Inadequate physician knowledge of the effects of diet on blood lipids and lipoproteins.

Background To assess the nutrition knowledge of physicians on the basic effects of diet on blood lipids and lipoproteins. Methods Anonymous mailed dietary knowledge surveys to 6000 randomly selected physicians in the United States licensed in either Internal Medicine or Cardiology. Results Response rate: 16% (n = 639). Half of the physicians did not know that canola oil and 26% did not know olive oil were good sources of monounsaturated fat. Ninety-three percent (84% of cardiologists vs. 96% of internists; p < 0.001) did not know that a low-fat diet, in general, would increase blood triglycerides. Approximately three-quarters (70% of cardiologists vs. 77% of internists; p < 0.01) did not know a low-fat diet would decrease HDL-c and almost half (45%) thought that a low-fat diet would not change HDL-c. Conclusions If physicians are to implement dietary and cholesterol management guidelines, they will likely need to become more knowledgeable about nutrition.


Background
Diet modification is recommended by the Third Report of the National Cholesterol Education Program (ATP3) as the initial treatment to lower blood cholesterol [1]. It is the physician's responsibility, according the ATP3, to encourage patients with elevated cholesterol to initiate a Therapeutic Lifestyle Changes Diet (TLC). The TLC diet recommends a range for the percent of calories from various sources including saturated fat (<7), total fat (25-35), and carbohydrate (50-60) [1]. Physicians' ability to effectively counsel patients to adopt a TLC diet, however, has never been examined. One of the major changes to ATP3 (v. ATP2) is to advocate lowering triglycerides, as a secondary target to lowering LDL, for persons with triglycerides (TG) >200 mg/dl. In this study, we set out to examine physician's understanding of basic nutritional physiology underlying recommendations in the TLC diet, particularly those recommendations specific to triglycerides, as triglycerides are a new prevention treatment target.

Methods
A random sampling mailing list of 6000 physicians licensed in the United States stratified to include physicians licensed in either Internal Medicine or Cardiology was purchased from Buckley Dement (Ridgewood, NJ) in April 1998. A brief questionnaire, nine items total, was sent to a random sample of 4000 physicians from the purchased list. Questions were designed by two of the authors (MF, KV), a Ph.D. nutritionist and a physician, respectively. The four survey items examined in this analysis are listed in Table 1. The remaining 5 items were based on dietary recommendation strategies that are no longer consistent with the ATP3, so they are not reported here. A cover letter indicated that the survey was anonymous and a postage paid return envelope was provided. Physicians who did not respond were not sent a second survey. Physicians were not paid for completing the survey.
Descriptive statistics are reported as percentages using whole numbers. The Chi-squared test was used to evaluate differences between cardiologists and internists. Significance was set at 0.05. All data were analyzed with SPSS, version 4.1.

Results
A total of 639 surveys were returned (16% response rate). Sixty-six percent were internists (n = 419), 19% were cardiologists (n = 120), 7% identified themselves as "other" and 8% did not indicate a specialty. The responses to these questions are in Table 1 for the total group and by cardiologist and general internist. Half of all those responding did not know that canola oil and 26% did not know olive oil were good sources of monounsaturated fat. Thirty percent mistakenly thought safflower, a highly polyunsaturated fat oil, was a good source of monounsaturated fat. Ninety-three percent (84% of cardiologists vs. 96% of internists; p < 0.001) did not know that a low-fat diet, in general, would increase blood triglycerides. Approximately three-quarters (70% of cardiologists vs. 77% of internists; p < 0.01) did not know a low-fat diet would decrease HDL-c and almost half (45%) thought that a low-fat diet would not change HDL-c. About onehalf (47%) (22% of cardiologists vs. 53% of internists; p < 0.001) did not know carbohydrate was the diet component most likely to raise triglycerides.

Discussion
The main finding in this study was that physicians might be lacking adequate nutrition knowledge to counsel their patients accurately about TLC diets, a diet recommended for all patients with high blood cholesterol [1]. Physicians showed a poor understanding of the effects of changing the relative intake of carbohydrates and fats on triglycerides and HDL. Only half of physicians identified canola oil as a good source of monounsaturated fatty acids and few physicians understood the relationship of changes in dietary fat with either triglycerides or HDL levels. On average, cardiologists in our study had a better understanding of nutrition than general internists, which we expected given their additional training and the fact that cardiologists often receive referrals to manage elevated cholesterol levels. * reflects significant differences between percentage of internists and cardiologists answering each question correctly.
Physicians note many barriers to counseling their patients about making health behavior changes [2]. These findings agree with Kushner and colleagues, who found that primary care physicians often note a lack of knowledge, in addition to a lack of time and physician confidence, as main barriers to diet counseling [2]. This is also in keeping with a U.S. Public Health Service study noting that physicians are "woefully undertrained in nutrition" [3]. Additional training may well be an important component, as cardiologists in our sample had a much better understanding of the effects of dietary changes on triglycerides and HDL than their general internist counterparts.
Our findings have several noteworthy limitations. First, the response rate of 16% raises the possibility of a significant response bias among physicians responding to the survey. As we did not evaluate a sample of non-responders to examine differences between responders and those who did not respond, the possibility of non-response bias remains. We believe, however, that this risk is minimal. As the survey was clearly to measure nutrition knowledge, we presume that biases inherent in a physicians desire to be knowledgeable about medical issues would increase the likelihood that the responding physicians were more knowledgeable about nutrition. No studies to our knowledge have attempted to measure nutrition knowledge of practicing physicians, so it is impossible to examine the comparative validity of these findings. Other studies are needed to understand whether or not these findings generalize to other groups of physicians (e.g., family physicians) or to other nutritional issues (e.g. dietary fiber).
Second, we did not examine other aspects (e.g., sources of dietary fiber) of a physicians' nutrition knowledge. The rationale for the study grew out of conversations with physicians and a continuing education seminar by one of the investigators (MF) for physicians, which identified a lack of understanding of the dietary aspects of cholesterol management. There are many other aspects of nutrition in medical practice and it is possible, though unlikely, that physicians have an excellent knowledge in these other areas.
Third, the term "low fat" was not defined, so it is left open to interpretation, which may not have been consistent across physicians. We chose not to define the term because, as it is frequently used without being defined, we wanted to assess the response of physicians using their working definitions of low fat diets and not impose a definition. This is in keeping with standard practice as a study of over 6000 adults with hypercholesterolemia revealed that 65% had been instructed by a physician to follow a "low-fat diet" without further defining the term [4].
Physicians showed a poor understanding of the effect of common dietary modifications used in TLC diets. Given the level of importance assigned to lifestyle changes in the TLC diet from the ATP3 report, it may prove more useful to develop counseling services from allied health care providers [3]. This is the standard of care for patients with diabetes, as the American Diabetes Association recommends yearly counseling with a nutritionist for all patients with diabetes [5]. Enhancing physician training about nutrition, in medical school, residency and beyond, may also be a reasonable approach, though modifying medical student curricula and changing physician practice are both important but notoriously resistant to change [6].
While recommendations have been made to increase nutrition counseling by physicians, physicians may not be receiving adequate training in nutrition. The National Nutrition Monitoring and Related Research Act of 1990, recommended "that students enrolled in United States medical schools and physicians practicing in the United States have access to adequate training in the field of nutrition and its relation to human health" [7], yet the number of medical schools that require, or even offer courses in nutrition was declining as of 1994 [8].

Conclusions
As physicians are viewed as reliable sources of information, including nutrition, and their nutrition counseling responsibilities have greatly increased in the ATP's TLC diet, it is imperative that physicians become more knowledgeable about nutrition.