Menu Back to Poster-Presentations-Details

PP12-88: Does Current Practice Reflect the Value of Rapid Diagnostics? US Patterns in the Time to Completion of Microbiology Results

Poster Presenter

Jennifer Nguyen

Director, Patient and Health Impact
Pfizer Inc
United States

Objectives

Timely microbiological diagnosis enables treatment optimization, may improve outcomes, and aids stewardship programs. This need led to a rise in rapid diagnostic technologies but delays in completion of testing are common in practice. This study examined factors that may contribute to such delays.

Method

This study used hospital data from 2015-2018 to examine time to test completion — defined as the time between specimen submission and when microbiology results are available — for complicated urinary tract infection (cUTI) and hospital-acquired pneumonia (HAP) due to gram-negative bacteria (GNB).

Results

There were 100,496 patients hospitalized in 215 different US hospitals with cUTI or HAP with available timestamped microbiology records, of which 94,306 had cUTI, 16,850 had HAP, and 10,660 had both cUTI and HAP. On average, testing results were available sooner for cUTI than for HAP in every US Census division, but there was considerable variability across geography (min and max mean difference for cUTI vs. HAP = 3.3 and 22.2 days). Time to test completion displayed a multi-modal distribution consisting of three peaks at ~42, ~60, and ~128 hours, which did not vary by hospital characteristics (teaching status, number of beds, urban/rural), infection type (cUTI/HAP) or type of GNB (K. pneumonia/E. coli/P. aeruginosa) but did vary by geography. Test results in the Mountain division lagged about 24 hours behind the other divisions, while the West North Central division had the earliest peak completion time. In contrast, human factors appeared to explain the multi-modal distribution for time to test completion. Graphical examination via a heat map showed distinct linear relationships such that incremental increases in the time of day of specimen submission were associated with shorter test completion times. Stated another way, there appeared to be time thresholds for test completion that did not depend on when the specimen was submitted, with one exception. If a specimen was submitted in the evening hours (7PM and later), the time to test completion was likely to increase by ~24 hours. Additionally, there were distinct diurnal patterns in specimen submission times and the times when results were available. Specimen submission times were normally distributed from 6AM-11PM, displayed a spike at 12AM, was lowest and uniform from 1AM-5AM, and most were timestamped on the hour, followed by the half hour. Microbiology result times were right-skewed with most results available from 5AM-1PM.

Conclusion

Despite a rise in rapid diagnostic tests that greatly speed up identification of pathogens and susceptibility to various antimicrobial drugs, time-saving microbiological diagnosis remains a key challenge for clinicians and patients. Standard techniques require ~48-72 hours for final results, while rapid tests have the potential to provide genotypic results in 5-15 hours and phenotypic results in 7-24 hours. In order to unlock the value of rapid diagnostics, new technologies must harmonize with health system factors, such as hospital workflow design and clinical microbiology laboratory schedules, to remove diagnostic delays. In this study, only 2.6% of cUTI and/or HAP patients had test results available in 24 hours or less, and only 0.8% had results in 6 hours or less. The real-world microbiology data also revealed a disproportionate number of specimen submissions timestamped at 12AM as compared to the hour before and after midnight. These timestamps may reflect workflow design, but when used to map the patient journey, may introduce measurement error of up to 24 hours per misreport. This could have implications for evaluating and interpreting patient metrics and outcomes with regards to timeliness of diagnostic testing. The consistent pattern of multiple thresholds for time to test completion observed across geography and various hospital characteristics suggests that health system factors play a key role in diagnostic delays. These delays may result in overuse of broad spectrum antibiotics, which can drive resistance, or delay appropriate treatment, which can increase morbidity, mortality, healthcare resource utilization and costs. Encouragingly, if health system factors are a significant driver of diagnostic test completion times, they are potentially modifiable and provide opportunities to integrate rapid diagnostic technologies into practice and realize their value in rapid time to completion of microbiology results to improve healthcare delivery.