The accuracy of tuberculin skin tests: self-assessment by adult outpatients.

ABSTRACT

Tuberculin skin testing is an accurate, inexpensive screening procedure for detecting tuberculosis infection. The return visit needed to interpret the reaction is inconvenient, costly, and may contribute to under-utilization of the test. Although some clinicians ask patients to read their own purified protein derivative (PPD) test results, patient accuracy and the degree of teaching needed to learn this skill are unclear. This study evaluated the accuracy with which 145 outpatients read their own Mantoux skin test (PPD) reactions and reported by postcard after brief training by nurse practitioners. A total of 89 instructed patients returned postcards and also returned for clinician readings; 46 submitted postcards without returning; 7 returned but did not complete postcards; and 3 neither returned postcards nor returned for readings. Ten of 135 postcards were uninterpretable. For 81 subjects with both interpretable tuberculin self-assessment postcards and clinician readings, overall PPD classification agreement was 88 percent; Kappaw = +0.905 (P less than .001). Compared to clinician readings, 1 of 53 patients falsely reported a positive reaction (greater than or equal to 10 mm) and 2 of 25 patients falsely reported negative PPD readings (0-4 mm). There was 100 percent agreement between postcard readings and clinician classifications in a subgroup of patients (N = 26), prospectively identified by nurse practitioners as capable of accurate tuberculin self-assessment. Inter-clinician reading agreement (N = 37) was 89 percent; Kappaw = +0.943 (P less than .001). The brief standardized teaching protocol described can enable most patients to measure and report their PPD results. Study results suggest thatpostcard reports, especially negative ones, from a subgroup of patients selected for their skill in measuring their initial PPD wheal and ability to paraphrase instructions, might be substituted for clinician readings.