How to improve monitoring and forecasting of disease patterns
"The answer, my friend, is blowin ' in lhe wind "

Throughout history, prophets and fortune tellers have pre- dicted the weather and disease. Recent scientific advances have made us less reliant on prophets for weather prediction, but our ability to forecast the weather has far outstripped our ability to predict occurrences of disease. Why is this so? Shouldn't we in health care strive to accomplish what those in weather forecasting have achieved: accurate monitoring and forecasting locally, nationally, and internationally?

To cite an illustration, last year's hurricane Andrew was the most destructive hurricane ever recorded in the United States. It cut a swath across some of the most densely populated areas of the country, destroying tens of thousands of houses and leaving 250 000 people homeless. (1,2) Despite the scale of destruction only 14 people were killed. Prophets might say that this was a miracle, but records show that during this century the number of deaths from hurricanes has fallen sharply. (1) The reasons lie in improvements in the weather forecasters' ability to monitor and predict the exact path, timing, and force of the hrricane. For Andrew, fore- casters were able to predict 10-14 days in advance where it would strike. In contrast, the Texas hurricane of 1900 struck with virtually no warning, resulting in 6000 deaths. During this century monitoring and forecasting the weather has become a science, permitting effective preventive measures. The meteorological model can provide us with many pointers to how we can improve the monitoring and forecasting of disease.

Classification systems and instruments - Instruments such as the thermometer and barometer enabled standardised, accurate, and cost effective measurements of weather con- ditions even in the most remote locations.(3,4) For diseases there are standardised and reasonably accurate classification schemes, such as the International Classification of Diseases, yet the current systems for monitoring health are either too inaccurate (for example, the communicable disease reporting systems), too costly (non-communicable disease screening or registries), or too slow (death certificates).(5-7)

Telecommunication systems - Part of the rationale for developing the global telegraph system in the mid-1800s was the transfer of information about the weather.(4) Using informa- tion supplied by these telegraphs, national weather bureaus provided timely forecasts to local areas. Interestingly, 150 years later we are just beginning to discuss and implement telecommunicaoon systems for monitoring diseases.(8)

Networkimg - Some of the first weather monitors were military doctors who reported weather conditions to the army's central weather centres.(4) A network of cooperative field stations developed for collecting standardised data on weather. National and intemational socieues sprang up. Currendy, people who count and forecast diseases are scattered across many different agencies with little communi- cation and no linking network.

Government recognition - Another reason for the success of meteorology was government recognition of the need for central weather bureaus to "provide useful warnings and information about various meteorological phenomena." Few governments have central bureaus dedicated exclusively to timely collection of standardized local and national data on all major diseases and rapid forecasting. National disease monitoring bureaus need to be established to perform similar functions to those of national weather bureaus.

Systems perspective - Forecasters view weather from a "systems" perspective, whereby patterns are examined, mapped, and used for forecasting. Thus an individual thunderstorm is important in the overall context of weather patterns and systems. In health care we have focused on individual diseases rather than taking a systems perspective. Nevertheless, the interrelations between diseases in a popula- tion follow systematic and predictable patterns.

Standardized approaches to mapping - In weather mapping, simple nomenclature and symbols evolved that the military, government, and local newspapers could all use. While weather maps in newspapers or on television help us to understand and predict the weather, such a standardised approach to portray disease has not yet evolved, probably because the data on morbidity have been so poor. Once better data are fed into the system, however, approaches to mapping will need to be established. Every day weather forecasters give us the odds of a storm. Shouldn't our disease forecasters every day give us the odds of an attack of hayfever, the odds of our children developing a cold and not going to school, or the odds of injury when driving to work on wet roads?

International cooperation - In 1873 the International Meteorological Organisation met in Vienna and agreed on standards. In health we have accomplished this for death certification and disease classification but for little else. An international society for disease monitoring and forecasting thus needs to be formed to set standards of monitoring morbidity.

User demand - Perhaps the most important reason for the rapid development of weather monitoring was that many diffcrent groups demanded weather forecasting for economic reasons. Farmers demanded weather forecasts because adverse weather could lead to economic doom if crops were not harvested before a hail storm. International maritime cooperation was stimulated with one of the first economic analyses, which showed that marine meteorological charts produced annual savings in the 1800s of between $40m and $60m worldwide. (9)

Airlines recognised the importance of accurate weather forecasts for saftey and profit. In contrast, governments, industry, and the public have not clamoured for better forecasting of disease as the economic importance of improved prediction may not be obvious. But, given the current costs of care, information concerning current and future disease patterns are critical for planning indiviudal and industrial strategy. Moreover, the accurate monitoring of disease is important for governments as they have to decide how to allocate their scant resources for health care.

We are fast approaching the twenty first century with techniques towards monitoring an dforecasting disease patterns that have changed little since the 1800s. These need revamping; the model of weather forecasting may be the the way forward.

Department of Epidemiology,
University of Pittsburgh
Pittsburgh, PA 15025

1 Facts on Files 1992; 27:629.
2 World Almanac. New York. Shripps Howard Company, 1993.
3 Whitnah DR. A history of the United States weather bureau.
Urbana: University of Illinois Press, 1961.
4 Hughes P. A century of weather service: a history of the birth
and growth of the national weather service, 1870-1970. New York
Gordon and Breach, Science Publichsers, 1970.
5 LaPorte RE, McCarty DJ, Bruno G, Tajima N, Baba S. Couting diabetes
in the next millennum: application of capture-recapture. Diabetes Care
1993; 16:528-34.
6 LaPorte RE, MCCarty DJ, Tull ES, Tajima N. Counting birds, bees
and NCDs. Lancet 1992; 339:494-5.
7 Directors of the WHO Non-Communicable Disease Collaborating Centres
and Key Officials. Needed: universal monitoing of all serious diseases
of global importance. AM J Public Health 1993; 83:941-3.
9 Birkhead G, Chorba TL, Root S, Klaucke DN, Gibbs NJ. Timeliness of
national reporting of communicable diseases: the experience of the
national electronic telecommuncations system for surveillance. Am J
Public Health 1991; 10:1313-5.