University of Minnesota

“It Isn’t Always Fun.” – Minnesota Heart Survey

As for the Minnesota Heart Survey (MHS), despite its extended and effective run, now at 24 years and counting, I find there’s not an anecdote in a barrelful about it. The action, the drama, and the humor of a large population study occurs at the level of the recruiters and the survey teams and the laboratories, not in administrative board rooms or analysis centers. I should have been where the action was, as I had been in the Seven Countries Study. I hope, however, that others will find and describe the flavor of this and other surveillance studies. I attempt here simply to summarize what our Minnesota baby, the MHS, was, and still is, all about:

In the Minnesota Heart Survey, we address the population-wide phenomena of heart attacks and strokes. We describe the components of mortality trends and try to explain them. For example, do the downward trends in deaths occur in both sexes, young and old, and in ethnic and social sub-groups?  Are the trends greater or less for sudden, out-of-hospital deaths or for those in-hospital? Are they associated with changes in population levels of risk factors, and, if so, what are the lag times between average risk factor change and change in death rates? Are the slopes of mortality rates constant or variable? Are changes in disease rates and risk factor levels in the community associated with quantifiable health behavior changes in, for instance, smoking, nutrient intake and dietary patterns, or physical activity? What is the relative contribution of changes in behavior to the public health versus those due to innovations in medical care of heart attacks? Is there a change in the manifestations and severity of attacks (e.g. In the proportion with sudden death, congestive heart failure, or arrhythmias)? Are there specific medical interventions (e.g. CABG, PTCA, thrombolysis, aspirin use) that measurably affect population mortality rates?

We felt in the late 1970s that we were ready and able to measure these trends and their components and possible causes in the population. We had worked in the Twin Cities’ population from the early CVD study of the 1940s through the National Diet-Heart Study of the 1960s and the clinical trials of the 1970s. Moreover, we had prepared the WHO Manual Cardiovascular Survey Methods with Geoff Rose, and thus knew something about the  principles and methods. Henry Taylor and I had accumulated a vast hands-on survey experience during the preventive trials, including savvy about census-taking, “cruising” of areas to complement the census, and about scheduling and organization of survey clinics. Further, we had the consultation of the “Michigan Marvel,” Les Kish, with unique skills in cluster sampling. We also  had Russell Luepker aboard to recruit for and direct the survey clinics, Ivan Frantz to assure accurate and stable chemistry laboratory values, Ron Prineas to assure standardized blood pressure measurement; Dick Crow, Prineas, me to assure standardized clinical and electrocardiographic data collection, and Rich Gillum, Ron Prineas, and me to provide models of trends and hypotheses of change and to prepare the proposals to NIH. We had an experienced staff of survey team trainers and supervisors residual from MRFIT days.

Ole Minnesota had The Right Stuff. And we were ready.

Our application to NIH to carry out the first survey and four-year surveillance in the Minnesota Heart Survey was submitted in spring of 1977 and reviewed by the EDC Study Section and its site visit team in the fall, right in the middle of my long-scheduled New Zealand Heart Foundation Lecture tour. Leaving the MHS site visit review in the capable hands of Ron Prineas and Rich Gillum as co-principals, I flew off to New Zealand, then proceeded, counter-current, around the world to Sydney, Perth, Johannesburg, Capetown, Rio, and home. Joseph Stokes, head of the NIH/EDC study section at the time and leader of its site visit team, was a long-term colleague who knew and understood what we were about. On the strength of his groups’ rigorous review and critique, we were able to revise and resubmit a proposal that was accepted and funded the following year. The study moved into two-year periods of surveys each quadrennium.

New sets of NIH reviewers over the years continued to find MHS methodology and its goals for detection, description, and explanation of cardiovascular disease trends both topical and worthwhile. MHS efforts continue, therefore, into the new millennium.


MHS has documented changes in rates and in the direction of trends for both in- and out-of-hospital deaths from coronary disease and stroke. It has shown strong correlations among disease trends and risk factor changes, and it has contributed to methods for national and international surveillance efforts undertaken since ours began (e.g. MONICA). Other efforts were stimulated by NIH conferences on the decline of coronary heart disease in 1978 and 1986. A major argument for continued support of this research has been on-going development of surveillance methodology, over and above measuring risk trends in Minnesota’s predominately white, middle-class population. In fact, Minnesota is in the lower third of the 50 states in cardiovascular disease death rates and among the top states in measures of good health.

But MHS has proved a bell-weather. It has gone on to work out difficulties in standardizing measures within and among surveys for such matters as blood pressure and blood lipid levels. It early broke the bad news that diagnosis-related hospital discharge codes (DRG), and their reimbursement, strongly influenced diagnostic patterns and apparent rates of specific cardiovascular disease diagnoses in hospital. We were among the first, along with the Mayo group in Rochester, to demonstrate the plateauing and then reversal of downward trends in death rates for stroke. We were early to document the relatively greater fall in out-of-hospital deaths than in total coronary deaths during the 1970s and 1980s, and similarly we were early to demonstrate a relatively greater fall of in-hospital deaths after the mid-1980s. These latter findings strongly challenged our earlier view that acute cardiac care could not strongly influence the population rate of CVD events. In fact, it can and does. We were early to demonstrate the backsliding in detection and control of hypertension in the community, despite a continued trend downward in mean blood pressure values in the population as a whole.

We worked hard to improve the methodology of our surveillance unit, along with procedural and mathematical ways of adjusting for drift in blood pressure and in laboratory values. And we added elements along the way to assess the impact of such new therapies as thrombolysis, coronary artery procedures and by-pass surgery, and changes in diet, health behavior, nutrient intake, and anti-oxidant serum levels. We early focused on ethnic minority populations and addressed the changing types and severity of coronary disease manifestations, particularly congestive heart failure.

Though we were clearly sanguine at the outset of MHS in the late 1970s, in what we thought could be learned from introducing this surveillance research strategy, and though the problems and complexity of surveillance continue to mount and evolve with the changing nature of disease and technology and privacy regulations, we believe that MHS research has strengthened the field. It has also surely advanced the skills and vitae of a whole generation of younger investigators at Minnesota (Prineas, Jacobs, Luepker, Gillum, Folsom, Pirie, Sprafka, Burch, Arnett, Shahar, McGovern, et alia).

Over the years, our frugality got us into hot water from carry-over funds. Red flags waved in Bethesda when our business manager bought a pack of computers for MHS staff at the end of a grant period. In punishment for our frugality, MHS funding was cut severely so that both effort and morale on the project were laid low for a time. Now they are quite recovered, and, needless to say, we no longer carry forward a penny.

Thus, we muddled through troubled years in MHS, and now, under youthful new management, MHS is presenting fresh ideas and sound plans for continued surveillance, with emphasis on researches in the rapidly changing picture of manifest disease, medical care, medical outcomes, and health behaviors in the community. The over-riding MHS idea today is that vascular diseases, and the influences on their mass occurrence, are in transition. We live during a technological revolution that not only reverses myocardial infarctions but also creates them, that reverses strokes and causes them, that leaves residuals of excess new and old phenomena such as chronic heart failure and arrhythmias, and where new concepts of disease and care stress our ability to standardize diagnoses or validate such simple issues as hospitalization rates for heart attack.

The MHS reapplication submitted in year 2000 stressed the recent trends found of a diminution in the rate of decline in CVD deaths and in the level and control of risk factors. Minnesota, through MHS and the Olmsted County Study, has become a sentinel for national trends demonstrated in the National Health and Nutrition Surveys (NHANES): faltering hypertension control and failure of health promotion to reach under-served groups. We think that event surveillance and research should go on in parallel with risk surveillance.

In sum, surveillance has historically been the foundation for successful public health efforts in communicable diseases. Today, the advantages of knowledge from monitoring for non-communicable disease trends are just beginning to be fully realized.

Click to go onto the next section.