Reflections on the Past and Challenges for the Future

Kerr L. White MD*

My association with the National Center for Health Statistics (NCHS) and later the U.S. National Committee on Vital and Health Statistics (NCVHS) goes back almost forty years when I first met Ossie Sagen, one the Center’s Associate Directors. Soon thereafter in the mid-1960s I encountered Forest Linder the first Director of the National Center for Health Statistics, Ted Woolsey, the Deputy Director, and I.W. Moriyama, an Associate Director and first Executive Secretary of the NCVHS. Ossie, also an Associate Director, was the initial statistical genius behind the whole enterprise. For the rest of his life, he was my closest friend. Together with his wife and mine, we visited frequently as well as travelling and vacationing on numerous occasions. Our initial encounter was followed by Ossie’s agreement to be the lead consultant for what became a large-scale seven-country collaborative survey of medical care use co-sponsored by the World Health Organization (WHO).(1) Ossie taught many of us a great deal about the theory and practice of health statistics, especially of surveys, as well about the NCHS. He spoke often about its origins, activities, and aspirations. These four leaders brought diverse talents to this ground-breaking institution; they were remarkable public servants each in his own way. To give you a flavor of their eminence and dedication you should know that Ossie turned down an appointment at Princeton University’s Institute for Advanced Studies in favor of public service. He thought the latter was a higher calling. Take heart all who labor at NCHS; you are doing the people’s work!

Here are several anecdotes that may be unknown to most members of the NCVHS and the NCHS’s current staff and others. The Center was the successor to the National Office of Vital Statistics and as such most of its staff was preoccupied with improving the country’s vital statistics systems. There is no gainsaying the importance of these but with the advent of Medicare and Medicaid in the early 1960s it became increasingly important to assess the needs and health care of the living in addition to counting the dead. Moriyama played a seminal role in developing the International Classification of Disease (ICD) and he understandably emphasized its central importance in the panoply of health statistics. There was little or no interest in developing better information about the country’s health services to say nothing about the outcomes of medical care in relation to the needs of individuals and populations. Hospital Statistics had attracted some sporadic attention from about 1955 but the emphasis was largely on the institutions’ overall activities than with their patients’ problems. The statistics described how busy they were but not what they were accomplishing by way of patient care.

In the mid-1960s, there were two or three small meetings at which several of us attempted to persuade Moriyama and others of the urgent need to develop information about the patients who use hospitals. This was essential, we argued, for policy-making, as well as for clinical and epidemiological studies, and eventually for the assessment of “outcomes”. To accomplish this required an appropriate clinical classification system for morbidity in contrast to mortality. For the most part the ICD provided codes for causes of death not for morbidity as experienced by hospitalized patients. We were unsuccessful at that time in persuading the NCVHS, most of the NCHS staff, or the WHO that theICD required expansion and modification for clinical applications. Nothing much happened for another decade. Vergil Slee of the Kellogg Foundation financed Commission on Professional and Hospital Activities (PAS) independently developed a clinical modification of the ICD, the ICD-9-CM. It was published by PAS in 1978.(2) In spite of the fact that Medicare and Medicaid were the largest financers of hospital care in the country NCHS had absolutely no money in its budget for creating or publishing ICD-9-CM. A year or two later this grievous deficiency was remedied but it illustrates our slow start in developing usable information about health services, their costs, and outcomes. We should be more far-sighted as we embark on statistical systems for the new millenium.

Vergil Slee’s Professional Activities Survey (PAS) was in use by many of the country’s hospitals but, in my view, it collected too much data and did not generate enough usable information. Accordingly, in 1969 Johns Hopkins University organized an international Airlie House conference on Hospital Discharge Abstracts. With ideas from the United Kingdom and Canada, as well as from this country we developed what became the first of several Uniform Minimum Data Sets. The first was for patients discharged from hospitals(3) and there were later conferences for patients receiving ambulatory care(4) and then long-term care.(5) These were not new ideas; Florence Nightingale outlined the scope of information to be gathered on hospitalized patients as the basis for generating statistics on each hospital’s accomplishments. Anyone interested in the history of health statistics should read what she advocated 130 years’ ago,(6)

The first survey mounted by the NCHS was the National Health Interview Survey, a major interest of Forest Linder. This was followed by the National Health and Nutrition Examination Survey for which Ossie Sagen did most of the original design. Both were important but what about that massive “black box” of medical care provided in doctors’ offices and outpatient clinics? The United States and the rest of the Western world were running massive medical education industries for future doctors with no information about the 85% of medical care provided to ambulatory patients. No business would contemplate such activities without a “marketing” survey; it was absurd.

In the 1950s, W.P.D Logan who was responsible for medical statistics in Great Britain initiated a series of annual reports on the content of patient encounters in the offices of a group of that country’s General Practitioners.(7) This landmark survey, however, was limited to a self-selected group of volunteers. Impressed with the information generated I took it upon myself, in a washroom of the old Health, Education, and Welfare building, to broach to Ted Woolsey the subject of a similar but properly designed survey of ambulatory care provided by office-based physicians in this country. He jumped at the idea and said that if he succeeded Forest Linder, who was soon to retire, as Director ofNCHS he would undertake such a survey and he did both. We had a contract at Johns Hopkins to design what in 1974 became the National Ambulatory Medical Care Survey (NAMCS).(8) It immediately became a “best-seller” with demand for the NAMCS reports far exceeding anything that NCHS had ever experienced.

In 1974, Senator Edward Kennedy was persuaded to have the Public Health Act amended to include two matters of some importance. For the first time, the NCVHS was made a statutory body with defined responsibilities. Prior to that time it had been a voluntary undertaking although most other industrialized countries had given their analogous committees legal authority. Second, the hearing resulted in a requirement that there be an annual report to Congress on the health of the American people.(9) This initiative had two purposes. First to disseminate and educate the population and its politicians on the country’s health status and our problems. Second, it was intended to improve coordination among the governments disparate data acquisition systems and the information generated. The NCHS staff, assisted by the NCVHS, labored mightily to produce this innovative annual report.

Publication of the initial volume of Health United States resulted in front page lead stories in the country’s major newspapers. I was visiting Dorothy Rice, Director of theNCHS at that time, when she received a call from the Office of the Secretary, David Matthews, of what was then the Department of Health, Education, and Welfare, relaying an urgent call from the White House asking in effect: “What the hell is this all about?” In spite of being duly forwarded up the governmental hierarchy and “signed off” at each appropriate level of the bureaucracy, no one seemed to have taken the trouble to read, let alone, assimilate the report. To this day, the annual report on the nation’s health and health services usually receives top billing by the press the day after its publication.

So much for the past. What about challenges for the future? My overall impression of the contemporary status of health information systems in the United States is that we have, as the New Yorker once put it, “roomful’s of data untouched by human thought!” or as Professor Finagle’s law states:

The information you have is not the information you want;

The information you want is not the information you need;

The information you need is not available from the present arrangements!

After all, it is wisdom we seek. One contribution to this idyllic state is knowledge; this in turn is based largely on information, which in turn is supported by data. The persistent preoccupation of the health statistics fraternity with collection and acquisition of “data” detracts from our enormous collective potential for generating information and even “intelligence” (in the military usage) about matters bearing on the health of individuals and populations. So what ideas might be considered? I suggest three that I believe deserve the characteristically diligent attention of the NCVHS.

First, as in the marketing of dog food, we need to know exactly what is consumed. Proposals for initiating or modifying health statistics systems, indicators, and related measures usually are generated by employees of governments, health services institutions, systems and related entities, and universities. This is all to the good but why not ask the users of the output from statistical endeavors about their requirements and priorities? Why not systematically survey decision-makers such as politicians, managers, and practitioners and their advisors at all levels of government and the health care establishment (both its private and public components) to specify:

  • The five measures that are or would be of greatest value to them in making the choices, decisions, and evaluations that confront them;
  • The accuracy required for each measure;
  • The timeliness required for each measure;
  • The frequency required for each measure;
  • For each measure provide an example that illustrates how it has been or could be of substantial importance in the respondent’s work.

Years’ ago we were in the process of planning such a survey in the WHO’s Division of Health Statistics when a change of personnel curtailed it. I believe it could be a useful exercise and would go a long way towards standardizing data acquisition instruments, empowering health statisticians, and building constituent support for this essential work. We may discover that it is better to be roughly right, timely, and useful than precisely accurate, late, and useless.

Secondly, we need to address the major paradigm shift that Western medicine will undergo in the 21st century. One example is the exponential growth of public interest in so-called alternative medicine and its practitioners. Belatedly this shift is being accompanied by intense scientific scrutiny and selected applications by orthodox western medicine. Of equal or greater importance is the long-overdue acceptance of the reality that the genesis of ill health and disease involves many factors other than germs and genes. In addition to nutritional, environmental, and occupational factors there are life situations and experiences that all too frequently predispose, precipitate, and perpetuate the manifestation of the patient’s underlying disorder.(10) For too long, aided and abetted by such ill-founded ideas as Diagnostic Related Groups (DRGs), the public, its health statisticians and, too often, many physicians have come to believe that diseases are “things”; they are reified artificially. We now seem to have a body-shop, nuts and bolts, pills and procedures model of medicine and health services. We should never forget as the late Sir Austin Bradford Hill, the doyen of medical statistics, warned us: “Health statistics represent people with the tears wiped off”,

For too long medicine and its statistical systems have relied on the outmoded 17th century reductionism view of the human condition. The new physics and quantum mechanics have changed the rest of scientific thinking. Now medicine must catch up.(11) Abstracting systems, surveys, nomenclatures, classifications, and codes should be revised and modified gradually to accommodate the realties of this new and broader paradigm.

Perhaps the best place to initiate the paradigm shift to the new broader view of health and disease is with the language of disease, the terms, and abstractions that patients use to express their distress and curiosity. The International Classification of Primary Care (ICPC)(12) recognizes that about sixty percent of encounters at the level of primary care can not be given a code provided by the International Classification of Disease (ICD). Patients at this level of care present themselves to caregivers and others with complaints, symptoms, feelings, problems, conditions, and questions in infinite varieties. They do not have what the medical profession calls “diseases”. “You don’t have ’it’ until I name ‘it’ “ commented one physician.

The ICPC has been widely adopted in Europe and elsewhere but not in the United States. I suppose this is because of the low regard in which primary care generally has been held by the dominant medical establishment and its political allies. This needs to change if the public and its politicians are to be provided with realistic information about the population’s health status and what is being done about it. For starters, the ICPC should be used in all health interview surveys and the NAMCS and whenever data about the patient’s or person’s initial complaint or problem are acquired. The persistent attempts to code all encounters with the health care system as “diseases” seriously distorts reality.

Politicians (special interest lobbying aside) and the public they serve primarily are interested in the extent to which the vast sums of money, time, and energy expended on health care improve the functional capacity, quality of life, and longevity of individuals and populations. It is end results that count, not how busy the health care enterprise is.

The third challenge I foresee is the need to continue the model reflected in the development of uniform minimum data sets and that is the standardization of many of the other terms and standards used in the collection, recording, and acquisition of health data and information. Such practices have been used for decades in accounting, commerce, finance, manufacturing, telecommunications, technology generally, and many scientific endeavors. It is time that the NCVHS and the NCHS, probably in collaboration withWHO, took on this task of setting national and international standards for terms, definitions, and related statistical conventions with provisions for their orderly periodic revision. It will not be easy and will require decades for completion but we should make a start. The alternative is increasing cacophony, confusion, and waste in the health information arena.

I am full of admiration for the remarkable strides made by the NCVHS and the NCHS in recent years and applaud their initiative in considering the outlines of a Health Statistics System for the 21st Century. I hope these gratuitous comments will contribute to your deliberations.


* Member Standing Committee on Public Health Records & Statistics 1966-73; Member USNCVHS 1974 and Chair 1975-80.

500 Crestwood Drive, #1410, Charlottesvile, VA 22903-4858; e-mail

(1)Kohn, R, White KL. Health Care: An International Study. London and New York: Oxford University Press, 1976.

(2) Commission on Professional and Hospital Activities. International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9CM). Ann Arbor, Michigan: Commission on Professional and Hospital Activities, 1978.

(3)Murnaghan, JH, White, KL (eds.). Hospital Discharge Data: Report of the Conference on Hospital Discharge Abstract Systems, June 1969. Philadelphia, PA: JB Lippincott Company, 1970, reprinted from Med Care 1970; 8 (4): (suppl.)

(4) Murnaghan, JH (ed.) Ambulatory Medical Care Data: Report of the Conference on Ambulatory Medical Care Records, April 1972. Philadelphia, PA: JB Lippincott Company, 1973, reprinted from Med Care 1973; 11 (2): (suppl.).

(5)Murnaghan, JH (ed.) Long-term Care Data: Report of the Conference on Long-term Health Care Data, May 1975. Philadelphia, PA: JB Lippincott Company, 1976, reprinted from Med Care 1976; 14 (5): (suppl.).

(6)Nightingale, F. Notes on Hospitals 3rd ed. London: Longman, Green, & Roberts, 1863.

(7)Logan, WPD. General Practitioners’ Records: Analysis of the clinical records during the period April 1951 to March 1952. London: Her Majesty’s Printing Office, 1953.

(8)Tenney,JB, White KL, Williamson, JW. National Ambulatory Medical Care Statistics: Background and Methodology, United States 1967-1972. Washington DC: US Department of Health. Education and Welfare, National Center for Health Statistics. DHEW Publication No. (HRA) 74-1335, 1975.

(9) Public Health Service Act, PL 93-353. Section 306, July 1974.

(10)White, KL. Viewpoint: Fundamental Research at the Primary Care Level. Lancet 2000; (In press)

(11)Pauli, HG, White, KL. Scientific Thinking, Medical Thinking and Medical Education: Questions about their Evolution in the 20th Century. Human Resources for Health Development Journal 1998; 2: (3) 156-167

(12)Lamberts, H, Wood, M, Hofmans-Okkes, I. The International Classification of Primary Care in the European Community. Oxford and New York: Oxford University Press, 1993