Volume 24-02
15 January 1998

[Table of Contents]

A SUMMARY OF THE 1996-1997 CANADIAN FLUWATCH PROGRAM

Background

FluWatch is a national surveillance project for influenzalike illness (ILI) that was piloted in 1995-1996, and became fully activated during the 1996-1997 influenza season. Prior to FluWatch, national influenza surveillance relied on aggregate laboratory data submitted to the Laboratory Centre for Disease Control (LCDC) from 21 laboratories across the country and case-by-case data from about one-half of these laboratories. LCDC also received isolates for virus characterization, and six provinces regularly reported data from their own surveillance programs. LCDC then compiled surveillance information and prepared monthly summaries for dissemination. The interpretation of influenza data at the national level was complicated; mechanisms used to measure influenza activity varied from province to province, and laboratory results were often delayed because of processing and reporting time.

FluWatch was developed to enhance the existing national influenza surveillance system by collecting consistent and timely national data. It is a collaborative project between the provinces and territories, College of Family Physicians of Canada (CFPC), sentinel physician reporting programs in British Columbia and Calgary, and LCDC.

Design

CFPC's National Research System (NaReS) recruited at least one physician from each of the 1991 census divisions across Canada. The exception was in British Columbia and the Calgary area where sentinel physicians were already involved in local surveillance programs. For one clinic day per week, between 1 October 1996 to mid-April 1997, physicians were asked to complete a report form with the number of patients seen and the number of patients meeting a standard definition for ILI. Both groups of patients were broken down by age category. Reports were either faxed or the information was conveyed via telephone to LCDC on a weekly basis. LCDC would then collate the data and prepare a report which would be distributed once every 2 weeks.

At the conclusion of its first full season, the FluWatch program was evaluated by participating physicians in order to measure whether it achieved stated goals, and to gauge the degree of user satisfaction with its design and implementation.

Results

Of the 290 census divisions across Canada, 273 had CFPC member physicians available for recruitment. Although 223 individual physicians reported at any time during the FluWatch season, on average 110 physicians (49%) participated in FluWatch on a weekly basis. The enlisted physicians were not equally distributed across the country. The percentage of census divisions by province and territory, with at least one physician reporting, ranged from 28% in Quebec to 100% in British Columbia and the Yukon Territory.

The physician response rate also varied between provinces (Table 1). For all of Canada, 41% of physicians submitted reports for at least 20 weeks (74%) of the FluWatch season.

Based on a separate evaluation questionnaire, the sentinel physicians saw an average of 32.8 patients per day during the FluWatch season. A total of 3,818 cases of ILI were diagnosed from 89,952 patients seen (42.5 per 1,000 patients seen).

The first peak in cases of ILI in Canada occurred during the Christmas holiday season. A second, smaller wave occurred mid-to late March. This trend is consistent with that observed in the laboratory-confirmed isolates that were reported to LCDC (Figure 1). This similarity remained after ILI rates were standardized to provincial populations. Overall, the greatest proportion of FluWatch cases occurred in the 20- to 44-year-old age group (33%), followed by those 45 to 64 years of age (19%). The largest rate of ILI was in the 0- to 19-year-old age group (50 per 1,000 patients seen).

Table 1 Breakdown of census divisions with at least one sentinel physician reporting and physician response rate, by province, FluWatch Program, 1996-1997

* excluding Calgary area sentinel physician reporting program

Province Total number of census divisions (1991 Census) Number of census divisions with NaReS physicians available for recruitment Number of census divisions with at least one physician reporting Response rate (% of physicians reporting a minimum of 20 weeks) Newfoundland 10 10 9 36   (4/11) Prince Edward Island 3 3 2 67     (2/3) Nova Scotia 18 17 12 31   (4/13) New Brunswick 15 15 8 42   (5/12) Quebec 99 94 26 20   (6/30) Ontario 49 49 43 43 (25/58) Manitoba 23 18 15 19   (3/16) Saskatchewan 18 15 10 73   (8/11) Alberta 19 19 13 33   (3/9)^* British Columbia 30 29 29 52 (29/56) Yukon 1 1 1 100     (2/2) Northwest Territories 5 3 2 0     (0/2) TOTAL 290 273 170   * excluding Calgary area sentinel physician reporting program

Reporting

During the influenza season, 15 FluWatch reports (one every 2 weeks) were prepared by LCDC. Each report included a map identifying the presence or absence of ILI activity within each census division for that 2-week period. The map was accompanied by graphics depicting rates of laboratory-confirmed isolates by province and territory and text describing local, national, and international influenza activity. The FluWatch report was disseminated to the sentinel physicians, NaReS representatives, federal and provincial epidemiologists and laboratories, the World Health Organization, and the United States Centers for Disease Control and Prevention. Other summary articles were prepared for the Canada Communicable Disease Report on a monthly basis, and information was also posted on the LCDC Website. The Website document received 700 to 800 hits per month.

Discussion

On average, 110 physicians participated in FluWatch on a weekly basis during the 1996-1997 influenza season. Over 76% of physicians who responded to the program evaluation questionnaire (121 of 132 physicians who received questionnaires responded) stated that they would be willing to participate again in the 1997-1998 season, and 84% found the case definition for ILI to be appropriate. Because physicians were either not available or not recruited in all census divisions, FluWatch data may not have been representative of influenza activity throughout Canada. The ability of FluWatch to provide consistent national data was also hampered by the variable response rate in some of the regions that did report. However, from those ILI reports that were submitted to LCDC, a bimodal pattern for the 1996-1997 influenza season was apparent. This trend was supported by the laboratory-confirmed data, evidence that FluWatch was an accurate indicator of national influenza activity (Figure 1).

Figure 1 Comparison of FluWatch ILI vs. laboratory-confirmed cases of influenza (IVR)* by reporting week, Canada, 26 October 1996-4 May 1997

* IVR = interactive voice response reporting system
SIRV = système de réponse vocale

Figure 2 Comparison of FluWatch ILI vs laboratory-confirmed cases of influenza (IVR)* by reporting week, Ontario, 26 October 1996-4 May 1997

* IVR = interactive voice response reporting system
SIRV = système de réponse vocale

When one considers national ILI and laboratory rates together, trends in activity appear to peak at the same time (Figure 1). However, when reviewing data from a province such as Ontario with regular reporting from a large number of census divisions, FluWatch rates did appear to anticipate the laboratory findings (Figure 2), in terms of peak activity.

The age distribution of ILI cases is not surprising when one considers that FluWatch captures the "walking population," i.e. people who visit a family physician's office or clinic. FluWatch does not capture children who visit pediatricians, emergency rooms, after-hours clinics, and the elderly in long-term care facilities; hence, the largest proportion of ILI cases were seen in the 20-to 44-year-old age group.

The majority of physicians (75%) who completed and submitted program evaluation questionnaires reported that the frequency of reports, issued once every 2 weeks, was acceptable, and stated that they liked the map. At the federal, provincial, and territorial Influenza Surveillance Meeting, held 5-6 June 1997, the provincial and territorial influenza surveillance representatives reported that they found the FluWatch text to be useful, but suggested that the map should be revised to increase its usefulness and ease of interpretation. The provincial and territorial representatives also requested that they receive the raw ILI data for their regions each week.

Conclusions

The similar trends observed between FluWatch and the laboratory data indicate the success of FluWatch as an indicator of true influenza activity. However, to provide consistent national data, FluWatch should be more representative of activity across the country. It is essential that all census divisions are represented. For the 1997-1998 season, the provinces are working with their local NaReS representatives to assist in the recruitment of physicians in all their census divisions. Efforts are also being made to include First Nations communities.

To ensure that FluWatch provides timely national data, methods to improve the weekly physician response rates have also been considered. These include simplifying the report form, faxing the report form out each week to act as a reminder, and providing special recognition to those physicians who do report regularly. It is also hoped that, by decentralizing parts of the program, local partnerships will develop between the physicians, public-health, and local NaReS representatives thereby fostering continued and regular participation.

Finally, the dissemination of the FluWatch data has been modified for next season. The provinces will receive the raw ILI data for their census divisions on a weekly basis. The FluWatch report will be prepared every 2 weeks, except during periods of elevated activity when it will be prepared on a weekly basis. The report will include a modified FluWatch map and information on local, national, and international influenza activity. The map will now reflect provincial and territorial levels of influenza activity. The influenza activity score will be assigned by the provincial or territorial influenza representative, and will take several sources of information into account: ILI data, laboratory data, other provincial and territorial data including school and workplace absentee rates, and any institutional outbreaks. The activity score will be a number between zero (no activity) and three (widespread activity). The FluWatch report will continue to be disseminated both nationally and internationally.

NOTE: Due to the success of the term "FluWatch," all future surveillance activities for influenza at the national level will be incorporated under this term.

Source:

M Litt, BScN, MHSc, Senior Epidemiologist, Health Programs Analysis Division, First Nations and Inuit Health Programs Directorate, Medical Services Branch, P Buck, DVM, MSc, Field Epidemiologist, J Hockin, MD, MSc, Director, Field Epidemiology Training Program, Bureau of Surveillance and Field Epidemiology, P  Sockett, PhD, Chief, Division of Disease Surveillance, Bureau of Infectious Diseases, LCDC, Ottawa.

[Previous] [Table of Contents] [Next]