Syndromic surveillance for emerging infections in office practice using billing data

Syndromic surveillance in primary care has not been widely tested. Existing systems often rely on emergency departments or hospitals. Primary care settings offer early detection potential. Prior research has shown emergency departments can detect outbreaks. No prior work had resolved whether office-based systems are viable. This gap motivated a study of billing data as a tool. Billing data are routinely collected and may include symptoms. This paper's contribution is a low-cost method using ICD-9 codes.

The study aimed to test if billing data could support syndromic surveillance. A primary care practice was selected for a 12-month trial. The goal was to assess feasibility and cost. Syndromic surveillance typically uses emergency data. This approach used office-based billing instead. The researchers wanted to compare patterns with emergency data. They also evaluated staff effort and system costs. The motivation was to expand surveillance beyond hospitals.

A custom program converted billing data into daily summaries. ICD-9 codes were de-identified and grouped by age and sex. A staff member prepared and emailed the summaries. The system used CDC's Early Aberration Reporting System. Syndrome thresholds were calculated using standard deviations. Respiratory, gastrointestinal, and fever syndromes were tracked. Data from emergency departments were used for comparison. The system cost $1,500 to develop and took 1-2 minutes daily.

The system reported an average of 253 ICD-9 codes per day. Respiratory illness was the most common recorded syndrome. Gastrointestinal illness and fever followed in frequency. Lymphadenitis, rash, and fever exceeded thresholds most. During the 2003-2004 flu season, emergency data matched state trends. Primary care data showed less consistency due to single clinic variation. However, spikes in the practice occurred before emergency data. This suggests earlier detection of influenza in the office setting.

The system proved feasible for office-based surveillance. Low cost and minimal staff time were key advantages. Syndromic patterns varied between practice and emergency data. The researchers propose further development is needed. They suggest including primary care in surveillance systems. The method may allow earlier detection of outbreaks. Authors state the system can be implemented with existing billing data. They recommend addressing remaining technical and logistical issues.