Accuracy of Laboratory Data Communication on ICU Daily Rounds Using an Electronic Health Record

Abstract

Objectives: Accurately communicating patient data during daily ICU rounds is critically important since data provide the basis for clinical decision making. Despite its importance, high fidelity data communication during interprofessional ICU rounds is assumed, yet unproven. We created a robust but simple methodology to measure the prevalence of inaccurately communicated (misrepresented) data and to characterize data communication failures by type. We also assessed how commonly the rounding team detected data misrepresentation and whether data communication was impacted by environmental, human, and workflow factors.

Design: Direct observation of verbalized laboratory data during daily ICU rounds compared with data within the electronic health record and on presenters' paper prerounding notes.

Setting: Twenty-six-bed academic medical ICU with a well-established electronic health record.

Subjects: ICU rounds presenter (medical student or resident physician), interprofessional rounding team.

Interventions: None.

Measurements and main results: During 301 observed patient presentations including 4,945 audited laboratory results, presenters used a paper prerounding tool for 94.3% of presentations but tools contained only 78% of available electronic health record laboratory data. Ninty-six percent of patient presentations included at least one laboratory misrepresentation (mean, 6.3 per patient) and 38.9% of all audited laboratory data were inaccurately communicated. Most misrepresentation events were omissions. Only 7.8% of all laboratory misrepresentations were detected.

Conclusion: Despite a structured interprofessional rounding script and a well-established electronic health record, clinician laboratory data retrieval and communication during ICU rounds at our institution was poor, prone to omissions and inaccuracies, yet largely unrecognized by the rounding team. This highlights an important patient safety issue that is likely widely prevalent, yet underrecognized.

Conflict of interest statement

Dr. Artis disclosed other support from the Agency for Healthcare Research and Quality (AHRQ). Her institution received funding from the AHRQ. Dr. Mohan’s institution received funding from the AHRQ. Dr. Gold’s institution received funding from the AHRQ. Dr. Dyer has disclosed that he does not have any potential conflicts of interest.

Figures

Figure 1.

A, Types and frequencies of laboratory misrepresentation events. Observation of 301 patient rounds presentations and 4,549 laboratory data points yielded 1,886 inaccurately communicated laboratory test results which were further classified by type of misrepresentation. Most misrepresentation events were omissions. B, Frequency ICU team caught laboratory misrepresentation events varied by type of misrepresentation. Overall frequency of detection of laboratory misrepresentation events was low, with the exception of “pending” type misrepresentation events, which were detected at a significantly higher frequency compared with all other misrepresentation types.

Figure 2.

Accuracy of laboratory communication on rounds and ICU team detection of misrepresented laboratory data by individual laboratory test. We observed the communication of 4,549 laboratory data points from 26 selected domains on daily ICU rounds. Communication accuracy and detection of misrepresented laboratory data varied by individual laboratory test. alk phos = alkaline phosphatase, ALT = alanine aminotransferase, AST = aspartate aminotransferase, Bcx = blood culture, BUN = blood urea nitrogen, Cr = creatinine, Hb/Hct = hemoglobin or hematocrit, HCO3 = serum bicarbonate, Phos = phosphate, Plt = platelet count, PT/INR = prothrombin time or international normalized ratio, PTT/hep = partial thromboplastin time or heparin level, ScvO2 = central venous oxygen saturation, t bili = total bilirubin, trop = troponin.

Figure 3.

A, Correlation between frequency a laboratory test was ordered and accurate communication of the laboratory results. Pearson correlation showed more frequently ordered laboratory tests were more often accurately communicated on rounds. B, Correlation between frequency a laboratory test was present on the artifact and accurate communication of the laboratory results. Pearson correlation showed laboratory tests more commonly present on the artifact were more commonly accurately reported on rounds. Alk phos = alkaline phosphatase, ALT = alanine aminotransferase, AST = aspartate aminotransferase, BUN = blood urea nitrogen, Ca/iCa = calcium or ionized calcium, Cr = creatinine, Hb/Hct = hemoglobin or hematocrit, HCO3 = serum bicarbonate, Phos = phosphate, Plt = platelet count, PT/INR = prothrombin time or international normalized ratio, PTT = partial thromboplastin time or heparin level, ScvO2 = central venous oxygen saturation, T.Bili = total bilirubin.