It’s Surveillance, Surveillance, Surveillance

 

This photograph of Ancel Keys showing the rising heart disease crude death rate in Minneapolis was taken by Wayne Bell for the Minneapolis Sunday Tribune of November 30, 1947.

 

Several prominent citizens of Rome drop dead in the streets during a brief period in the 17th Century. The Pope, fearing the Almighty’s displeasure about Roman morals, commissions Giovanni Maria Lancisi to document and explain the terrifying phenomenon. (Lancisi 1971)

A century or so later, John Snow, anesthetist to the several accouchements of Queen Victoria, confronts epidemic cholera in London and is commissioned to explain the phenomenon—then attributed to bad air. Snow, of course, finds that it was bad water. He counts cases, identifies the different pumps and water providers for victim and non-victim alike, and, as told in that classic of epidemiology, “Snow on Cholera,” we see how he wraps up the source of infection as the polluted Thames and activates successfully for its public health implications. (Snow 1936)

Still another century later, clinicians and pathologists on the continent, internists and pathologists in Scandinavia, a cardiologist in London, a physiologist in Minneapolis, and a public health authority in Washington, D.C. independently recognize the post-war epidemic of heart attacks, the essence of an epidemiological transition. For the Twin City physiologist, it is a repeat of the Roman phenomenon: prominent Minnesota citizens, captains of industry and deans of the University, are dying unexpectedly in the streets. To the Europeans, it is an opposite phenomenon that commands attention, heart attacks disappear from their clinics and autopsy tables during the privations of World War II, and then reemerge vigorously in the immediate post-war years.

Thus, surveillance of illness, and vital statistics on causes of death, often thought a weak link among medical practice, epidemiology, and the public health, can have the greatest portent. History shows how clinical impression, when confirmed by systematic case-counting, with good numerators imposed on sound denominators, can play a decisive role in social awareness and the public health: detection and definition of the epidemic; sounding the alarm, inducing investigation, enhancing understanding, and effecting a societal response: disease control and prevention.

Yet epidemiological surveillance tends to be derogated among the medical elite and the ordinary practitioner, as illustrated by the shocking comment of the scholarly Lewis Thomas who once called epidemiology “mere leafing through death certificates” (personal communication), and by the general skepticism in which vital statistics are held by those who fill out the death certificates and bear their uncertainty; doctors.

Despite this nattering negativism, one finds at every turn how invaluable is surveillance. The post-World War II events in the streets led Ancel Keys to seek crude local mortality data to illustrate the implications for theMinnesota scene in 1947. That awareness led to his organization of the Minnesota Business and Professional Men Study, the first CVD prospective study among a cohort. The lessons of that inadequate study brought about, in the mid-1950s, the first systematic cross-cultural observational study of CVD, the Seven Countries Study.

Surveillance evidence on U.K. deaths sent Jeremy Morris, the young cardiologist turned social epidemiologist, back to the pathology laboratory of London Hospital to systematically compile 40 years of autopsy reports on ischemic heart deaths, confirming the dip in ischemic cardiomyopathy during the War and its rise post-war. Morris’s findings and his international influence led to two of the more notable contributions to CVD epidemiology. In the 1960s, the multi-center WHO Myocardial Infarction Registers, the first systematic “hot pursuit” of coronary deaths, brought together hospital and community data on incidence and cardiac care. This preparation led, in turn, to development of wider interest and skills throughout Europe and the mounting of MONICA, the multi-national program of cardiovascular disease surveillance that characterized WHO efforts of the 1980s and 1990s, effecting a new world-wide sophistication in epidemiology.

Systematic surveillance of heart attack trends came to the fore in the U.S. in the mid-‘70s, just before and after the bombshell of Weldon Walker’s New England Journal editorial: “Changing United States lifestyle and declining vascular mortality: cause or coincidence?” (Walker 1977) This article aroused consciousnesses that the United States was in the midst of a major decline in coronary heart disease death rates. It took Walker, a hard-headed observant practitioner, to point out the falling mortality rates to all the epidemiologists and vital statisticians who had, in essence, missed or downplayed it. In fact, however, it was a young public health officer, Nemat Borhani, and colleagues working in the State Health Department of California, who first published the significant downward trend in cardiovascular disease deaths starting among California women in the early 1960s. The California group failed to get credit for this observation, and it may be that Bohani was never asked why he neglected to ballyhoo the primacy of this California finding. [It was uncharacteristic, in that Borhani was so without reserve in all other aspects of his life and career!]

Action at the Center

Once alerted, the science bureaucrats at NIH mobilized vigorously and called a major “Decline Conference” in Bethesda in 1978, inviting researchers and health authorities from around the world to document the epidemic and its trends and inspiring national and international surveillance in search of explanations. In fact, in 1977, the Minnesota Heart Study (MHS) had already proposed regional surveillance to NIH as a methodological research program. A critical first part of the 2-part Minnesota community strategy was to document cardiovascular disease conditions and trends in the 7-county Greater Twin Cities, followed by studies of community health promotion and measuring their impact.

Minnesota had the community experience in various components of such surveillance to launch a program not only as surveillance but as research on surveillance methods [that study is in its 6th round of surveys and accumulation of hospital morbidity and mortality data over 26 years.]

Henry Blackburn outlined in a memoir the Minnesota picture at the outset of MHS:

“We had worked in the Twin Cities’ population from the 1940s through the preventive trials of the ‘70s. We had some experience with cardiovascular survey methods, having written a standard text in the field, and therefore presumably knew something about the principles and methods of such undertakings. The Laboratory of Physiological Hygiene had considerable hands-on survey experience from the Seven Countries Study, the U.S. Railroad Study, and during recruitment for prevention trials. This included savvy about census taking, “cruising” the area to complement an outdated census, and scheduling and organizing survey contacts and clinics. It had experience in design and analysis and had consultation from the more knowledgeable person on cluster sampling, Les Kish from Michigan. It had strong faculty to direct survey clinics, to assure stable laboratory values; faculty to train and standardize blood pressure measurement and to assure standardized clinical and electrocardiographic data collection. It had an experienced staff of survey team trainers and supervisors from prevention trial days. And it had leadership to provide models for the trends and hypotheses of change, to prepare the proposals, and to direct the study. Moreover, the job clearly needed to be done both for new knowledge about trends and their causes and for improved methods. We were ready for it. We thought that Minnesota had ‘the right stuff.’ (Blackburn 2004)

MHS is ongoing, with added components of research among ethnic minorities, youth, and older ages, and monitoring for stroke and heart failure as well as for coronary events.

At that moment in history, with the mortality decline in evidence and the firm charge of the Bethesda Decline Conference to mount CVD surveillance in the Institute’s program, Paul Leaverton came to head the Epidemiology Branch at NHLBI. He realized that a pilot study was necessary to demonstrate feasibility of an effective regional surveillance and so initiated the Community Cardiovascular Surveillance Program (CCSP), with four pilot field centers involving pioneer epidemiological researchers: Kuller at Pittsburgh, Tyroler at Chapel Hill, Borhani at Davis, and Labarthe in Houston, with Canner as coordinator at Maryland. The administrative difficulties of the feasibility study, along with uncertainties about the data acquired, led “Bethesda Central,” despite the confidence of the investigators and consultants to the study, to turn from primary surveillance to a new and combined “Framingham-type” cohort study with parallel hospital surveillance of CVD incidence, the ARIC study (Atherosclerosis Research in Communities).

Meanwhile, other morbidity-mortality surveillance got under way with different strategies in Massachusetts and in Olmsted County, Minnesota, while the Minnesota Heart Study in the Twin Cities was implemented in late 1979. They asked these sorts of questions about the nature and causes of CVD trends:

• Do the downward trends underway in heart and vascular death rates occur in both sexes, young and old, and in all ethnic and social subgroups?
• Are the trends greater or less for sudden, out-of-hospital deaths (attributable to public health phenomena), or for deaths in-hospital (mainly medical care effects)?
• Are the trends in mortality rates paralleled by those of attack rates, hospitalization, or incident events?
• Are the death rates associated with changes in population levels of single and multiple combined risk factors and if so, what are the lag times between average risk factor change and change in death rates?
• How variable are the slopes of mortality rates?
• Are changes in disease rates and risk factor levels in the community associated with health behavior changes such as smoking, nutrient intake, dietary patterns, or physical activity?
• Is the basic severity of the underlying coronary or cerebral atherosclerosis or the clinical events modified along with the risk factor changes?
• Is there a change in severity of the attacks measured as the level of cardiac enzymes produced, the proportion of events manifest as sudden death, or heart failure or arrhythmias?
• What is the proportionate contribution to mortality trends of changes in behavior and risk levels versus application of cardiac care innovations?
• What is the proportionate contribution to mortality trends of changes in behavior and risk levels versus application of cardiac care innovations?
• Do specific medical interventions (for example bypass surgery, angioplasty and stents, thrombolysis, or aspirin and beta blocker use) affect in-hospital mortality rates and subsequent survival?

The Decline Conference also led to the appointment of Manning Feinlieb as director of the National Center for Health Statistics and strengthening of its research components and staffing. Its surveillance and serial national health surveys, plus in-house reports from NHLBI statistician T. Thom, provide still much of the U.S. data and impetus to CDC and state health control and prevention programs. NCHS and the SEER cancer monitoring centers provide the basic data on non-communicable diseases for such public health undertakings as the anti-tobacco trials, innovative regional disease mapping, and original analyses and modeling of mortality change in this country and abroad. (Epstein and Thom 1994)

And so we arrive at summer 2007, with this distressed memorandum from colleagues in the National Center for Health Statistics:

As of yet, the budget shortfall has not been realized for NCHS (National Center for Health Statistics). Our budget problems impact all our data systems, not just Vitals [Vital Statistics]. However, both the House and Senate Appropriations Committees have recently recommended increases in our budget for Fiscal Year 08 by an amount which will not make us well, but at least will keep us alive for another year. Whether the resulting appropriations both for Health and Labor once sent to the President will be signed is another matter.

If help does not come, then we are faced with a bunch of bad options throughout NCHS, such as:

• Taking surveys out of the field; or
• Drastically reducing sample size of surveys; or
• Collecting all or a portion of vitals.

For vitals I think we would just try to delay the purchase of vitals from the states, which would mean our reporting would become even later and data quality will suffer; state vital registration activities will be hurt, and this in turn would have an impact on the NDI (National Death Index). But not as bad as missing a year or a portion of a year.

I might add that NCHS is not the only activity in CDC and HHS that is faced with some difficult budget options so it is not that we are being singled out; these are just very difficult budget times for many fine programs. (Charlie Rothwell, CDC/CCHIS/NCHS)

[ed. Epidemiology is # 3. We try harder!]

References

Blackburn, H. 2004. It isn’t always fun. Privately published.

Lancisi, G. M. 1971. Translation of De subitaneis mortibus (On sudden deaths) trans. A. V. Boursy and P. D. White. New York: St. John’s University Press.

Snow, J. 1936. Snow on cholera, being a reprint of two papers. New York: The Commonwealth Fund.

Thom, T.J. and F. H. Epstein. 1994. Heart disease, cancer and stroke mortality trends and their relationships: An international perspective. Circulation 90: 574-582.

Walker W.J. 1977. Changing United States lifestyle and declining vascular mortality: Cause or coincidence? NEJM 297(3): 163-165.