Yin P, Kiss A, Leis JA. Urinalysis Orders Among Patients Admitted to the General Medicine Service. JAMA Intern Med. 2015;175(10):1711-1713. doi:10.1001/jamainternmed.2015.4036
Urinalysis (UA) is a frequently ordered rapid screening test to exclude the presence of a urinary tract infection (UTI) among patients admitted to the general medicine (GM) service from the emergency department. Despite its excellent negative predictive value,1 a positive UA result is nonspecific because it occurs in as many as 90% of asymptomatic elderly patients.2,3 We hypothesized that overuse of UA in the emergency department contributes to overdiagnosis and excessive use of antibiotics for UTI among patients admitted to the GM service.
We conducted a prospective cohort study of consecutive adult patients to assess the appropriateness of UA orders on admission to the GM service of a large tertiary care center for 4 consecutive weeks in September to October 2014 and 3 consecutive weeks in January 2015. Each patient was assessed within 24 hours for indications for UA, including symptoms of UTI based on guidelines for patients with and without urinary catheters3,4 or acute kidney injury, defined as a 2-fold rise in serum creatinine levels.5 We recorded the frequency of empirical therapy for UTI, orders for urine culture (UC), and antimicrobial prescriptions based on UC results. We compared the proportion of patients who underwent UC or received antibiotic treatment (empirical or by UC result) using χ2 tests based on a positive or negative UA result. We assessed predictors of UA orders without indication (UTI or acute kidney injury) using a multivariable logistic regression model with 75 years or older, sex, residence in long-term care, diabetes mellitus, dementia, or presence of 3 or more comorbidities as variables. We obtained approval from the research ethics board of Sunnybrook Health Sciences Center. All data were deidentifed and informed consent was waived.
Of 403 GM patients (median age, 79 years; 212 [52.6%] women), 250 (62.0% [95% CI, 57.3%-66.7%]) underwent UA on admission at the discretion of the emergency department or GM physicians. Of these patients, 211 (84.4% [95% CI, 79.9%-88.9%]) lacked symptoms of UTI and 198 (79.2% [74.2%-84.2%]) lacked UTI and acute kidney injury. The presenting complaints of the 198 patients are listed in Table 1. For all patients undergoing UA, the frequency of UC orders and antimicrobial therapy for UTI (empirical or by UC result) are listed in Table 2. Positive UA results were associated with increased probability of UC orders (P < .001) and antibiotic prescription (P < .001) among asymptomatic patients. After controlling for other independent variables, only the presence of multiple comorbidities was significantly associated with UA orders without indication (odds ratio, 5.3 [95% CI, 2.5-11.0]: P < .001).
More than half of our patients underwent UA at admission although most lacked an appropriate clinical indication. Positive UA results from these asymptomatic patients significantly increased their probability of receiving additional low-value care, including UC and antibiotics for asymptomatic pyuria or bacteriuria. Conversely, appropriate UA orders among symptomatic patients were used effectively to exclude UTI and withhold antimicrobial therapy.
These findings highlight the harms of UA overuse in this patient population because positive UA results can introduce cognitive biases in favor of a UTI diagnosis even when patients lack accepted guideline-based criteria.6 After controlling for other variables, those most likely to undergo UA without an appropriate clinical indication had multiple comorbidities. One possible explanation is that patients with complex medical problems are more likely to undergo a broader net of investigations.
This patient cohort from a single academic center may not reflect the ordering patterns of other institutions. A more liberal definition of appropriateness of UA orders could have included tangentially related complaints, such as abdominal pain or fever without localizing symptoms; however, these additional indications accounted for the minority of presenting complaints in our cohort. The use of a guideline-based definition for UTI may have overestimated the proportion of asymptomatic patients, especially among those who could not communicate their symptoms reliably. However, the presence of dementia was not associated with increased UA orders among patients lacking this UTI definition. Limiting indiscriminate UA ordering has the potential to improve UC and antimicrobial prescribing practices among GM patients.
Corresponding Author: Jerome A. Leis, MD, MSc, Division of Infectious Diseases, Department of Medicine, Sunnybrook Health Sciences Centre, Room H463, 2075 Bayview Ave, Toronto, ON M4N 3M5, Canada (firstname.lastname@example.org).
Published Online: August 17, 2015. doi:10.1001/jamainternmed.2015.4036.
Author Contributions: Drs Yin and Leis had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Yin, Leis.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: Yin, Leis.
Statistical analysis: Kiss, Leis.
Study supervision: Leis.
Conflict of Interest Disclosures: None reported.