|Diabetes? Exercise and then die just as soon.|
It turns out that a just-published and high quality research study shows it's not so simple. What's more, the Disease Management Care Blog brazenly suggests that the disease management/population health vendors discovered this years ago.
The just-published study is here in the prestigious New England Journal of Medicine. The DMCB suspects that, thanks to the mainstream media's fixation on Snowden, SCOTUS, and Shakira possibly hawking Obamacare, this important research may not get the front-page attention it deserves. Considering that it was ten-year, prospective, randomized multi-center academic study involving over 5000 patients, that'd be a shame.
Here's the DMCB's summary:
Eligibility: Participants had to be between 45 and 75 years of age with adequately controlled (A1c less than 11) "type 2" diabetes, an "overweight" body mass index (BMI) of 25 or more, blood pressure less than 160/100, an ability to exercise and access to a primary care provider.
Recruitment: This went from August of 2001 through April of 2004. It was also tailored to keep insulin-using participants to less than 30% of the study group.
Interventions That Were Compared: Participants were randomly assigned to an "intensive lifestyle intervention" study arm or a "support and education" study arm. The intensive group received weekly group and individual counseling for six months that subsequently tapered over the subsequent duration of the study. The counseling included a 1200-1800 calorie diet plus 175 minutes of moderate physical activity per week that was aimed at achieving a weight loss of at least 7% of body weight. The support group got only three group sessions per year. Medicines and their doses were generally left to the primary care provider.
Outcomes Studied: Participants' waist circumference, weight, blood pressure, medications and exercise tolerance were assessed once a year. Hospital and other medical records were reviewed to assess the number of deaths and cardiovascular events, such has heart attack or stroke.
The Study Population: 5,124 persons were enrolled; 2570 were randomly assigned to the intensive group while 2575 were assigned to the support group. The average age was 59 years, 60% were women, the median duration of the diabetes diagnosis was 5 years and the average body mass index was a hefty 36. Only 4% were lost to follow-up.
Outcomes: After a median of 9.6 years of follow-up......
- patients assigned to the intensive group lost approximately three cm. from their waist and six kg. in weight vs. zero cm. and four kg., respectively, in the support group. This translated to a weight loss of 6% of body weight (vs. the target of 7%) in the intensive group vs. about 3.5% in the support group.
- the A1c, which is a test of overall blood sugar control, was about two tenths of a point (7.4% vs. 7.2%) lower (i.e. better) in the intensive group. LDL cholesterol was also lower. Better control of the diabetes meant that the persons in the intensive group were taking fewer medicines at lower doses.
- But it was all for naught. During the course of the study, there were 403 cardiovascular deaths, non-fatal heart attacks or heart-related ("angina") hospitalizations in the intensive group, vs. 418 in the support group. The calculated rates of 1.8 vs. 1.9 events per 100 person years was too small to be statistically significant and was more likely the result of chance or randomness.
The early painful lesson of the "disease management" industry was that a broad life-style intervention applied to a large group of diabetics was not going to meaningfully improve outcomes. Critics believed that while the interventions were conceptually sound (diet, exercise, weight loss), the delivery was flawed.
This just published NEJM study would suggest the intervention itself is futile. If so, that is bad news.
"Not so fast!" says the DMCB.
In addition to renaming itself (now "population health"), the industry responded to the science and the critics by retooling. It learned to channel tailored interventions at population sub-segments who are most likely to experience a specific benefit. Instead of an "intensive" weight loss intervention for all overweight diabetics, population health can use baseline survey, insurance or clinical data to spot (risk stratify) those diabetics who are most likely to achieve a specific benefit that could range from (for example) a sustained 7% weight loss to reduced readmissions.
This NEJM study tried to benefit all diabetics. A better approach is to find which diabetics will benefit.
As an aside there were some other issues with the study to bring up when debating the study with colleagues and foes:
The BMI of 36 suggests this was a very obese study population that lost only 6% of their body weight during the course of the study. Since weight was still a health risk at the end of the study, the DMCB wonders if the intervention would have shown more benefit with a less heavy population.
The support group also lost weight and lowered their A1c, which could have obscured the clinically significant benefit in the intervention group.
This accompanying editorial points out that lower statin and ACE drug use in the intervention group could have paradoxically increased their risk, since these drugs are known to lower the incidence of stroke and heart attack.
The editorial also points out that spin-off studies have already shown that the intervention group benefitted from higher quality of life.