The use of electronic health record data extraction for estimating influenza vaccination coverage and effectiveness
Influenza vaccination is the primary method for preventing influenza and is recommended annually for all individuals six months or older. Reliable and timely estimates of influenza vaccine effectiveness (VE) and uptake are necessary to better inform influenza prevention programs. Despite the importance of such data, limited vaccine uptake estimates are available in Australia, and reliable VE estimates are often not possible until post-season.
METHODS:
In 2013, the Western Australian Department of Health piloted a project to extract de-identified data from electronic outpatient records from general practices in Western Australia. Individual patient data from these practices were aggregated to estimate the proportion of the population immunised against seasonal influenza. Pathology results associated with these records were used to measure VE. Information extracted included the number of patients vaccinated, total number of patients within each practice, specimen collection date, vaccine brand, vaccination date, and laboratory testing results as they were returned to the general practice. A test negative design was used to calculate VE, based on the estimated odds (OR) of influenza infection (VE=1-OR).
RESULTS:
Preliminary analysis of 54 participating practices indicated that 18.8% of the population in Western Australia were immunised against influenza in 2013; 52.6% of adults ≥65 years, 52.2% of patients with coronary heart disease, 29.7% of those with asthma, and 47.9% of those with diabetes received a seasonal influenza vaccine. Results from 2,731 individual pathology tests indicated that influenza vaccination in 2013 was 56.0% effective in preventing infection (OR: 0.44 [95% CI: 0.24-0.79]).
CONCLUSIONS:
These preliminary findings indicate that systems which extract data from existing electronic outpatient health records can provide rapid estimates of influenza vaccination coverage and effectiveness. Such timely data could improve the public health response to influenza and public health prevention programs. Similar methods of data capture in other countries could also improve the surveillance of influenza globally.