Comparison of Electronic Laboratory Reporting and Laboratory-Based Manual Reporting — Pennsylvania, 2007–2014

BACKGROUND: Standards-based electronic laboratory reporting (ELR) system has been in place since 2005 through Pennsylvania’s electronic disease surveillance system (PA-NEDSS); however, it was not available for most reportable conditions until 2007. Laboratories that do not participate in ELR must key-enter data manually into PA-NEDSS. Although laboratories undergo a rigorous ELR onboarding process, the quality of ELR data in Pennsylvania has not been systematically evaluated and routine feedback is not provided to laboratories. We compared the timeliness and completeness of ELR data with laboratory-based manual entry reports (non-ELR).

METHODS: We extracted ELR and non-ELR reports from 2007 through early December 2014 from PA-NEDSS relational databases. Lead, tuberculosis, sexually transmitted infection and human immunodeficiency virus reports were excluded. Odds ratios and confidence intervals were used to assess associations between missing data and report type (non-ELR versus ELR) by specimen source, specimen collected and received dates, test completed date, and patient address. Timeliness (number of days to report) was calculated by subtracting test result date from the date the report was received by PA-NEDSS.

RESULTS: From January 1, 2007 through December 2, 2014, a total of 1,074,521 laboratory reports were received by PA-NEDSS; of these, 531824 (49.5%) were received via ELR. The proportion of reports received by ELR has increased over time from a low of 6.2% in 2007 to a high of 66.2% in 2014 with a simultaneous 400% increase in total laboratory reports received (45,527 to 227,798). In 2007, two laboratories used ELR for at least some conditions; in 2014 the number of laboratories participating in ELR had increased to 22. Non-ELR reports were more likely to have missing data for specimen source (16.33% versus 4.10%; OR: 4.57; 95% CI: 4.50–4.64), specimen collected date (7.43% versus 0.46%; OR 17.19; 95% CI: 16.50–17.91), laboratory received date (15.49% versus 0.15%; OR: 124.99; 95% CI: 116.47–134.14), and test completed date (12.37% versus 0.18%; OR: 80.41; 95% CI: 75.37–85.79); however, ELR reports were more likely to have missing address data (1.23% versus 4.83%; OR: 0.25; 95% CI: 0.24–0.25). ELR reports were more timely compared with non-ELR reports (median: 1 and 2 days respectively).

CONCLUSIONS: Automation of reporting through ELR has resulted in more complete and timely disease reporting in Pennsylvania. However, despite rigorous ELR onboarding, missing data, particularly for patient address, remains an area of concern. Laboratories should receive standardized feedback to assure the quality of laboratory reporting data.