Putting medical practice guidelines into practice: the cholesterol model.

As more and more medical practice guidelines are developed in the United States, commensurate evaluation efforts should assess their impact on professional practice and patient outcomes. We describe an ongoing research program designed to develop and test practice models for applying the 1988 Adult Treatment Panel Guidelines for the clinical management of high blood cholesterol. Four studies are evaluating different models to assist nonacademic community practices in the detection, evaluation, and treatment of high blood cholesterol. We have designed randomized controlled trials set in solo and small-group primary care practices of family or general practitioners and internists situated in rural, suburban, and urban settings. Patients include adult men and women who represent diverse socioeconomic and ethnic backgrounds. We are measuring rates of cholesterol screening; dietary and drug treatment and follow-up; changes in dietary intake and compliance with drug therapy; changes in quality of life and cost of intervention; and reduction in cholesterol level. Scheduled for completion in 1994, this program will provide insights into practical and effective methods of lipid management. It serves as a model for studying the application of health guidelines in the context of nonacademic primary care practices serving diverse patient populations.

Change in cholesterol awareness and action. Results from national physician and public surveys.

The National Heart, Lung, and Blood Institute, Bethesda, Md, sponsored national telephone surveys of practicing physicians and the adult public in 1983, 1986, and 1990 to assess attitudes and practices regarding high serum cholesterol levels. Each time, approximately 1600 physicians and 4000 adults were interviewed. Trends show continuing change in medical practice and public health behavior relating to serum cholesterol. In 1990, physicians reported treating serum cholesterol at considerably lower levels than in 1986 and 1983. The median range of serum cholesterol at which diet therapy was initiated was 5.17 to 5.66 mmol/L (200 to 219 mg/dL) in 1990, down from 6.21 to 6.70 mmol/L (240 to 259 mg/dL) in 1986 and 6.72 to 7.21 mmol/L (260 to 279 mg/dL) in 1983. The median ranges for initiating drug therapy were 6.21 to 6.70 mmol/L (240 to 259 mg/dL) in 1990, 7.76 to 8.25 mmol/L (300 to 319 mg/dL) in 1986, and 8.79 to 9.28 mmol/L (340 to 359 mg/dL) in 1983. The number of adults who reported having had their cholesterol level checked rose from 35% to 46% to 65% in 1983, 1986, and 1990, respectively. Between 1983 and 1990, the number of adults reporting a physician diagnosis of high serum cholesterol increased from 7% to 16%; the number reporting a prescribed cholesterol-lowering diet increased from 3% to 9%. Reports of self-initiated diet efforts reached a high of 19% in 1986 and decreased to 15% in 1990 compared with 1% in earlier years. In 1990, over 90% of physicians reported awareness and use of the recommendations from the Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults, and the public reported marked increases in awareness of dietary methods to lower serum cholesterol. These changes suggest educational gains; the data also suggest areas for continued cholesterol educational initiatives.

Addressing unanswered questions about population cholesterol screenings: the Model Systems for Blood Cholesterol Screening Program.

Elevated blood cholesterol is one of the three major modifiable risk factors for heart disease. Almost 60% of adults in the United States have an elevated blood cholesterol level, yet most adults are unaware of their level. The National Heart, Lung and Blood Institute (NHLBI) and other organizations have now recommended that all adults be tested to assess their blood cholesterol level. New portable blood cholesterol analyzers have recently been designed and are being promoted widely for cholesterol screening. However, there are many unanswered questions about the reliability of these devices and about the usefulness of mass cholesterol screening programs. The Model Systems for Blood Cholesterol Screening Program, an NHLBI-funded effort consisting of three research projects designed to provide a systematic evaluation of these devices and of mass cholesterol screening, is described. This research will contribute to a data base from which recommendations regarding public cholesterol screening will be made.

Screening for high blood cholesterol.

The detection and treatment of high blood cholesterol will require a sizeable screening effort. To a large extent, the screening can be accomplished through routine medical care, but screenings conducted at public or community sites also offer promise. The NCEP will ultimately recommend whether or not public screening should supplement that performed in the physician's office. In the meantime, the National Heart, Lung and Blood Institute has issued screening guidelines that address critical issues including reliability of cholesterol measurement, compliance to the education and referral advice given at screening, and physician readiness to evaluate and treat those referred. Findings from the Model System for High Blood Cholesterol Screening and other research projects provided a basis for such recommendations. It is hoped that these guidelines will ensure high quality public screening.

Change in physician perspective on cholesterol and heart disease. Results from two national surveys.

The National Heart, Lung, and Blood Institute sponsored national telephone surveys of practicing physicians in 1983 (N = 1610) and 1986 (N = 1277) to assess attitudes and practices regarding elevated serum cholesterol levels. The 1983 survey was conducted just before the release of the results of the Lipid Research Clinics Coronary Primary Prevention Trial, which showed that a reduction in the blood cholesterol level reduced coronary heart disease. In 1986, 64% of physicians thought that reducing high blood cholesterol levels would have a large effect on heart disease, up considerably from 39% in 1983. Whereas in 1983, physicians attributed considerably less preventive value to reducing the cholesterol level than to reducing blood pressure or smoking, this disparity was substantially smaller in 1986. The median range of blood cholesterol at which diet therapy was initiated was 6.21 to 6.70 mmol/L (240 to 259 mg/dL) in 1986, down from 6.72 to 7.21 mmol/L (260 to 279 mg/dL) in 1983; the median for drug therapy was 7.76 to 8.25 mmol/L. (300 to 319 mg/dL) in 1986 and 8.79 to 9.28 mmol/L (340 to 359 mg/dL) in 1983. In 1986, 87% of physicians surveyed felt that medical evidence warranted the recommended treatment levels set forth in the 1984 National Institutes of Health Consensus Conference on Lowering Blood Cholesterol. These changes indicate that by 1986, physicians were more convinced of the benefit of lowering high blood cholesterol levels and were treating patients accordingly. The data also suggest areas for continued educational initiatives.

Change in public perspective on cholesterol and heart disease. Results from two national surveys.

The National Heart, Lung and Blood Institute, Bethesda, Md, and the Food and Drug Administration, Washington, DC, sponsored two national probability telephone surveys (N = 4000) of adults to assess attitudes and knowledge about heart disease risk from high blood cholesterol levels and the public's efforts to lower blood cholesterol levels. The first survey was conducted in 1983, before release of the results from the Lipid Research Clinics Coronary Primary Prevention Trial, which showed that a reduction in the blood cholesterol level reduced coronary heart disease; the second survey was conducted in 1986. The percentage of adults who believed that reducing high blood cholesterol levels would have a large effect on heart disease increased from 64% in 1983 to 72% in 1986, so that the importance attached to reducing high blood cholesterol levels approached that attributed to reducing smoking and high blood pressure. In 1983, 35% of adults reported that they had their cholesterol level checked vs 46% in 1986. In both years, diet changes were most frequently chosen (greater than 60%) as ways to control the blood cholesterol level; reducing dietary fat was believed to be as important as reducing dietary cholesterol. By 1986, 23% of adults reported that they made dietary changes specifically to lower their blood cholesterol level, up from 14% in 1983. These comparative data show gains in public awareness and action relating to high blood cholesterol level risk. The data can be used to develop education programs.

Nonverbal behavior of the type A individual.

Forty-one health professional students were videotaped during three consecutive conditions: a 5-min wait for a tardy interviewer, the structured interview for determining the Type A behavior pattern, and a 5-min relaxation period. Afterward, subjects were classified as Type A or B based on audiotapes of the interview. The total amount of activity and the frequency or duration of Type A behaviors were compared between Type A and Type B subjects. In all three experimental conditions, Type A subjects moved their arms more than Type B subjects. During both the waiting and the relaxation periods, Type A subjects sat still less and spent more time exploring than did Type B subjects. During the interview, Type A subjects gestured more frequently than Type B subjects. Nonverbal behaviors correctly identified behavior pattern in 71% of the subjects. The addition of nonverbal behaviors improved the discrimination of behavior pattern above that obtained from verbal behavior alone. The greater activity, restlessness, exploratory behavior, and gestures of Type A persons are consistent with the two major etiologies proposed for the Type A behavior pattern.

Assessment of behavioral risk for coronary disease by voice characteristics.

Behavior characterization is increasingly used and studied in the pathogenesis and management of coronary disease. A specific behavior pattern called Type A has been found to be independently associated with fatal and nonfatal coronary disease. The assessment of Type A is made without quantitative guidelines based on verbal and nonverbal behaviors in a standardized interview. This study has attempted to make this classification solely on the basis of quantifiable verbal behaviors. Two naive persons were trained to code speech behaviors from tape recorded behavior pattern interviews on which Type A had previously been assessed. Intercoder and intracoder reliability was acceptable. Using regression analysis 87% of the interviews were classified in agreement with the previous, standard assessment. Volume of voice and speed of speech were found to make the greatest predictive contribution. Other verbal characteristics including plosiveness were also important. It is concluded that Type A can be validly and repeatably measured by verbal behaviors in the standardized interview.