Assessing the validity of tuberculin skin test readings by trained professionals and patients

Study objectives: To assess the validity of purified protein derivative (PPD) readings by patients and trained health-care professionals as compared with a calibrated model. Design and participants: Survey of a group of patients, nurses, medical assistants, and physicians at five neighborhood health centers in the Bronx, NY. Interventions: Participants were asked to read a calibrated model with four PPD indurations measuring 0 mm, 3 mm, 7 mm, and 13 mm. Indurations ≥ 5 mm were to be considered 'positive' reactions. Measurements and results: Data were obtained from 233 patients and 80 trained professionals. All patients correctly measured the 0-mm induration site and were able to detect the presence of an induration in 99.3% of possible observations. Compared with professionals, patients had more variability in measurements and interpretations of the 3-, 7-, and 13-mm sites. Professionals detected 100% of all indurations. Patients' specificity for the 0- and 3-mm sites was 97.4% and 62.7%, respectively; whereas sensitivity for the 7- and 13-mm sites was 68.2% and 89.3%, respectively. Professionals' specificity for the 0- and 3-mm sites was 98.7% and 65.3%, respectively; their sensitivity for the 7- and 13-mm sites was 86.7% and 97.3%, respectively. Seventy percent of professionals agreed that the model was a realistic representation of PPD indurations. Conclusions: Patients can reliably distinguish between the presence and absence of an induration at a PPD injection site. They are not as reliable in the measurement and interpretation of test reactions. Professionals had considerable variability in their assessments of PPDs but were more precise overall in their assessments than patients.