A multivariate generalizability model for clinical skills assessments.
Students; Medical; Models; Statistical; Clinical Competence – Evaluation; Multivariate Analysis – Evaluation
Assessments of clinical skills of medical students rely increasingly on standardized patients demonstrating medical cases with faculty rating performance. The common finding of inconsistency of scores across cases is often referred to as case specificity. A multivariate generalizability model reveals that overall case specificity cannot explain inconsistency in this example. Results from the multivariate model suggest directions for enhancing the table of specifications, optimizing scoring, and improving case selection to reduce inconsistency in scores across cases.
Jarjoura D; Early L; Androulakakis V
Educational & Psychological Measurement
2004
2004-02
Article information provided for research and reference use only. All rights are retained by the journal listed under publisher and/or the creator(s).
<a href="http://doi.org/10.1177/0013164403258466" target="_blank" rel="noreferrer noopener">10.1177/0013164403258466</a>
Evaluation Of Acetaldehyde-modified Hemoglobin And Other Markers Of Chronic Heavy Alcohol Use: Effects Of Gender And Hemoglobin Concentration
abuse; acetaldehyde; adducts; alcohol; carbohydrate-deficient transferrin; cation-exchange chromatography; consumption; drinking; gamma-glutamyl-transferase; gender; hemoglobin; HPLC; liquid-chromatography; liver-disease; serum; Substance Abuse
The present study examined whether measurement of hemoglobin-acetaldehyde (HbA1-AcH) using an improved methodology may be useful as a biological marker of alcohol abuse. Red blood cell hemolysates of 182 patients consecutively admitted to the drug and alcohol treatment unit of our institution were analyzed for HbA1-AcH concentration using cation exchange HPLC. Mean HbA1-AcH of those who claimed to drink greater than or equal to 6 drinks/day [mean = 0.055 (% total hemoglobin), SD = 0.051] was significantly higher than the mean of those who drank <6 drinks/day (mean = 0.026, SD = 0.0174). The greatest sum of sensitivity (67%) and specificity (77%) came with a cut-score of 0.030 area% of total hemoglobin. A cut-score of 0.080 produced a 100% specificity, but lowered the sensitivity to 20%. The Pearson product moment correlation (r) between HbA1-AcH and reported drinks per day was r = 0.30 (p < 0.001). There was no significant difference in the association of HbA1-AcH end reported drinking between males and females, and the small difference observed was shown to be entirely associated with differences in hemoglobin levels between the sexes. Cocaine use did not significantly alter the correlation between reported drinking and HbA1-AcH levels. Hemoglobin levels were shown to have a significant correlation with HbA1-AcH independent of drinking. HbA1-AcH was shown to have a better sensitivity and specificity than gamma-glutamyltransferase, ALT, AST, or mean corpuscular volume in this population. The results suggest that HbA1-AcH may be a useful marker to help detect alcohol abuse, especially in populations where other markers have been shown to fail.
Hazelett S E; Liebelt R A; Brown W J; Androulakakis V; Jarjoura D; Truitt E B
Alcoholism-Clinical and Experimental Research
1998
1998-11
Journal Article or Conference Abstract Publication
<a href="http://doi.org/10.1097/00000374-199811000-00029" target="_blank" rel="noreferrer noopener">10.1097/00000374-199811000-00029</a>