Abstract
OBJECTIVE: Evaluate the prevalence and risk factors for inappropriately ordered urine cultures in primary care. DESIGN: Cross-sectional study using chart reviews. SETTING: Two primary care, safety-net clinics in Houston, Texas. PATIENTS: Non-pregnant adults without a urinary catheter who had a urine culture and a primary care visit between 11/2018 and 3/2020. METHODS: We classified patients lacking physician documentation of the following symptoms as having an inappropriately ordered urine culture: dysuria, frequency, urgency, hematuria, fever, chills, costovertebral angle tenderness, nephrolithiasis, and pain (suprapubic, pelvic, or flank). We extracted patient demographics, visit-related diagnostic codes, past medical history, and urine culture results. Diagnostic codes were grouped based on body system, visit type (e.g. routine visit), or sign or symptom clusters. We evaluated the relationship between these factors and inappropriately ordered cultures using generalized estimating equations logistic regression. RESULTS: We included 807 patients who had 870 visits. Most patients were Hispanic (66.3%) or African American/Black (24.8%) females (76.1%) with a median age of 50 years. Among 870 cultures, 210 (24%) were ordered inappropriately. We found having an abnormal urinalysis or urine characteristic (adjusted odds ratio (aOR): 13.66), acute low back pain (aOR: 4.88), a cardiovascular-related (aOR: 1.68) or gynecological/family planning visit (aOR: 10.84), being evaluated at the non-teaching clinic (aOR: 6.03), or having a routine health visit (aOR: 1.81) within the non-teaching clinic (interaction aOR: 4.27) were significantly associated with inappropriate urine cultures. CONCLUSIONS: Our study revealed factors associated with inappropriately ordered urine cultures that may be unique to ambulatory settings and can help design outpatient diagnostic stewardship interventions.