In my inaugural blog I asked whether getting into couples therapy could reduce the risk of a coronary event. I assembled what evidence I could from the literature to answer this question. I concluded that a large sample of couples would have to greatly improve their relationships in order to prevent a few coronary events. Of course, couples therapy has other benefits, but people concerned mainly about avoiding a coronary event would be better off seeking interventions to reduce other risk factors for cardiovascular disease.
Now I'm going to turn the Skeptical Sleuth on my own words and see what evidence there is that behavioral counseling reduces risk factors for cardiovascular disease.
Fortunately, we have the benefit of an exceptionally comprehensive report from the US Preventive Services Task Force, which you can access yourself and compare your assessment to mine (http://tinyurl.com/behavcounseling). Two investigators from the task force reviewed 13,562 abstracts and 481 articles using a set of criteria for methodological quality and excluded studies of lower quality. On this basis, they identified 73 separate studies reported in 109 articles for inclusion in a meta-analysis.
A meta analysis statistically integrates the results of individual studies and when properly conducted is considered the strongest form of evidence, certainly stronger than just counting the number of positive versus negative studies.
I should caution that the studies were quite heterogeneous in their results: technically speaking, that means that the findings varied greatly from study to study, with more intensive behavioral interventions showing the greatest effects on cardiovascular risk factors. And evidence was found for publication bias: the pattern of published results suggested that studies with less positive findings had been withheld from publication. This is quite common in the medical and psychology literatures and needs to be taken into account as part of our judging how much confidence to have in a set of findings from the published literature. Finally, only 11 trials could be located that followed behavioral and physical health outcomes for longer than a year, so we are mainly discussing short term effects that would require behavior change to be maintained after the program ended and there is a lack of evidence that occurs.
Interventions were classified as low intensity if they lasted only 0 to 30 minutes, medium if they lasted 31 minutes to 6 hours, and high intensity if they lasted more than 6 hours.
The results of this enormous undertaking are either exciting or discouraging, depending on the expectations that you have for what can be gotten from such interventions, but I think most people would be disappointed if they achieved the average results of participating in even a moderate to high intensity intervention.
The review found that long-term follow-up showed behavioral counseling reduced sodium intake and led to a decrease in the incidence of cardiovascular disease. However, there was no direct evidence of any reduction in deaths due to cardiac disease. In fairness to these interventions, it is enough of an achievement to be able to demonstrate that the intervention reduced disease, not just changed behavior or reduced sodium intake. But maybe it is too much to expect that less disease would be reflected in significantly fewer deaths. To show that deaths were reduced as well would take a much larger sample followed for a longer time, so that there were sufficient cardiac deaths to register an effect.
The exercise interventions tended to be of low to moderate intensity and I was surprised that they did not yield significant weight loss. High-intensity dietary counseling, regardless of whether was provided in combination with counseling to increase exercise, reduced body mass index .3 to .7 kilograms (.7 to 1.5 pounds), reduced systolic blood pressure 1.5 mm/L, distolic blood pressure .17 mm/L, total cholesterol level 6.56 mg, and low-density lipoprotein cholesterol levels 5.02.
Overall, these are statistically significant difference in cardiovascular risk factors, but they are quite modest. On the other hand, the authors of the report note that at little a change in systolic blood pressure as 2mm can reduce risk for coronary heart disease by 6%.
If we just focus on behavior, the medium to high intensity behavioral counseling yielded moderate to large changes in self-reported dietary and exercise behaviors. This is yet another illustration that getting a substantial change in behavior does not necessarily lead to large changes in health outcomes.
In the typical restrained language of the US Preventive Task Force report, it was concluded that "Counseling to improve diet or increase physical activity changed health behaviors and was associated with small improvement in adiposity [body-mass index], blood pressure, and limited levels." Hmm, not a lot of excitement here.
Practically speaking, what can we make of this massive effort to locate and synthesize evidence concerning the effects of behavioral counseling on cardiovascular risk factors? First, the effort did not locate many high quality studies with an adequate duration of follow up. I hesitate to draw the conclusion that "more research is needed," but this time, it is true. Second, we have a long way to go to improve behavioral counseling if it is to produce clinically significant changed in cardiovascular risk factors. Maybe we need to be realistic that existing programs are not going to produce changes in risk factors needed to shift most people's risk status from high to moderate or moderate to low, and so we cannot rely on such inventions alone. And we certainly need more intensive and effective interventions.
Finally, if we or a family member enroll in such an intervention, we should be easy on ourselves or our family member if dramatic improvement does not occur. It probably cannot be expected.
