Interfacing of Point-of-Care Urinalysis Results and Errors
Supplementary File 1: Data for point-of-care urinalysis results entered into electronic medical record either manually or by electronic interface. The urinalysis analyzer was Clinitek Status Plus (Siemens Healthineers, Cary, NC) using either Multistix 10 SG or Uristix 4 (Siemens Healthineers) urinalysis strips. [Tab A]: Overview of classes of errors and examples of errors seen in the data. [Tab B]: Interfaced data. Urine color and clarity are interpreted by the operator, with the choices selected within the urinalysis instrument software and then transmitted by interface to the EMR. The remaining urinalysis results are read by the instrument and directly interfaced with the EMR electronically. [Tab C]: “Manual Pre-simplification” – These results are interpreted by the staff performing the test and entered into the EMR by selecting options from a series of drop-down menus. These tests were performed between January 2019 and September 2019. In October 2019, we restricted the number of options available in the provided drop-down menus to simplify data reporting and improve consistency. [Tab D]: Manual – Simplification” – These data are similar to “manual pre-simplification” but occurred after the restriction in drop down menu options. [Tab E]: “Enter-Edit” – These results are interpreted by the staff performing the test and then manually entered into the EMR using free text without character restrictions. The following data (when available) were collected and found in Tabs B, C, D, and E: deidentified sample number, order description (either “Multistix 10”, “Uristix 4”, or “individual component"), age at time of testing (90 years or older reported as > 90), legal sex (male, female, or unknown), and urinalysis component data (color, clarity, glucose, bilirubin, ketones, specific gravity, blood, pH, protein, urobilinogen, nitrite, leukocyte esterase).
Steps to reproduce
The retrospective analysis period was January 1, 2019 to June 9, 2021. The project had approval from the University of Iowa Institutional Review Board. An electronic medical record (EMR) reporting program, known as Epic Reporting Workbench, was used to identify and retrieve data from the study as we have described in a previous report . All data was for patient data in the EMR from the University of Iowa Hospitals and Clinics (Iowa City, Iowa, United States). No data was obtained from EMRs of other hospitals. Urinalyses were performed using the Clinitek Status Plus (Siemens Healthineers, Cary, NC)  with one of the following test strips: Multistix 10 SG (Siemens Healthineers) – glucose, bilirubin, ketones, specific gravity, blood, pH, protein, urobilinogen, nitrite, and leukocyte esterase or Uristix 4 (Siemens Healthineers) – glucose, protein, nitrite, and leukocyte esterase. Data were extracted between January 1, 2019 and June 9, 2021 using a component search for results associated with point-of-care urinalysis that were manually input into the EMR using drop-down choices or interfaced from the instrument. Point-of-care urinalysis results that used Enter/Edit (free text entry of results) were retrieved by search for test results that included free text. “Typos” were defined as errors that did not change the interpretation of the data (spelling errors, single character substitutions, “++” instead of 2+, etc.). Errors “outside reportable range” were defined as values that fell outside of the reportable range defined in the EMR that were not due to simple typographical error. As an example, the reportable range for specific gravity was 1.005 to 1.030 for the methods used in the study. Thus, a value of 1.035 entered in the EMR would be outside the defined reportable range for the specific gravity component and should have been reported as “> 1.030”. “Uninterpretable” errors were defined as those which were not interpretable either because they were so severely misspelled or because they conveyed no clinical meaning (i.e., “normal” for a quantitative variable, “trace” for clarity, “increased” for urobilinogen, or a numeric value for a qualitative variable, etc.). Categories were mutually exclusive and thus no single error is represented more than once. References:  Z.J. Fleishhacker, P. Rastogi, S.R. Davis, D.R. Aman, C.S. Morris, R.L. Dyson, M.D. Krasowski, Impact of Interfacing Near Point of Care Clinical Chemistry and Hematology Analyzers at Urgent Care Clinics at an Academic Health System, J Pathol Inform 13 (2022) 100006. https://doi.org/10.1016/j.jpi.2022.100006.  Q. Zhang, X. Zong, P. Li, G. Sun, J. Lin, Performance Evaluation of Four Commercial Point-of-Care Testing Urine Analyzers for Diagnosing Urinary Tract Infection, Clin Lab 67 (2021). https://doi.org/10.7754/Clin.Lab.2021.210214.