Journal of Health Care for the Poor and Underserved

Brief communication FOLLOW-UP OF TUBERCULIN SKIN TESTS After many years of decline, incidence rates of tuberculosis are again increasing12, particularly, as Neims3 points out, in homeless and other underserved populations. This increase, especially with the emergence of strains resistant to multiple antibiotics, has focused attention on prevention.4 The cornerstone of early intervention is the tuberculin skin test. There is a substantial literature about the type, placement, and interpretation of tuberculin skin-test results5,6, but much variation in the methods of following up on skin tests once they are placed. Some experts suggest that all skin tests should be read by trained and experienced clinicians7; this is the follow-up method used in our local health department and in our hospital's adult medicine clinic. Unfortunately, return rates for appointments to read the skin test may be lowest in those populations at highest risk of tuberculosis, prompting others to suggest that some follow-up, even by parents, who may lack the skills to assess skin-test results accurately, is better than no follow-up at all. In our pediatric clinic which serves poor urban children at high risk of tuberculosis, families were asked to return to clinic if they did not speak English and follow-up was uncertain or if they noticed any redness or swelling at the test site. However, follow-up rates were only about 10 percent, and we were concerned about the potential for missing positive results. Two middle-ground strategies which are commonly employed are mailing postcards to families8"10 or having nurses phone families to ask about test results. We compared four follow-up strategies to determine which method had the highest rate of any type of follow-up and which strategy had the highest rate of clinician assessment of skin-test results: 1) asking families to assess the skin test and return if any redness or other reaction was noted (Family); 2) reappointment to have all tests assessed by a trained clinician (Reappointment); 3) nurse followup by telephone (Nurse-call); and 4) asking parents to return a preprinted postcard indicating the appearance of the test site (Postcard). Methods The study was conducted prospectively over 10 months. Because large numbers of staff members participated in the study, each follow-up method was used for two to three months rather than being randomized for each patient. The Journal of Health Care for the Poor and Underserved · Vol. 5, No. 1 · 1994 Follow-Up of Tuberculin Skin Tests two strategies used in our community, Family and Reappointment, were each used for three months, and the other strategies, Nurse-call and Postcard, were each used for two months. Throughout the study, all non-English-speaking parents were asked to return (Reappointment); this affected seven to 10 percent of all patients tested during each period. The Mantoux tuberculin test was administered by one of four trained nurses. Information about the child's age, gender, race, payer status, birthplace, and reason for administering the test was obtained for all patients. Outcomes were recorded as "any follow-up" (by phone call, postcard, or return visit) or "clinician assessment" (the patient returned between 48 and 72 hours after placement of the skin test, and had the site evaluated by a clinic nurse or, on weekends, by a pediatrician in the emergency department). Patients who did not return for a scheduled appointment or who did not return postcards were called by a nurse to ask families about reactions at the test site. Reactions were judged positive if there were >10 millimeters of induration (hardening of the skin).7 Results were analyzed using one-way analysis of variance. Results Over the 10-month study, 274 tuberculin skin tests were administered. Patients' demographic factors were similar for all four follow-up methods. The patients' median age was 19 months; 49 percent were male; 80 percent were on Medicaid; 41 percent were African-American, 25 percent were Asian, 23 percent were non-Hispanic Caucasian, eight percent were Hispanic, and three percent were Native American. Eighty percent of tests were administered for routine screening, 13 percent were administered to foreign-born children, four percent were administered as part of contact evaluations, and three percent were...