Remediation of residents is a common problem and requires organized, goal-directed efforts to solve. The Council of Emergency Medicine Residency Directors (CORD) has created a task force to identify best practices for remediation and to develop guidelines for resident remediation. Faculty members of CORD volunteered to participate in periodic meetings, organized discussions and literature reviews to develop overall guidelines for resident remediation and in a collaborative authorship of this article identifying best practices for remediation. The task force recommends that residency programs: 1. Make efforts to understand the challenges of remediation, and recognize that the goal is successful correction of deficits, but that some deficits are not remediable. 2. Make efforts aimed at early identification of residents requiring remediation. 3. Create objective, achievable goals for remediation and maintain strict adherence to the terms of those plans, including planning for resolution when setting goals for remediation. 4. Involve the institution's Graduate Medical Education Committee (GMEC) early in remediation to assist with planning, obtaining resources, and documentation. 5. Involve appropriate faculty and educate those faculty into the role and terms of the specific remediation plan. 6. Ensure appropriate documentation of all stages of remediation. Resident remediation is frequently necessary and specific steps may be taken to justify, document, facilitate, and objectify the remediation process. Best practices for each step are identified and reported by the task force.
The Cuban national Integrated Medical Emergeny System or "Sistema Integrado de Urgencias Medicas" (SIUM) was formed in 1997. In 1998, the SIUM began an active out-of-hospital thrombolysis program using Heberkinasa, the only streptokinase obtained through recombinant DNA techniques, produced by the Cuban Center for Genetic Engineering and Biotechnology. An active community training program has also been implemented, standardizing training for the almost 20,000 members of the national emergency medical services.
The objective was to compare the image quality, diagnostic accuracy, radiation exposure, and contrast volume of "triple rule-out" (TRO) computed tomography (CT) to other diagnostic modalities commonly used to evaluate patients with nontraumatic chest pain (dedicated coronary, pulmonary embolism [PE], and aortic dissection CT; invasive coronary angiography; and nuclear stress testing). Four electronic databases were searched, along with reference lists and contacted content experts, for relevant studies from inception until October 2012. Eligible studies enrolled patients with nontraumatic chest pain, shortness of breath, suspected acute coronary syndrome (ACS), PE, or aortic dissection; used at least 64-slice CT technology; and compared TRO CT to another diagnostic modality. Eleven studies enrolling 3,539 patients (791 TRO and 2,748 non-TRO) were included (one randomized controlled trial and 10 observational). There was no significant difference in image quality between TRO and dedicated CT scans. TRO CT had the following pooled diagnostic accuracy estimates for coronary artery disease: sensitivity of 94.3% (95% confidence interval [CI] = 89.1% to 97.5%), specificity of 97.4% (95% CI = 96.1% to 98.4%), positive likelihood ratio (LR+) of 17.71 (95% CI = 3.92 to 79.96), and negative likelihood ratio (LR–) of 0.08 (95% CI = 0.02 to 0.27). There were insufficient numbers of patients with PE or aortic dissection to generate diagnostic accuracy estimates for these conditions. Use of TRO CT involved greater radiation exposure (mean difference [MD] = 4.84 mSv, 95% CI = 1.65 to 8.04 mSv) and contrast exposure (MD = 38.0 mL, 95% CI = 28.1 to 48.0 mL) compared to non-TRO CT patients. Triple rule-out CT is highly accurate for detecting coronary artery disease. Given the low (<1%) prevalence of PE and aortic dissection in the included studies, and the increased radiation and contrast exposure, there are insufficient data to recommend use of TRO CT in the diagnosis of these conditions. Comparar la calidad de imagen, la certeza diagnóstica, la exposición a radiación y el volumen de contraste de la tomografía computarizada (TC) para "triple descarte" (TD) con otras modalidades diagnósticas utilizadas frecuentemente para evaluar a los pacientes con dolor torácico no traumático (TC dirigida a coronarias, embolismo pulmonar (EP) y disección aórtica; angiografía coronaria invasiva; pruebas de estrés isotópicas). Se buscaron los estudios relevantes publicados hasta octubre de 2012 en cuatro bases de datos electrónicas, en las listas de la bibliografía y en los expertos de contenido contactados. Los estudios elegibles incluyeron pacientes con dolor torácico no traumático, disnea, sospecha de síndrome coronario agudo, EP o disección aórtica; utilizaron TC con tecnología de al menos 64 cortes; y compararon la TC TD con otra modalidad diagnóstica. Se incluyeron once estudios (uno aleatorizado y 10 observacionales) con 3.599 pacientes (791 TD y 2.748 no TD). No hubo diferencias significativas en la calidad de la imagen entre TC TD y los dirigidos. La TC TD tuvo las siguientes estimaciones de certeza diagnóstica para la enfermedad de las arterias coronarias: sensibilidad 94,3% (intervalo de confianza [IC] 95% = 89,1% a 97,5%), especificidad 97,4% (IC 95% = 96,1% a 98,4%), razón de probabilidad positiva 17,71 (IC 95% = 3,92 a 79,96), y razón de probabilidad negativa 0,08 (IC 95% = 0,02 a 0,27). No hay suficiente número de pacientes con EP o disección aórtica para generar estimaciones de certeza diagnóstica para estas enfermedades. La utilización de la TC TD supuso una mayor exposición a la radiación (diferencia de la media [DM] 4,84 mSv, IC 95% = 1,65 a 8.04 mSv) y al contraste (DM 38,0 mL, IC 95% = 28,1 a 48,0 mL) en comparación con los pacientes con una TC no TD. La TC TD es altamente certera para detectar la enfermedad coronaria. Dada la baja prevalencia (<1%) de EP y de disección aórtica en los estudios incluidos, y el incremento de radiación y exposición a contraste, no hay suficientes datos para recomendar el uso de TC TD en el diagnóstico de estas dos enfermedades. Chest pain is the second most common reason patients present to emergency departments (EDs) across the United States and accounts for over 6 million annual visits.1 Information obtained from the history, physical examination, electrocardiogram (ECG), and cardiac biomarkers is often insufficient for clinicians to safely distinguish patients who require further testing or hospital admission from those who can be safely discharged from the ED. This leads to high rates of testing and negative inpatient cardiac evaluations in low-risk patients,2 with admission rates as high as 96% reported in a recent investigation.3 "Triple rule-out" (TRO) coronary computed tomography (CT) angiography has recently emerged as a technology that noninvasively evaluates the coronary arteries and simultaneously visualizes the pulmonary arteries, thoracic aorta, and other intrathoracic structures. TRO CT, which has potential to identify both coronary and other life-threatening etiologies of chest pain such as coronary stenosis, pulmonary embolism (PE), and aortic dissection, is emerging as a diagnostic modality in some clinical settings for patients at low to moderate risk for acute coronary syndromes (ACS) in whom PE or aortic dissection are also being considered in the differential diagnosis. To perform TRO CT, either ECG-gated 64-slice multidetector or dual-source CT technology is required.4 Dual-source CT uses two x-ray tubes and two detectors arranged at 90° angles, allowing reconstruction of cross-sectional images at one-quarter of the gantry rotation time and improving the temporal resolution and diagnostic image quality of coronary artery examinations without requiring preexamination β-blockade.5 Currently, there is a paucity of data comparing the performance of the TRO CT to other diagnostic imaging modalities commonly used to evaluate patients with chest pain such as nuclear perfusion imaging, coronary CT angiography, or dedicated PE or aortic dissection CT. As TRO CT becomes increasingly available as a diagnostic imaging modality to evaluate patients with chest pain, it will be critical for emergency physicians to know its potential utility in practice. The objective of this study was to compare the image quality, diagnostic accuracy, radiation exposure, and contrast volume of TRO CT to other diagnostic modalities commonly used to evaluate patients with nontraumatic chest pain. This was a systematic review and meta-analysis, and it adheres to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) as applicable to diagnostic accuracy reviews.6 We included original research studies that enrolled adults with nontraumatic chest pain; shortness of breath; or symptoms suggestive of ACS, PE, or aortic dissection that used 64-slice CT technology (or greater) to compare TRO CT to another diagnostic modality. Studies that enrolled patients under 18 years of age or trauma patients were excluded. An expert librarian (PJE) designed a comprehensive search strategy with input from the clinical lead author (EPH). The electronic search included Ovid MEDLINE, Ovid EMBASE, Web of Science, and Scopus from inception until October 2012 (see Data Supplement S1, available as supporting information in the online version of this paper, for the MEDLINE search strategy). We made adjustments to the search strategy to account for differences in indexing between databases. Web of Science and Scopus depend heavily on text words, so acronyms were included in the strategies. We did not apply a language restriction to the search strategy. We also reviewed the "related citations" section of PubMed and reference lists of included studies and contacted a content expert (JEH) to identify additional articles for review. Two investigators (DA, MFB) independently screened the titles and abstracts of all records identified from the search strategy (phase I). If either reviewer thought the study might be eligible, we obtained the full report. The same two investigators then independently assessed the eligibility of each full report (phase II). We used Cohen's unweighted kappa to measure chance corrected agreement between reviewers for each phase of study selection. Any disagreements were discussed with the clinical lead author (EPH) and resolved by consensus. Data regarding study quality were abstracted for each study by one author (MFB). We assessed the quality of randomized clinical trials with the Cochrane Collaboration's tool for assessing risk of bias.7 The quality of case–control and cohort studies was assessed with the Newcastle-Ottawa quality assessment scale,8 and the quality of studies of diagnostic accuracy with the revised Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool.9 One author (DA) extracted data from each included article using a standardized data extraction form. Data were reviewed for accuracy by another author (MFB). We extracted the following data from each study: study design, patient demographics, imaging technology and parameters, and the clinical setting. Data were also extracted for both TRO CT and control study populations to assess diagnostic accuracy, image quality, radiation exposure, contrast volume, imaging time, cost, length of stay (LOS), and admission rate. When data were not sufficiently reported, we contacted the corresponding study author by e-mail twice over a period of 2 weeks to acquire missing information. Image quality was assessed using different scales across the studies. Therefore, we standardized these scores by dividing the difference in means by its standard deviation (i.e., made the scales unitless). The standardized difference in means10 is comparable across studies and can be expressed as an odds ratio (OR) and pooled using a random effects model11 as implemented in comprehensive meta-analysis (version 2, Biostat, Englewood, NJ). Results are presented as ORs with values over 1.00 indicating better image quality with TRO. Diagnostic accuracy measures were pooled using random-effect meta-analysis as implemented in MetaDiSc12 and also tested in a bivariate mixed effects regression model.13 Continuous outcomes were pooled as a difference in means using a random-effects model as implemented in RevMan (version 5.1, Cochrane Collaboration).14 Heterogeneity was assessed using the I2 statistic.15 Figure 1 shows the study selection process. The search strategy yielded 733 records. Of these, 692 were not relevant to the study question and were excluded. Review of the titles and abstracts in phase I identified 41 potentially relevant studies for further review. Of the 41 articles, 38 were in English, two were in Chinese,16, 17 and one was in German.18 The Chinese articles were reviewed for eligibility by a physician who was fluent in Chinese (see acknowledgements). After further review of the title and abstract, the German article was found to include a subset of the same cohort published in another article19 and was excluded. Seven articles were excluded because they compared different CT protocols instead of comparing TRO CT to a separate diagnostic modality.20-26 Nine studies did not compare the TRO CT to another control group.16, 17, 27-33 Four articles did not involve TRO CT but focused on CT for PE or coronary artery imaging instead.34-37 A total of 11 studies were included in the systematic review. Observed agreement for phase II of the review was 92.7% with a kappa of 0.81 (95% CI = 0.60 to 1.02). Table 1 describes the 11 included studies, which enrolled 3,539 patients (791 TRO CT [intervention] and 2,748 patients receiving another diagnostic modality [control]). Seven studies had case–control designs, three had cohort designs, and one was a randomized clinical trial. In six studies,38-43 the cohort consisted of patients at low to intermediate risk for ACS; in four studies,44-47 the patients had undifferentiated chest pain with suspicion for PE, aortic dissection, or ACS; and in one study,42 the cohort consisted of a subset of patients who underwent both TRO CT and invasive angiography (see Data Supplement S2, available as supporting information in the online version of this paper, for a complete description of the eligibility criteria for each of the included studies). The diagnostic modalities to which TRO CT was compared included nuclear stress testing,38 dedicated PE CT, aortic dissection CT,44, 45, 47 dedicated coronary CT,40, 43, 45, 46 and coronary angiography.19, 39, 41, 48 Four studies compared the diagnostic accuracy of TRO CT to coronary angiography,19, 39, 41, 42 and the remaining seven studies reported a number of outcomes including utilization (LOS, cost, and admission rate) and technology-specific outcomes (image quality, radiation exposure, and volume of contrast administered). One study evaluated a composite yield for the diagnosis of coronary artery disease (CAD), PE, or aortic dissection.47 Six studies used 64-slice CT technology, four studies used dual-source CT,19, 39, 43, 44 and one study included 64-slice CT technology and dual-source CT.47 Four of the six studies that used 64-slice CT technology used beta blockers to slow the heart rate and improve image quality. One study performed D-dimer measurements in all the patients and performed TRO CT instead of coronary CT angiography in those with elevated D-dimer.43 There were 13.9% in the TRO group and no patients in the coronary CT angiography group with PE. Table 2 shows patient demographics by intervention and control arm. Patients were predominantly men with a mean age of 49 to 67 years. The intervention and control groups were well balanced in regard to age, sex, and body mass index, where reported. The one randomized clinical trial included in the review did not report blinding of the participants or outcome assessors, but otherwise met criteria for appropriate sequence generation, allocation concealment, and outcome reporting.44 Table 3 shows the quality assessment for the case–control and cohort studies according to the Newcastle-Ottawa scale for case–control and cohort studies (maximum score of 9). The studies by Shapiro and colleagues,45 Madder and colleagues,47 and Gruettner and colleagues43 received scores of 6*, and the study by Takakuwa and colleagues38 a score of 5*; the remaining studies received scores of 2*, suggesting a substantial risk for bias. Table 4 shows the QUADAS-2 scores for the four diagnostic accuracy studies. All the studies were at a high risk of bias in flow and timing because not all the patients received the reference standard test (only a selected group of patients undergoing TRO CT also underwent coronary angiography). Two studies did not report blinding to the results of the reference standard test when interpreting the results of the index test. Three of the four studies consecutively recruited the patient cohort and were at low risk of bias.19, 39, 42 Most of the studies were classified as low risk for concerns regarding the applicability of the studies. Four studies evaluated the image quality of TRO CT compared to dedicated CT (Figure 2).40, 44-46 Image quality was not significantly different between the two groups (OR = 0.78, 95% CI = 0.58 to 1.06; I2 = 0%). When the two studies with the lowest methodologic quality scores were removed from the analysis,40, 46 the pooled effect estimate was similar. Four studies evaluated the diagnostic accuracy of TRO CT using coronary angiography as the criterion standard test for the detection of CAD (Table 5).19, 39, 41, 42 In each of these studies a subset of the cohort underwent TRO CT scan followed by coronary angiography, with each patient serving as his or her own control. The pooled diagnostic accuracy estimates were sensitivity of 94.3% (95% CI = 89.1% to 97.5%; I2 = 58.2%), specificity of 97.4% (95% CI = 96.1% to 98.4%; I2 = 91.2%), positive likelihood ratio (LR+) of 17.71 (95% CI = 3.92 to 79.96; I2 = 98.6%), and negative likelihood ratio (LR–) of 0.08 (95% CI = 0.02 to 0.27; I2 = 70.7%). Sensitivity and specificity estimates obtained using a bivariate mixed-effects regression model provided similar point estimates with slightly wider 95% CIs. One study47 reported no difference in the percentage of patients diagnosed with CAD with TRO CT compared to coronary angiography (13.2% vs. 16.1%, p = 0.22), but did not report diagnostic accuracy estimates comparing the two modalities. There were eight patients diagnosed with PE.39, 43, 44 In each case patients were considered to have PE without confirmation by dedicated CT for PE or invasive pulmonary angiography (no criterion standard test), so diagnostic accuracy data were not reported. There were no patients diagnosed with aortic dissection among the included studies. Five studies provided data on radiation exposure that were amenable to meta-analysis (Figure 3).43-47 A total of 2,307 patients were included: 377 patients in the TRO CT group and 1,930 patients in the control group. Patients who underwent TRO CT were exposed to more ionizing radiation (mean difference [MD] = 4.84 mSv, 95% CI = 1.65 to 8.04 mSv). Although this estimate was associated with significant statistical heterogeneity (I2 = 97%), the direction of effect indicates that TRO CT was associated with more radiation exposure in all studies. When the study with the lowest methodologic quality score46 was removed from the analysis, the pooled effect estimate was similar and the statistical heterogeneity was lower. Five studies provided data on contrast volume administration (Figure 4).43-47 A total of 2,307 patients were included: 377 patients in the TRO group and 1,930 patients in the control group. Patients undergoing TRO CT received greater volumes of intravenous (IV) contrast (MD = 38.0 mL, 95% CI = 28.1 to 48.0 mL). Although this estimate was associated with significant statistical heterogeneity (I2 = 96%), the direction of effect indicates that TRO CT was associated with a greater volume of IV contrast in all studies. When the study with the lowest methodologic quality score46 was removed from the analysis, the pooled effect estimate was similar. Two of the 11 studies reported data on LOS.38, 44 Takakuwa and colleagues38 reported that, compared to patients who underwent nuclear stress testing, TRO CT patients had significantly shorter mean lengths of stay (MD = 6.5 hours, 95% CI = 5.0 to 8.0 hours; p < 0.001). Rogers and colleagues,44 who compared TRO CT to dedicated PE, aortic dissection, or coronary CT, did not observe a significant difference in the median LOS for TRO CT compared to dedicated CT (8.2 hours vs. 7.6 hours, respectively; p = 0.79). Two studies reported data on cost.38, 44 TRO CT costs an average of $1,306 per scan, compared to $945 for nuclear stress testing (p < 0.0001).38 Although Rogers and colleagues reported a higher median cost for TRO CT, the difference was not statistically significant ($1,898 for TRO CT vs. $1,724 for dedicated CT; p = 0.16). One study conducted an economic analysis comparing the use of TRO CT to a local protocol that involved hospital admission and invasive coronary angiography for exclusion of coronary artery stenosis in patients at intermediate risk for ACS.49 The median cost of care per patient was significantly lower in the TRO CT group compared to the standard care group ($687 vs. $2,522; p < 0.001). In the subgroup of patients with significant coronary artery stenosis, the TRO CT group had a higher cost of care than the standard care group ($3,831 vs. $3,020; p < 0.001). Takakuwa and colleagues reported that patients receiving TRO CT were significantly more likely to be admitted to the hospital compared to patients receiving nuclear stress testing (11.7% vs. 5.5%, 95% CI = 1.0% to 11.4%).38 Rogers and colleagues, however, did not observe a significant difference in the rate of hospital admission for TRO compared to dedicated CT (31% vs. 30%, respectively; p = 0.99).44 Overall, there were insufficient data reported on LOS, cost, and admission rate to draw any meaningful conclusions. Triple rule-out CT was found to be highly accurate for the diagnosis of CAD. Although emergency physicians may be most interested in the diagnosis of ACS, if a TRO CT is obtained in a patient with nontraumatic chest pain and no significant coronary stenoses are identified, ACS is effectively ruled out. The opposite, however, is not true; a patient with a potentially significant coronary stenosis on TRO CT may require invasive coronary angiography to more precisely define the coronary lesion, which involves additional radiation and contrast exposure. We observed greater radiation exposure with TRO CT in our review, with a MD of 4.8 mSv. Although this amount of additional ionizing radiation is relatively small, the median radiation exposure from commonly performed CT scans ranges from 2 to 31 mSV for a single scan,50 and it is not uncommon for patients' cumulative radiation doses to exceed the 50 mSv threshold associated with cancer risk in atomic bomb survivors.51 Because cumulative radiation exposure can add up over one's lifetime, imaging tests that use radiation should only be done if there is a compelling rationale to do so. A similar line of reasoning can be applied to additional contrast exposure; although it is unclear how much an additional 38 mL of contrast increases the risk of contrast-induced nephropathy, the incidence of contrast-induced nephropathy associated with chest CT is not negligible,52 and every effort should be made to limit the amount of contrast administered to patients to minimize the risk of complications. The number of cases of PE and aortic dissection in each of the studies was too low to compare the diagnostic accuracy of TRO to dedicated CT. Although the data in this review indicate that the image quality between TRO and dedicated CT is comparable, at this point in time there are insufficient data to support the use of TRO CT for the diagnosis of PE or aortic dissection. Overall, given the increased radiation and contrast exposure and lack of diagnostic accuracy data for PE and aortic dissection, there are no grounds to recommend use of TRO CT in the diagnosis of these conditions. The majority of the studies included in the review enrolled patients with chest pain at low to moderate risk for ACS. Only four of the 10 studies enrolled patients with chest pain suspected to be due to PE or aortic dissection, and the prevalence of these conditions was exceedingly low. To accurately determine the utility of TRO CT in clinical practice, future studies will need to estimate the prevalence of each of these three diagnoses prior to conducting the study to have sufficient power to compare the diagnostic accuracy of TRO CT to dedicated PE and aortic dissection CT. In addition to careful patient selection, future trials should ascertain a number of outcomes of interest such as diagnostic accuracy, cost, ED LOS, frequency of incidental findings, and the clinical yield and cost of subsequent investigations initiated to follow-up incidental findings. Finally, outcomes will need to be carefully selected to generate meaningful inferences relevant to patients. Diagnostic randomized controlled trials in which patients are randomized to TRO CT or to a specific diagnostic modality commonly used in practice, prospectively followed to ascertain patient-important outcomes beyond the conventional diagnostic accuracy measures such as sensitivity and specificity, and analyzed on an intention to treat basis, will be critical to determine the effect of alternative diagnostic approaches on outcomes that matter to stakeholders such as patients, clinicians, payers, and policy-makers. The strengths of this review included a search strategy that involved four electronic databases, searching the bibliographies of the included articles, and contact with a content expert and authors of the included studies. This minimized the potential for publication bias. We contacted study authors to obtain data not included in the original reports. We also used sound methodology in conducting the review, including assessment of interrater reliability for study selection. Interrater reliability was high. This systematic review and meta-analysis was limited by the relatively small number of studies included in the review. The degree of details reported in each of the studies with regard to patient characteristics was limited, as several of the studies were published in the radiology literature and were more focused on technology than patient outcomes. In some studies only a subset of control patients underwent reference standard diagnostic testing, increasing the risk of bias. Two case control studies40, 46 were of poor methodologic quality. As recommended by the Working Group of the Grading of Recommendations, Assessment, Development and Evaluation (GRADE),53 sensitivity analyses should be conducted excluding these studies. If the pooled effect size and the conclusions change, such studies should be excluded. In our case, the main findings of this meta-analysis were robust to these assumptions, justifying pooling the whole body of evidence and providing a summary estimate. Although there was statistical heterogeneity in the analyses of radiation and contrast exposure, effect estimates in each of the studies were in the same direction, supporting greater radiation and contrast exposure with TRO CT. There were only eight patients diagnosed with PE among the included studies,39, 43, 44 and no cases of aortic dissection, limiting our ability to generate inferences regarding the diagnostic accuracy of TRO CT for PE or aortic dissection. Triple rule-out computed tomography is highly accurate for detecting coronary artery disease. Given the low (<1%) prevalence of pulmonary embolism and aortic dissection in the included studies, and the increased radiation and contrast exposure, there are insufficient data to recommend use of triple rule-out computed tomography in the diagnosis of these conditions. The authors appreciate Dr. Yue Dong's assistance in translating the Chinese articles identified in the process of conducting the review. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
As the marketplace for academic positions in emergency medicine grows more competitive, it becomes increasingly important for residents who desire academic careers to distinguish themselves during their residency. This report attempts to outline a road map for department and residency program leaders to help their houseofficers become successful candidates for an academic emergency medicine position. Specific ways a resident can enhance his or her “academic marketability” include 1) involvement in research, 2) establishment of a track record of productivity via scholarly writing, 3) awareness of the literature in the specialty, 4) involvement in specialty organizations and hospital committees, 5) competition for national awards, 6) gaining education skills, 7) developing an academic niche, and 8) fellowship training.
Several factors influence the final placement of a medical student candidate on an emergency medicine (EM) residency program's rank order list, including EM grade, standardized letter of recommendation, medical school class rank, and US Medical License Examination (USMLE) scores. We sought to determine the correlation of these parameters with a candidate's final rank on a residency program's rank order list.We used a retrospective cohort design to examine 129 candidate packets from an EM residency program. Class ranks were assessed according to the instructions provided by the students' medical schools. EM grades were scored from 1 (honors) to 5 (fail). Global assessments noted on the standardized letter of recommendation (SLOR) were scored from 1 (outstanding) to 4 (good). USMLE scores were reported as the candidate's 3-digit scores. Spearman's rank correlation coefficient was used to analyze data.Electronic Residency Application Service packets for 127/129 (98.4%) candidates were examined. The following parameters correlated positively with a candidate's final placement on the rank order list: EM grade, ρ = 0.379, P < 0.001; global assessment, ρ = 0.332, P < 0.001; and class rank, ρ = 0.234, P = 0.035. We found a negative correlation between final placement on the rank order list with both USMLE step 1 scores, ρ = -0.253, P=0.006; and USMLE step 2 scores, ρ = -0.348, P = 0.004.Higher scores on EM rotations, medical school class ranks, and SLOR global assessments correlated with higher placements on a rank order list, whereas candidates with higher USMLE scores had lower placements on a rank order list. However, none of the parameters examined correlated strongly with ultimate position of a candidate on the rank list, which underscores that other factors may influence a candidate's final ranking.
