PHLT 8270 Week 9 Lab Interoperability Analysis Example

Valid Messages, Unusable Data: An Interoperability Analysis of Three Reporting Choices That Pass Every Technical Check Yet Force Hours of Work on a Composite State's Surveillance Analysts

Student Name

Walden University

PHLT 8270: Health Informatics and Surveillance

Instructor Name

Month Day, Year

Valid Messages, Unusable Data: An Interoperability Analysis of Three Reporting Choices That Pass Every Technical Check Yet Force Hours of Work on a Composite State's Surveillance Analysts

The Problem

Every electronic laboratory report the composite state receives for carbapenem-resistant organisms passes the checks built into its message gateway: the message is well formed, required segments are present and the sending laboratory is recognized. Yet the state's analysts spend hours each week cleaning these reports before they can be used. This analysis follows three reporting choices, each permitted by the message standard, to the work each forces on the people downstream. Dixon et al. (2013) note that structural validity does not ensure that electronic data are usable for surveillance, and these three choices show how the gap arises. All details are composite.

Choice 1: How the Carbapenemase Result Is Coded

A laboratory can send the carbapenemase result as a separate observation with a standard code that names what was tested, as a generic observation that says only that a carbapenemase test was done, or as a comment attached to the susceptibility results with no code at all. All three are allowed. In the composite state, about half the laboratories use a specific code, a quarter use a generic code with the gene or enzyme named only in the result text, and a quarter use comments.

The Work It Forces

For the coded half, the result flows directly into the case record. For the rest, analysts must read text to determine what was tested and what was found. A comment saying "carbapenemase producer detected by phenotypic method" and another saying "KPC gene detected by PCR" carry different information about the mechanism, and only a human reading both can tell. When a comment is dropped by the laboratory's interface, as described in the Week 3 data flow map, the result never arrives.

Choice 2: Results Carried as Text

Even where a result has a code, its value may be sent as text rather than as a coded answer. Values arriving in the composite state include "POS," "positive," "Detected," "KPC POSITIVE," "Carbapenemase present" and "see comment." Each is valid in the message.

The Work It Forces

Analysts maintain a lookup table of more than 60 text values mapped to positive, negative or indeterminate, and every new laboratory or software update adds variants. Unmapped values default to unknown, which, as earlier weeks showed, means no containment response. Coded answers drawn from a standard vocabulary would remove this work entirely.

Choice 3: Facility Identifiers and Addresses in One Line

The ordering facility, and for nursing home residents the facility where the patient lives, can be sent as structured fields or as a single text line combining name and address. Many laboratories send one line, such as "Riverbend Care Ctr 1400 Mill Rd Westfield."

The Work It Forces

The containment response depends on knowing which facility to call and whether the patient lives in a nursing home. Analysts must match free-text facility lines to the state's facility registry by hand, resolving abbreviations, misspellings and facilities that changed names. Matching errors send containment calls to the wrong place, and unmatched records leave facility type blank, the weakest field identified in the Week 5 appraisal.

Why These Choices Persist

Laboratories make these choices for reasons that make sense locally. Their information systems were built for clinicians, who read comments easily. Building specific codes and structured fields for public health reporting costs money and staff time, and the message standard does not require them. Electronic reporting has nonetheless improved surveillance dramatically compared with paper; in one comparison, it captured 4.4 times as many cases and arrived about eight days sooner (Overhage et al., 2008). What remains is a question of form: data now reach the state, but too often in shapes a computer cannot read.

Estimating the Cost of the Work

The time analysts spend on these three choices can be estimated. In the composite state, analysts reported spending about six hours a week reading comments and free-text results, updating the lookup table and matching facility lines, roughly 300 hours a year. That time comes from the same staff who would otherwise review clusters and support facilities. Each report that arrives with a specific code, a coded result and structured facility fields requires no manual handling, so the savings from improving the largest reporters would return to surveillance work quickly. The estimate also gives the program a concrete figure to present when asking laboratories and vendors to change their messages.

A Note on Electronic Case Reporting

Electronic case reporting, in which clinical systems send case information directly from the electronic health record, could eventually supply some of what laboratory messages lack, such as where the patient lives. It does not remove the need for well-coded laboratory results, because the carbapenemase result originates in the laboratory, and case reports drawn from clinical records often reproduce the same free-text values. The program should treat the two streams as complements and apply the same coding expectations to both. Doing so now avoids repeating, in a new reporting stream, the problems this analysis traced in the old one. Each fix described here also makes the data easier to share with partners under the governance rules set out in Week 8. Cleaner data are safer data to share.

What the Analysis Recommends

The state should publish an implementation guide that specifies codes for carbapenemase observations, a coded answer list for results and structured fields for facility identifiers, then work with the laboratories that send the most reports to adopt them first. The message gateway should flag, not reject, messages that use comments or free text for these elements, so analysts know which to review. Whether to mandate a single coding standard is a harder question: mandating one improves consistency but may delay reporting from laboratories that cannot comply quickly. A phased requirement, with technical assistance, balances the two. Because confirmed carbapenemase results trigger containment of the kind that has stopped spread elsewhere (Woodworth et al., 2018), the result field deserves priority over all others.

References

Dixon, B. E., Siegel, J. A., Oemig, T. V., & Grannis, S. J. (2013). Electronic health information quality challenges and interventions to improve public health surveillance data and practice. Public Health Reports, 128(6), 546-553. https://doi.org/10.1177/003335491312800614

Overhage, J. M., Grannis, S., & McDonald, C. J. (2008). A comparison of the completeness and timeliness of automated electronic laboratory reporting and spontaneous reporting of notifiable conditions. American Journal of Public Health, 98(2), 344-350. https://doi.org/10.2105/AJPH.2006.092700

Woodworth, K. R., Walters, M. S., Weiner, L. M., Edwards, J., Brown, A. C., Huang, J. Y., Malik, S., Slayton, R. B., Paul, P., Capers, C., Kainer, M. A., Wilde, N., Shugart, A., Mahon, G., Kallen, A. J., Patel, J., McDonald, L. C., Srinivasan, A., Craig, M., & Cardo, D. M. (2018). Vital signs: Containment of novel multidrug-resistant organisms and resistance mechanisms, United States, 2006-2017. MMWR. Morbidity and Mortality Weekly Report, 67(13), 396-401. https://doi.org/10.15585/mmwr.mm6713e1