Depression screening in cardiac patients has been recommended by the American Heart Association, but the best approach remains unclear.
To evaluate nurse-administered versions of the Patient Health Questionnaire for depression screening in patients hospitalized for acute coronary syndrome.
Staff nurses in an urban cardiac care unit administered versions 2, 9, and 10 of the questionnaire to 100 patients with acute coronary syndrome. The Depression Interview and Structured Hamilton was administered by advanced practice nurses blinded to the results of the Patient Health Questionnaire. With the results of the Depression Interview and Structured Hamilton as a criterion, receiver operating characteristic analyses were done for each version of the Patient Health Questionnaire. The Delong method was used for pairwise comparisons. Cutoff scores balancing false-negatives and false-positives were determined by using the Youden Index.
Each version of the questionnaire had excellent area-under- the-curve statistics: 91.2%, 92.6%, and 93.4% for versions 2, 9, and 10, respectively. Differences among the 3 versions were not significant. Each version yielded higher symptom scores in depressed patients than in nondepressed patients: version 2 scores, 3.4 vs 0.6, P = .001; version 9 scores, 13 vs 3.4, P < .001; and version 10 scores, 14.5 vs 3.6, P < .001.
For depression screening in hospitalized patients with acute coronary syndrome, the Patient Health Questionnaire 2 is as accurate as longer versions when administered by nurses. Further study is needed to determine if screening with this tool changes clinical decision making or improves outcomes in these patients.