Systematic Utilization of a Stool Color Card to Improve the Accuracy of Inpatient Melena Identification on Physical Exam in Post-Graduate Training

Melena is often misidentified by resident physicians early in their careers; in addition, rectal exam (RE) completion rates are low among residents, contributing to inaccurate gastrointestinal (GI) consultations for melena. Precise melena identification can have important diagnostic and therapeutic implications. We designed a stool color card (SCC; Figure 1) and prospectively examined whether SCC implementation, as part of residency training, will improve percentage of REs completed, accurate melena identification, and RE concordance between residents and GI consultants. A 1-year prospective study at a tertiary care academic hospital in which internal medicine and emergency medicine residents were oriented to utilize the SCC as part of the evaluation of all inpatient cases of GI bleeding (overt or occult) and anemia. Findings were compared with a control cohort from the previous year via a retrospective analysis. Data collection focused on documented clinical history and RE, subspecialty, and melena concordance rate (CR) on exam, and subsequently correlated with the appropriateness of proton pump inhibitor (PPI) use. Statistical significance was analyzed with Fisher’s exact probability testing. Of the 576 patients examined, 338 (183 [57%] control cohort and 153 [60%] SCC cohort) had documented REs and were included for analysis. The RE completion rates between the control and SCC cohorts were resident physicians 90% vs 76% (P = .001), emergency medicine 95% vs 100% (P = .08), internal medicine 72% vs 41% (P = .004), and GI 55% vs 67% (P = .03). Sixty-three (20%) and 59 (23%) had REs by both resident and GI in the control and SCC cohort, respectively. Of cases with 2 REs, the CR improved from 52% to 70% after SCC implementation compared with control (odds ratio = 2.1; P = .04). The prevalence of utilization of SCC is 11%. Of 37 melena patients, RC increased from 38% (control) to 57% (SCC) cohort (P = .31). Among patients with melena on oral iron, RC increased from 0% to 100% (P = .008). In the SCC cohort, those with concordant REs were more likely to have appropriate use of PPI than those with discordant exams 69% vs 44% (P = .07). The systematic utilization of SCC is associated with increased RE concordance between residents and GIs. SCC use also increased the frequency of RE performed by GI consultants, but its impact on resident physicians is subspecialty dependent. Melena concordance did not significantly improve with SCC application. However, there is a role in SCC for patients with melena on iron. Overall PPI appropriateness was not associated with utilization of SCC; however, a trend toward more appropriate PPI use was seen among cases with concordant exams. The overall direct uptake of SCC utilization is low; however, the systematic education and focus on rectal exam accuracy likely contributed to improved rectal exam quality. Further education or more stringent protocols may improve uptake of SCC and accuracy of melena identification.