As the visitor proceeds through the program entering data regarding symptoms, food items eaten, and food sources, increasingly specific comparisons with other users’ data are available. The objective is to help each visitor determine what he might have in common with other persons, including symptoms, time of illness onset, and consumption of the same food item from the same food source.

Figure 2

Figure 2. Comparison report displays each risk factor the visitor reported during his target period along with a comparison period.

The summary report is a univariate descriptive analysis showing the number and percentage of past visitors who reported the same risk factors (foods, food sources, nonfood exposures) as the current visitor, who can use the summary report to select individual reports for viewing. A comparison report analyzes the visitor’s risk factors during an adjustable “target” period of onset dates, compared to an adjustable historic “comparison” period. Subsequent retrievals can be restricted on the basis of risk factors, and data can be viewed as a case report or output in a format accessible by most spreadsheet, database, and statistical programs. Figure 2 shows an example of the comparison report.

Visitors may choose to enter their data and leave the investigation to the health departments to determine whether a cluster exists. They may also view a descriptive table of recent reports and see whether the source they originally suspected was mentioned by other Forum visitors. Persons who are satisfied that they are not part of an outbreak may leave the Forum without requesting further output. Those who see common exposures may pursue more sophisticated output. At any time in the process, Forum visitors can request the aid of their health department.

The Forum does not investigate outbreaks or replace the current systems used by local health departments. Rather, the Forum increases reporting of foodborne illness and makes identifying suspicious clusters that may warrant further investigation possible. By collecting numerous variables, the Forum delivers information on a large number of risk factors to local health departments to assist in an investigation.

A moderator views each report after it is entered and conducts appropriate follow-up. For example, if a report from outside the three-county pilot area is entered, the moderator will alert the proper health department that a report from their jurisdiction has been entered. Like most Web-based forums, the moderator also reviews data for reasonableness, profanity, or other infringements of posted Forum rules. Records can be excluded from analysis, and each health department may similarly reject records.

RUsick2 visitors cannot retrieve other visitors’ personal identifiers, narrative testimony, or names of restaurants, stores, or other food sources. In contrast, health departments have access to all data fields for visitors who report being residents of their jurisdiction. Local health departments in the three-county pilot area around Lansing, Michigan, were involved in developing the Forum and have helped design methods by which local health departments can monitor the Forum. During the pilot phase, local health departments have had password access to the database to check reports from their residents. After expansion of the Forum, counties with less activity will be able to request automatic email notification from the Forum’s computer.

Several features of the Forum are designed to prevent phantom outbreaks caused by the power of suggestion. Visitors cannot implicate any particular food item or exposure. Second, visitors enter data before being given the opportunity to view output that might influence their own reports. (While visitors may modify their responses on subsequent visits to the Forum, very few do.) Third, each visitor only views output for food sources, foods, and other risk factors they have already indicated in earlier data entry screens. Fourth, food stores, restaurants, and other food sources are identified to Forum visitors only by abbreviations, which may not be specific to particular establishments. Abbreviations are sufficient to identify suspicious clusters, but only health departments see the entire names of commercial establishments. Finally, suspicious clusters must be investigated by the local health department to determine if clusters are due to foodborne outbreaks, chance, confounding, pranks, or normal changes in diet.

The Forum is being pilot tested in the tri-county area of Clinton, Eaton, and Ingham counties, which make up the Greater Lansing, Michigan area. The Forum was implemented in November 2002, involving three local health departments: the Barry-Eaton District Health Department, the Ingham County Health Department, and the Mid-Michigan District Health Department. For comparison purposes, we evaluated previous foodborne illness reports from the population of these counties for the years 2000 and 2001 (4–6).

Table 2 depicts the Michigan Department of Agriculture data collected for 2000 and 2001 by year, month, and county. From these data, we predicted that approximately 22 reports would have been expected during the comparable months of 2002 and 2003. The RUsick2 data are displayed by month and county in Table 3 and show that 93 reports were obtained with the Web-based system.

From November 2002 (the time the Forum was implemented) until February 2003, a total of 93 reports to the RUsick2 Forum reached at least the entry level of reporting (which begins by identifying foods consumed). In the first 17 weeks of the program, an average of 5.37 cases were reported each week. Based on the previous years’ reports, we calculated an expected number of 1.31 cases per week; hence, the ratio of reported cases to expected reports (5.37/1.31) was 4.10.

Figure 3

Figure 3. Comparison of foodborne disease reports from Ingham, Eaton, and Clinton Counties before and after implementing the RUsick2 Forum; MDA, Michigan Department of Agriculture.

Figure 3 shows the weekly average of foodborne complaints reported to the Forum from the three pilot counties during the first 17 weeks of operation. Also shown on this graph is the weekly average number of foodborne disease complaints reported to the state of Michigan during the corresponding months of January, February, November, and December 2000–2001. Approximately 22 reports would have been expected during the 17-week period, based on reports from previous years. However, 93 reports were received during the first 17 weeks of operation, more than a fourfold increase in reporting.

Local health departments monitored reports to the Forum and contacted reporting persons by telephone or email to verify the accuracy of the report and authenticate the complaint. Contacts usually required a few minutes. Reports delivered from the Forum to local health departments were treated as traditional reports. The Forum identified two foodborne outbreaks that would likely not have been identified. One fictitious report was identified and rejected.

Similar to other Web-based forums, RUsick2 allows persons with a common health problem to examine one another’s information. In an attempt to determine if they share any common exposures, visitors can view risk factors such as food histories, food sources, and other exposures entered by previous visitors. The Rusick2 Forum acts as a “front end” to existing surveillance by increasing reporting and identifying suspicious clusters that warrant a full investigation.

The Forum has recently been adapted for national usage so that the technical aspects of data entry and information retrieval will function identically for residents of all states. The publicity campaign conducted in the pilot counties will be too expensive to reproduce on a national level, so the Forum will only gain national prominence if local health departments and consumer advocacy groups use and publicize the Web site. The input screens of this program are easily modifiable or removable. For example, the symptoms screen can be altered to include symptoms specific to a non-foodborne disease, or the food history screens can be deleted entirely. Thus, this program could be adapted for use in other disease outbreaks.