2.3.1 Dietary Interventions
Evaluation of dietary advice to reduce disease risk factors or to treat an existing disease is based on six reviews with approximately 44,000 participants. Outcomes of these reviews are listed in Table 62.1 (SR1-6). Usually, they were physiological or anthropometric risk factors. While behavior was reported in some trials as an indicator of intervention compliance, it was rarely measured as an outcome. Quality of life and cost-effectiveness were not reported. Overall, small but statistically significant improvements were found in physiological and anthropometric outcomes (Table 62.1, SR1-6). Only one review (Brunner et al, 2007) reported on behavioral outcomes: in comprehensive dietary interventions for reducing CVD risk, modest beneficial changes in behavior were translated into statistically significant improvements in physiological and anthropometric outcomes. Beneficial disease outcomes included reduction of cardiovascular events, but this effect could only be established for interventions to reduce fat (Hooper et al, 2000). Another disease outcome was reduction of type 2 diabetes incidence (Nield et al, 2008), but the finding was based on just one trial and therefore provided only weak evidence. Outcome evaluation of diet only interventions in treatment of type 2 diabetes (Nield et al, 2007) was rendered impossible by the heterogeneity and poor quality of the studies. The only review on dietary interventions capable of reporting on mortality (Hooper et al, 2000) found no statistically significant effects on either cardiovascular or total mortality.
Effects of intervention delivery and duration were evaluated in two of the dietary intervention reviews. While dietary advice from a dietitian was more effective than advice from a doctor, it was no more effective than fairly simple self-help material. Comparison between dietitians versus nurses was not possible because of limited data (Thompson et al, 2003). The beneficial effect of fat reduction to cardiovascular events was limited to trials with intervention extending over 2 years (Hooper et al, 2000). Although one review (Hooper et al, 2000), p. 18) specifically referred to the potential benefits of applying a behavioral theory for improving intervention outcomes, none actually discussed the interventions within any sort of theoretical framework.
Evaluation of exercise interventions covers over 10,000 participants but it is mainly based on reviews among patient populations. The intervention programs were heterogeneous (e.g., exercise alone or as part of comprehensive rehabilitation program including educational or psychological interventions) - often meaning that the independent effect of exercise could not be separated. A range of outcome measures was covered but typically the focus was on physiological measures and as a secondary outcome, quality of life. Morbidity and mortality were rarely studied, and no review reported on cost-effectiveness. Behavior was regarded as a compliance factor rather than a primary intervention outcome. None of the reviews discussed behavioral theories used in the interventions although the comprehensive rehabilitation programs often included an educational or psychosocial component.
Overall, exercise training was found to improve several of the measured physiological or anthropometric factors (Table 62.1, SR7-10), such as glycated hemoglobin (HbA1c) among patients with type 2 diabetes (Thomas et al, 2006) and lipid profile among patients in comprehensive cardiac rehabilitation (Jolliffe et al, 2001). No reduction was found in body mass index (BMI), but body composition changed significantly, with adipose tissue decreasing and fat-free mass increasing. In most cases only short-term effects on risk factors could be established due to lack of long-term follow-ups. However, the one review with a longer follow-up was able to show reduction in both all cause and cardiac mortality (Jolliffe et al, 2001). Findings on quality of life were mixed, although most studies tended to have found small improvements.
Duration of the interventions ranged widely, and one review evaluated its effect on outcomes. Among patients with type 2 diabetes, decrease in HbAic was greater for briefer (< 6 months) interventions than for longer interventions (6-12 months) (Thomas et al, 2006). One review also evaluated the effects of intervention delivery on outcomes (Ashworth et al, 2005). Assessment of home- versus center-based physical activity programs in older adults showed that on the short term, center-based training produced better outcomes. However, two studies in the review were able to evaluate longer-term adherence (12 years) which was shown to be better in the home-based program.
2.3.3 Combined Diet and Exercise/Weight Reduction Interventions
Altogether 15,000 patients are included in the assessment of the efficacy of combined diet and exercise or weight reduction interventions on prevention (Norris et al, 2005b; Orozco et al, 2008) and treatment (Norris et al, 2005a) of type 2 diabetes. Prevention interventions included participants with elevated risk for type 2 diabetes, who typically (but not necessarily) had an impaired glucose tolerance. Behavior was not reported as a specific outcome measure although one review included physical activity and diet as indicators of compliance (Orozco et al, 2008). Typically, outcomes included incidence of type 2 diabetes and physiological and anthropometric risk factors (Table 62.1, SR11-13). None of the reviews were able to report on behavior or mortality. No aggregated data was provided for cost-effectiveness although two trials in one review provided support for cost-effectiveness (Orozco et al, 2008). Length of follow-up ranged from 12 months to 10 years (Norris et al, 2005b). Although most of the reviews mentioned the use of behavioral theories and/or specific intervention strategies (including goal setting, self-monitoring and feedback, and stress management and coping), these were not systematically evaluated.
Small improvements were found in weight, BMI, and waist circumference (Table 62.1, SR11-13), although statistical heterogeneity for these outcomes was high and effects were minimized by significant weight loss in the comparison groups (Norris et al, 2005a). Modest, but statistically significant improvements were shown on many physiological and anthropomet-ric outcomes. Both reviews on prevention of type 2 diabetes reported statistically significant reductions in the incidence of type 2 diabetes, but only one provided a pooled effect (Orozco et al, 2008).
Intervention duration per se was not shown to effect outcomes although the number of contacts correlated positively with weight loss among adults with pre-diabetes (Norris et al, 2005b). Furthermore, examination of different intervention arms suggested that multi-component interventions with low or very low calorie diets might help to achieve weight loss among patients with type 2 diabetes (Norris et al, 2005a).
Evaluation of tobacco control interventions is based on nearly 110,000 participants including mainly healthy adults. Without exception, the outcomes were always measured in terms of behavior. For interventions addressing cessation the outcome was abstinence, in prevention interventions it was smoking behavior (Table 62.1, SR14-18). None of the reviews reported on physiological or anthropometric outcomes or quality of life, and only one study reported on disease or mortality outcomes. Cost-effectiveness was not commonly reported although three reviews estimated the number required to treat one individual successfully (NNT). Unlike the dietary and exercise interventions, smoking interventions were commonly based on theoretical models, especially if they were delivered by professionals other than physicians or nurses. However, none of the reviews compared different theoretical approaches to behavior change.
Nursing interventions (Rice and Stead, 2008), physician advice (Stead et al, 2008), psychosocial interventions (Barth et al, 2008), and interventions delivered by oral health professionals in connection with oral examination (Carr and Ebbert, 2006) were all shown to be effective. The overall likelihood for quitting was 28-66% higher in these interventions in comparison to usual care. Heterogeneity of community interventions prevented pooling of those data, so overall quitting rates could not be established. However, only 2 out of the 13 community interventions were more effective than no treatment (Sowden et al, 2003). Estimated NNT ranged between 10 and 120, the lowest NNT being in CHD patients who tend to have high unassisted quit rates anyhow (30-50%) (Barth et al, 2008) and the highest NNT was in primary care patients, whose unassisted quit rate was estimated to be only 2-3% (Stead et al, 2008). The minimum follow-up time for all interventions was 6 months and the majority had 12 month or longer follow-up.
Higher intensity interventions increased the effectiveness of interventions. Psychosocial interventions including behavioral therapies outperformed usual care, as did those based on self-help or telephone support, but none of these were found superior to each other. Furthermore, a subgroup analysis among patients in the dental setting showed the interventions effective regardless of whether participants had actively sought treatment (Table 62.1).
One extensive review with almost 150,000 participants evaluated the effects of multiple risk factor interventions among adults without clinical evidence of established CVD. All the trials compared an intervention comprising some form of education or counseling targeting combinations of diet, exercise, weight loss, smoking cessation, diabetes management, and use of medication with control groups receiving either usual care or no treatment. Behavioral theories underlying the interventions were rarely specified, with a few studies using the Transtheoretical Model of Stages of Change as an exception (DiClemente et al, 1991). Smoking was the most common behavioral outcome included, other reported outcomes included blood pressure, cholesterol, and mortality. Quality of life outcomes or cost-effectiveness were not reported.
Overall, the interventions had a small, positive effect on smoking prevalence (Table 62.1, SR19). Also, modest but statistically significant improvements were shown in blood pressure and cholesterol, but these were most likely related to pharmacological treatments rather than the lifestyle interventions used. Ten of the trials provided data on CHD mortality and total mortality, but overall, no effect could be established. Studies where participants had highest initial risk factor levels demonstrated larger improvements in these factors (Ebrahim et al, 2006).
Interventions for improving risk or disease management were assessed among almost 30,000 patients with vascular conditions. Included were interventions with a narrow focus to adherence to treatment recommendations (Table 62.1, SR20-21) as well as broader self-management education and support programs delivered by professionals (SR22-25) and peers (SR26-27). Reporting on outcomes focused on physiological and anthropometric risk factors but also behavior and quality of life were included. Only one review reported on disease outcomes and mortality.
The disease management interventions were probably even more heterogeneous in content, intensity, and duration than the other lifestyle interventions already described in this chapter. While some tackled only relatively simple behaviors (such as blood glucose monitoring or taking a lipid lowering medicine), others addressed very complex sets of behaviors (e.g., comprehensive disease management including lifestyle, self-care, and adherence to medical care). Rather surprisingly, the interventions targeting simpler behaviors were often more heterogeneous and more poorly described in the reviews. They also typically lacked any description of the behavioral component(s). The more comprehensive programs, however, were often theory based, well described, and they also allowed comparison of different modes of delivery.
Behavioral outcomes were rarely established in interventions targeting adherence (Table 62.1, SR20-21) and improvements in physiological outcomes tended to be small at best. None of the few studies including quality of life measures showed any significant effects on it. Morbidity and mortality outcomes were not reported for adherence interventions. The effect of intervention characteristics on the outcomes was not evaluated in any of the reviews.
The more comprehensive self-management programs delivered by professionals (Table 62.1, SR22-25) provided heterogeneous results in terms of behavior change. In terms of physiological or anthropometric outcomes among patients with type 2 diabetes, individual patient education was no better than usual care or group education (Duke et al, 2009). However, group training resulted in moderate improvement in most of these risk factors (Deakin et al, 2005). Disease and mortality outcomes were only measured among CHD patients. No effect was found on mortality, and the significant reduction in the number of non-fatal reinfarc-tion was found to be influenced by publication bias (Rees et al, 2004a). Quality of life outcomes in interventions among type 2 diabetes patients were mixed, but CHD patients were shown to gain modest psychological benefit from the interventions in terms of reductions in anxiety and depression (Rees et al, 2004a).
Self-management programs by lay leaders had very little effect on behaviors, physiological and anthropometric outcomes, or quality of life (Table 62.1, SR26-27). None of the studies reported on morbidity or mortality (Dale et al, 2008; Foster et al, 2007).
Was this article helpful?