Latest Mediterranean Diet Study Results: Much Ado About Nothing

Olive oil

Fearing that it would have to write about the Oscar award show’s vulgar Hollywood humor or Washington’s budgetary caterwauling, the Disease Management Care Blog was happily distracted by a flurry of splashy news reports on a just-published primary prevention trial:

“Mediterranean diet….showed 30% lower risk of having a heart attack, stroke or dying of heart disease…..”

“Tasty diet cuts heart disease, study finds.”

” ….diet rich in olive oil, nuts, beans, fish, fruits and vegetables prevents about 30% of heart attacks, strokes and deaths from heart disease…”

Fighting the urge to run to the pantry and add years to its life by ingesting some olive oil-drizzled walnuts, the DMCB did something that mainstream news reporters seem incapable of doing: it asked what the diet actually entails and whether the impact was really all that.

Here’s the original New England Journal of Medicine study. Disappointingly, the absolute risk reduction was quite small and, for the typical DMCB primary care patient, would involve a major lifestyle change.

The DMCB explains.

The PREDIMED trial (“Prevención con Dieta Mediterránea”) was a prospective and randomized clinical trial in which male (age 55 to 80 years) and female (age 60 to 80 years) volunteers with either:

1) diabetes or

2) three other risk factors, such as tobacco use, high blood pressure, abnormal cholesterol levels, obesity and a worrisome family history of premature heart disease,

were randomly assigned to one of three treatment arms:

1) A “Mediterranean diet” supplemented with extra-virgin olive oil (a liter was delivered to the home each week), or

2) A “Mediterranean diet” supplemented with nuts (30 grams of almonds, walnuts and hazelnuts per day) or

3) A control diet.

What’s just what is a Mediterranean diet?  According to the authors:

A) lots of olive oil for cooking and dressing dishes;

B) consumption of ≥ 2 daily servings of vegetables (at least one of them as fresh vegetables in a salad), discounting side dishes;

C) ≥ 2-3 daily servings of fresh fruits (including natural juices);

D) ≥ 3 weekly servings of “legumes” (kidney beans, pinto beans, black beans, chickpeas, lima beans, black-eyed peas, split peas, and lentils);

E) ≥ 3 weekly servings of fish or seafood (at least one serving of fatty fish);

F) ≥ 1 weekly serving of nuts or seeds;

G) selecting white meats (poultry without skin or rabbit) instead of red meats or processed meats such as burgers and sausages;

H) cooking regularly (at least twice a week) with tomato, garlic and onion adding or not other aromatic herbs, and dress vegetables, pasta, rice and other dishes with tomato, garlic and onion adding or not aromatic herbs. This sauce is made by slowly simmering the minced ingredients with abundant olive oil.

I) the limitation of cream, butter,margarine, cold meat, pate, duck, carbonated and/or sugared beverages, pastries, industrial bakery products (such as cakes, donuts, or cookies), industrial desserts (puddings, custard), French fries or potato chips, and out-of-home pre-cooked cakes and sweets.

By the way, the study organizers used genuine olive oil, not this fake stuff.

Patients assigned to the diet had to see a dietitian at the outset of the study and participate in every 3 month group education sessions.  There were no recommendations to lose weight or participate in exercise.

The results?

8713 persons were screened and 7447 were assigned to one of the study arms.  After a median of 4.8 years, 523 persons were lost to follow-up. 288 persons had either a heart attack, stroke or
cardiovascular death.  There were 96 deaths in the extra virgin olive oil group, 83 in the nut group and 109 in the control group. The calculated percents were 3.8%, 3.4% and 4.4%, respectively.  Versus the control group that’s an absolute difference of 0.6% and 1.0%. There didn’t appear to be any surprises in the subgroups’ outcomes.

Using a number needed to treat analysis, these data basically mean 100 to 166 persons would need to live on fish, nuts, lima beans, and olive oil without burgers, starches or soda for 5 years to prevent a single heart attack, stroke or death.

The DMCB’s take:

1. The small amount of benefit is not enough, in the DMCB’s estimation, to warrant routine inclusion of a “Mediterranean diet” in disease management care planning for persons with diabetes or multiple cardiac risk factors.

2. The diet was “all or none.”  There is no evidence, based on this trial, that substituting crab cakes for that steak once a week and using olive oil on your iceberg lettuce tonight will help you live longer.

3. The “number needed to treat” of 100-166 over 5 years is in the same performance range as statin drugs with a NNT that, according to this study, ranges from 77 to 150 over four years.  No wonder many persons choose to continue with their more tasty burgers and fries along with a daily pill.

4. It’s unlikely that persons who are not already on a Mediterranean diet will chose to switch, and given the amount of sacrifice involved and the small reduction in risk, the DMCB can’t blame them.

Image from Wikipedia

A Summary of the Latest Population Health Management Journal

It’s that time again. The latest issue of Population Health Management is out and you’d read it if you weren’t so busy with other stuff. After all, PHM is your window into the latest goings-on in the disease management community, its information gives you and your company a competitive advantage and quoting from it impresses policy makers, bosses and colleagues. Good thing you read the Disease Management Care Blog: it has the information you want in a format you can quickly use.

Check it out and decide just which articles you really need to read and which ones are just FYI. So, without further ado……

In this ‘Point of View,’ the veteran Robert Stone of Healthways discusses the maturation of the disease management industry with a special emphasis on the insurers’ eternal choice of ‘build or buy.’ According to Mr. Stone, insurers want mutually supportive and broad-based health, wellness, prevention, case and disease management on one platform that are all built to last and are adequately capitalized. For those that are foolish enough to think about building, he cautions the availability of tools is not synonymous with an ability to use them. A telling quote: ‘Price is not the best indicator of ultimate value.’

In this article, Harry Leider of Ameritox, David Mirkin of Milliman and Christobel Selecky of LifeMasters reminisce about the recently concluded Ninth Population Health and Disease Management Colloquium. Harry pointed out there were presentations about conditions that have been largely ignored by the industry, such as chronic pain, autism, migraine and psychiatric conditions. David reviewed how unsettled the science is of using actuarial trends to estimate the economic impact of disease management programs. Christobel detailed how there is a growing emphaisis in her company and among others in maximizing patient activation. Good quote from Ms. Selecky about trending: “I wonder if people arent’ just shell-shocked with trying to come up with a methodology – once you think you have it nailed down, something squirts out the other end.”

Thomas Foels, Sharon Hewner: Integrating pay for performance with educational strategies to improve diabetes care. This describes how Independent Health of Western New York State compensated physicians (60 to 70 cents PMPM plus CME) to conduct reviews of their own charts for diabetes care quality. Physicians were then provided summary data that included an estimate of the patient’s overall burden of illness along with suggestions for improvement. 84% of the physicians participated, and over time there were at least 10 percentage point gains in the usual measures of blood pressure (less than 130/70), LDL (less than 100) and A1c (less than 7). The authors say – with very little detail – that they saved money. The DMCB thinks this is was an interesting article because this was more of a pay for ‘quality improvement program’ (? P4QI?) than a typical pay for performance (P4P) program: that seems unique. Kudos to the authors for this quote: ‘There were several limitations to this study, thanks to the lack of a control group, a small sample size per practice site, underrepresentation of rural and small practices and the selection of patients used in the survey based on a claims profile.’ The DMCB says this is promising and some more research is needed.

George Ioannidis, Alexandrea Papaioannou, Lehana Thabane, Amiram Gafni, Anthony Hodsman, Brent Dvern, Eleksandra Walsh, Famida Jiwa, and Jonathan Adachi. Family Physicians’ Personal and Practice Characteristics that Are Associated with Improved Utilization of Bone Mineral Density Testing and Osteoporosis Medication Prescribing The authors used a physician questionnaire from 225 Canadian docs to assess personal and practice characteristics and then correlated those results to the likelihood of ordering osteoporosis testing and treatment. Being female, not having hospital privileges, not being a recent medical school graduate correlated with ordering bone density testing, having an electronic health record was associated with treatment for osteoporosis. Best quote: ‘This is not surprising…..’

Susan Robinson, Robert Baron, Bruce Cooper and Susan Janson: Does health service use in a diabetes management program contribute to health disparities at a facility level? Optimizing resources with demographic factors. These researchers from the University of California followed 315 persons with diabetes for 18 months. Since all had equal access to the clinic, the authors were interested in knowing whether demographic factors correlated with healthcare utilization. Persons with Medicare and Medicaid as well as persons of Hispanic heritage used the emergency room more frequently. Women were more likely to be hospitalized and Hispanics less so. Telling quote: Clearly ‘disparities in utilization’ of health care services continue to exist within demographic subpopulations.” And we thought disparities hindered access to care.

Micah Throp, Jessica Weinstein, Jason DeVille Eric Johsnon, Amanda Petric, Xiuhai Yang, David Smith: Comparison of renal replacement therapy and mortality using 1 versus 2 estimated glomerular filtration rates. Using data from their electronic health record (EHR), these Kaiser researchers found that when when physicians ordered an ‘estimated glomerular filtration rate’ (eGFR)and found evidence of chronic kidney disease, the appearance of a second eGFR in the order data sets predicted a worse outcome. If you use the electronic record to find patients for a CKD disease management program, you may want to use this signal as evidence of a higher burden of illness. The bottom line quote: For the purpose of disease management, it is less important that the second low eGFR demonstrate [sic] an independent prediction of mortality…. Disease managemer can take advantage of its superior predictions….”

Yiduo Zhang, Timothy Dall, Yaozhu Chen, Alan Baldwin, Wenyua Tyng, Sarah Mann, Victoria Moore, Elisabeth L Nestour, William Quick: Medical cost associated with prediabetes. Using a ‘Cost of Diabetes Model,’ NHANES, Census Bureau Data and the Ingenix Research DataMart, these Lewin Group researchers ultimately estimate there is a total cost of $ 25 billion or $ 443 per person with prediabetes, mostly due to increased outpatient utilization, not from emergency room or inpatient care. What’s more, ‘these cost estimates understate the true cost of prediabetes to society…[due to] increases in missed work days and lower productivity.’ The DMCB says use this paper when you want to market your metabolic syndrome disease management program.

Yaozhu Chen, William Quick, Wenyua Yang, Yiduo Zhang, Alan Baldwin, Jane Moran, Victoria Moore, Navita Sahai, Timothy Dall: Cost of Gestaional diabetes mellitus in the United States in 2007. This is another fact-fest from the Lewin Group who, this time, relied on discharge data to that gestational diabetes to estimate that the cost is $ 636 million, or $ 3,305 per pregnancy.