Implementation of an algorithm for testing, diagnosis, and antibiotic stewardship of asymptomatic bacteriuria in pregnancy

Abstract

BACKGROUND

Asymptomatic bacteriuria affects 2% to 15% of pregnant women, with 20% to 40% developing symptoms later. Symptomatic urinary tract infections are more common in pregnancy, with a prevalence of 33%, posing risks, such as preterm delivery, low birthweight, and maternal pyelonephritis. The gold standard for urinary tract infection detection is a urine culture, but point-of-care urinalysis dipsticks are frequently performed as screens during regular obstetrical visits. Leukocyte esterase has been used to justify the treatment of asymptomatic bacteriuria, even with low sensitivity and specificity. Confirmatory tests are crucial for avoiding false positives and ensuring optimal outcomes. Current guidelines for urinalysis dipstick interpretation and the decision to treat asymptomatic bacteriuria in pregnancy are limited. It remains unclear whether an evidence-based algorithm can improve test utilization, diagnosis, and treatment decisions for asymptomatic bacteriuria in pregnancy.

OBJECTIVE

The primary objective of our study was to develop, implement, and evaluate an evidence-based algorithm to guide urinalysis interpretation, culture, diagnosis, and antibiotic stewardship of asymptomatic bacteriuria in pregnant patients during routine obstetric visits.

STUDY DESIGN

The project involves both retrospective and quasi-experimental prospective medical record reviews of pregnant patients aged ≥18 years, beyond 20 weeks of gestation, from routine obstetrical visits with urinalysis dipstick tests. A doctorate in clinical laboratory sciences student developed an educational algorithm to guide urinalysis dipstick interpretation, culturing necessity, and treatment decisions based on evidence-based practice. Our study considered patient records from February 1, 2022, to February 28, 2022, as retrospective (prealgorithm implementation) data and January 24, 2023, to February 22, 2023, as prospective (postalgorithm implementation) data. Data collected from the electronic medical record included deidentified patient information, urinalysis results, culture dates and outcomes, antibiotic prescriptions, urinary tract infection or asymptomatic bacteriuria diagnoses, provider details, adverse pregnancy outcomes, and demographics. Data analysis using SPSS (version 29; SPSS IBM, Armonk, NY) involved chi-square tests, likelihood ratios, and effect size calculations, with P values of <.05 considered statistically significant.

RESULTS

This study examined a total of 1176 patient records. Preimplementation data included 440 records, of which 224 were abnormal urinalyses and 216 were normal urinalyses. Postimplementation data included 736 records, of which 255 were abnormal urinalyses and 481 were normal urinalysis. The patient demographics predominantly featured White individuals (87%), with a median maternal age of 27 years and a gestational age of 32 weeks. Our preimplementation analyses revealed significant associations of algorithm deviations with both culture utilization (P<.001) and antibiotic stewardship (P<.001). However, no significant association was observed between algorithm deviations and adverse patient outcomes. Culture underutilization decreased significantly from 43.0% (189/440) before implementation to 29.5% (217/736) after implementation (P<.001). The overall reduction in asymptomatic bacteriuria prevalence from 16.3% (8/49) to 6.7% (10/67) suggests a decrease of nearly 60.0%. In addition, antibiotic overprescription decreased significantly from 1.6% (4/258) before implementation to 0.8% (4/522) after implementation (P=.003), with a reduction from 7.1% (3/42) to 2.4% (1/41) among abnormal urinalyses.

CONCLUSION

Our findings show a strong alignment between the use of the algorithm and subsequent clinical decisions, underscoring its potential to enhance patient care and management in obstetrical settings. Adherence to the algorithm was higher among providers displaying prudent antibiotic use.