losses of the two adults (
Tables
Combination of the additional strategies markedly improved the dropout rates and the losses of weight. Figure
Weight loss in groups of parents with different combinations of weight-reduction strategies. Statistically significant differences are indicated by *(
In this randomized and controlled trial, we evaluated three additional strategies aimed at reducing weight in obese families. The major results are as follows: (
The better effect of the dual diet compared with calorie restriction alone could not readily be foreseen. Studies of the effects of carbohydrate diets with different glycemic index have given contradictory results in the past particularly in respect to weight loss [
It might well have been expected that financial incentive would improve both weight loss and compliance. However, the implications of this observation remain to be debated. Further studies should investigate if this improvement is lasting and if the health benefit justifies the costs. Its effectiveness in adults could be of interest for health insurance companies, which might offer bonuses to obese individuals to encourage them to lose weight and consequently to improve risk markers. Further studies should determine the magnitudes and frequencies of such bonuses necessary to motivate weight losses.
Telemetric monitoring of weight and physical activity is a new and effective strategy. Two separate mechanisms are involved here. The first acts via the continuous feed back from the accelerometer telling the user how effective physical activity has been in terms of the distances covered and—more importantly—in terms of the calories used up. We noticed that the users pay considerable attention to this information—they try to increase their daily activity and to maintain it at a high level. This instrument therefore works like a personal coach, who steadily brings to the user’s awareness his or her physical activity and the associated benefit. The second mechanism is the regular feedback from the person in charge. The families using these devices were asked to use the scales every day. They reported unanimously that the knowledge that their body weight was continually under observation provided an additional stimulus to control food intake and to increase activity during the day. They also reported that they eagerly awaited the weekly letters, which commented on their progress and encouraged them to continue. Taken together, these strategies proved to be an effective tool enhancing their physical activity and boosting their motivation. This technology does both without frequent and time-consuming meetings and also over long distances. It must also be taken into account that telemonitoring is comprised of several elements all of which can contribute to a better weight reduction: daily self-weighing, [
The children who completed the study lowered their BMI-SDS by 0.16 to 0.21 largely independent of the strategies. Because BMI-SDS is an overweight measure which is not easily understood, we illustrate the BMI-SDS change in a hypothetical child: a 10 years old girl with a body weight of 59 kg and 150 cm height would be on the 99th BMI percentile of her age group and her BMI-SDS would be 2.32. Six months later she would have grown by 3 cm. Assuming that her BMI-SDS would be unchanged, her weight gain due to body growth would be 3.6 kg. A reduction of her BMI-SDS by 0.2 (as in this study) would lower this weight gain to 0.5 kg. Therefore, this girl would have avoided an increase of weight by 3.1 kg which has about the same order of magnitude of weight loss as in their parents, namely by 5.1%.
In parents, the combinations of the strategies were more effective than each strategy alone. This applied both to the dropout rates (Figure
Are weight losses after these alternative strategies more sustained than those after calorie restriction? While the data cannot answer this question, we are optimistic that the accelerometers open a new road to their users by improving their awareness of everyday physical activity. Progress in weight control by increased physical activity has been difficult in the past, but may become easier with the use of the telemonitoring. In fact, the dual diet and the enhancement of physical activity are key elements of the desired change in lifestyle, and they are continually reinforced by telemetric control, a kind of lifestyle training.
Summing up, this study shows that in adults weight reduction by fat-calorie restriction can be improved by three additional strategies. Combining these strategies enhances the weight loss remarkably. In children, however, the collective family endeavour seems to be more important than the chosen strategy.