Evaluating the Timeliness and Completeness of Electronic Laboratory Reporting (ELR)

BACKGROUND: The Houston Department of Health and Human Services (HDHHS) collaborated with local hospitals and clinics to implement electronic laboratory reporting (ELR) using HL7 2.5.1 messaging.  Previous research studies assessed ELR utility and demonstrated mixed findings. The mixed results may be due to different disease incubation periods, various reporting volumes, and inconsistency of evaluation methods. This study aims to use a systematic approach and focuses on a single reportable condition, HIV, to evaluate the utility of ELR against the traditional reporting method.

METHODS: A large safety-net hospital system conducted parallel testing with the HDHHS in 2014 by sending surveillance data using two methods: ELR and CD-ROM. To measure timeliness, the number of days elapsed between specimen collection date and report date was calculated for both methods and was analyzed using Access and SAS. The completeness was evaluated at variable-level and dataset-level after joining the ELR dataset to the CD-ROM dataset by linkage variables. Variable-level completeness was measured by the proportion of complete data by test type, using the traditional method, CD-ROM, as the gold standard. Dataset-level completeness was measured by the number of records in ELR divided by the number of records on CD-ROM for each test type.

RESULTS: A total of 2,185 reportable records of HIV, representing 733 unique persons, were submitted to the HDHHS by the hospital between December 1^st and December 15^th, 2014 by a CD-ROM. Among these unique persons, 635 (87%) were found in the HDHHS disease surveillance system (Maven) by name and date of birth. There was 7.1 days of delay in reporting via CD-ROM, compared to 2.2 days via ELR. The variable-level completeness is 80% for HIV multispot, 51% for 4^th generation HIV enzyme-linked immunoassay, 90% for CD4 percent, and 74% for real-time polymerase chain reaction. The genotype results were excluded from the analysis because the data was not available on the CD-ROM.

CONCLUSIONS: Although ELR augments disease reporting, data completeness and accuracy need to be evaluated systematically. This study demonstrates a methodology to assess ELR completeness, timeliness, and accuracy. The methods can be used for other reportable conditions. The results will be used to promote and enhance ELR adoption in the region.