Establishing goals, techniques and priorities for national communicable disease surveillance.

Which communicable diseases should be routinely monitored? How should they be monitored? Why are we monitoring them at all? The national Advisory Committee on Epidemiology set up a subcommittee to ponder these questions in 1987. The conclusions this committee reached led to recommendations on communicable disease surveillance which were subsequently endorsed by the Advisory Committee on Epidemiology. These recommendations, summarized in this article, will have a significant impact on microbiologists and clinicians in this country both from the standpoint of their work-load and data they may have available for the study of communicable diseases.

routinely monitored? How should they be monitored? Why are we monitoring them at all? The national Advisory Committee on Epidemiology set up a subcommittee to ponder these questions in 1987. The con clusions this committee reached led to recom mendations on communicable disease surveillance which were subsequently endorsed by the Advisory Committee on Epidemiology. These recommendations, summarized in this article, will have a significant impact on microbiologists and clinicians in this country both from the standpoint of their workload and data they may have available for the study of communicable diseases.

GOALS
The first requiremen t was to answer, Why are we doing this? Without goals, one cannot make decisions on priorities or methods. The committee established the followin g goals: • To facilitate th e prevention and control of the disease u nder surveillance by identifying: prevailing incidence levels, impacts and trends for the development of feasible objectives for prevention and control of the disease and th e evaluation of control programs; epidemiologic patterns and risk factors associated with the disease to assist in the development of intervention strategies; and outbreaks for timely investigation and control.
• To satisfY the needs of government (especially regulatory programs). health care professionals, voluntary agencies and the public for information on risk patterns and trends in the occurrence of communicable diseases.
It was necessary to define the quality requirements of the surveillance system if these goals were to be met. The committee determined these requirements to be as follows : • The use of a uniform case definition across Canada and definition of a preventable case, if applicable. • The collection of sufficient, appropriate epidemiologic data on a case to fulfill the goals and identifY preventable cases. • The timely transmission of these data from local to provincial and national agencies for analysis and timely reporting of analysis to clinicians and public health officials. • The use of data to enhance disease control programs and to assist in the development of realistic objectives for reducing the number of preventable cases. • Evaluation of the surveillance system every five years for effectiveness, cost and progress towards control of the disease .

TECHNIQUES
A 'generic' methodology was developed that met these goals and quality standards. This consisted of n otification of cases to local public health authorities, collection of data on the case, and transmission of these data on a case-by-case basis to provincial and n ational surveillance agencies. As each disease was examined individually, the 'generic' method was altered to suit the particular characteristics of the disease. Not all diseases can be effectively and efficiently kept under surveillance by a case-by-case notification system. Other methods are available for keeping track of a disease. One can also use laboratory-based information, hospital reports, sentinel physicians, epidemiologic investigation, mortality data, special surveys, or a combination of these.

PRIORITIES
After the committee had decided on the goals and methods of communicable disease surveillance, the question of what diseases warranted surveillance was addressed. This was done by developing 12 criteria that contribute to the importance of a disease from a national disease surveillance perspective (using the goals previously defined). Each disease was given from zero to five points for each criterion depending on how important that criterion was for that disease. The sum of the points gave the estimate of the importance of that disease for national surveillance purposes.
Some may wonder why all infectious diseases of general interest are not nationally notifiable. First, there isn't enough mon ey, time or energy for health care providers, local health units, provinces, territories or Health and Welfare Canada to report and collect data on every communicable disease. Second, to make a disease nationally notifiable means every provin ce and territory needs to go throu gh the legis lative or regu latory process of making it reportable in their jurisdictions. That takes time and effort! The disease must be important enough to b e worth this effort. Having priorities helps us decide where to put our greatest effort.
The following are the criteria used to determine the importance of a disease: WHO interest: The World Health Organization (WHO) is interested in all communicable diseases to some degree; however, ch olera, plague and yellow fever are subject to the 1983 International Health Regulations. Canada must collect and report information about cases of these diseases as an international duty. Five points were awarded to each of these d iseases. Three points were Agriculture Canada interest: Agriculture Canada and federal food regu latory agen cies spend su bstantial amoun ts of money on inspecting anim als, food production facilities and products. Human surveillance helps to monitor the effect of such work. The committee gave five points to diseases directly prevented by this work and two for diseases which have been or could be prevented.
Incidence: The committee classified the disease incidence rates into quintiles and gave one through five points to each disease depending on the quintile of the rate into which its incidence fell. Zero points were given if no cases were being reported. Quintiles were used s ince the incidence rates of infectious diseases cover a wide range, from zero for yellow fever to 174 per 100,000 population for gonorrh ea, and the rates are not normally distributed.
Morbidity: For morbidity, the committee assigned points as above, by qu in tile, based on h ospital days and short term disability. If data were not available, the disease was subjectively compared to those with data.
Mortality: The committee used the actual number of deaths recorded by Statistics Canada and divided the diseases into quintiles assigning points as above. The number of deaths varied from less than one annually to 443 reported for the acquired immunodeficiency syndrome ( Table 2 for definition of 'Disease Rank' Case fatality rate: Because some organisms may cause few deaths but have a high case fatality rate for the cases that do occur, the committee rated the death/case ratio by quintiles , with the highest being 100% for rabies.

Communicability and potential for outbreaks:
These were rated more subjectively based on the literature and experience of subcommittee members. AIDS illustrates the difference between the two criteria, as it rated two out of five points for communicability, but four out of five points in potential for an outbreak.
Socioeconomic impact: For this criterion, the committee consid ered factors such as the cost of immunization and food inspection programs, nonhospital health care and long term disability. Economic impact analyses have not been prepared for many communica ble diseases, so subjective comparison was often n ecessary.
Public perception of risk: This was even more subjectively based on subcommittee members' experience and media interest. It included the following factors: •  Necessity for an immediate public health response: This criterion evaluated the need for health officials to act immediately upon learning of a case of the disease in order to prevent further transmission. It measured the effectiveness of immediate case and contact management measures and was based on the members' knowledge and subjective comparison between diseases.
The committee did not give weights to the various criteria-in other words, give more importance to one criterion compared to another. Perhaps weighting will be done in the future as experience is gained with this approach. A second set of criteria would have to be developed for assigning weights to the first. Table 1 illustrates the points scored in the 12 categories for measles, AIDS and tuberculosis, which were ranked as the top three diseases by the present system.
Looking at all 60 diseases which have been evaluated (Figure 1, Table 2). the diseases notifiable in 1988 are distinguished on the basis of points scored from those added in 1989 to 1991 and from those considered but not recommended for addition.
What is important in this exercise is not the exact number of points given to a disease, but rather the relative position of the disease to others, and the cutoff point chosen. The committee decided to use a cutoff of 18 points to recom-40 mend a disease for national surveillance (Table 2). This requires diseases to rank somewhat higher to initiate than to continue surveillance. The system has worked well for diseases proposed in 1989-91 and provides a methodology to judge additional disease suggestions in the future.
Disease surveillance is much more than collecting numbers. It also means compilation, consolidation and analysis of data, and the development of conclusions. Then and only then will it lead to better prevention and control of the diseases being counted in the first place. The provincial epidemiologists and the LCDC are attempting to meet this challenge by developing standardized case definitions for the notifiable diseases (in final draft, to be published this year) and instituting case-by-case disease reporting to the federal level. With everyone's cooperation, valuable information should be readily available to help in the prevention and control of communicable disease in Canada.
If anyone would like a copy of the detailed ratings given to each disease or the case definition booklet, please write to the author at the Laboratory Centre for Disease Control, Health and Welfare Canada, Tunney's Pasture, Ottawa, Ontario KlA OL2.