Poor Validity of Parent Reading of Skin Test Induration in a High Risk Population

Between 1985 and 1992 there was a 20% rise in cases of tuberculosis overall and a 35% rise in cases in children in the United States.1 With tuberculosis on the rise renewed emphasis has been placed on selective or periodic tuberculin skin testing. The American Academy of Pediatrics 1994 policy statement on Screening for Tuberculosis in Infants and Children recommended annual Mantoux tuberculin tests for children at high risk including “poor and medically indigent city dwellers.” These recommendations stated that the use of multiple puncture tests (MPTs) should be “severely restricted, if not eliminated” and Mantoux tests should be “used and read by qualified medical personnel.”2 The CDC and the American Thoracic Society have supported the recommendations that children be tested depending on their epidemiologic risk group and that Mantoux tests be used.3 Questions regarding the feasibility of implementing this policy have been raised especially concerning compliance in reading the test at 48 to 72 h. Currently many physicians are screening with MPTs and allowing parent/guardian reading of results and return of a postcard. One survey of practicing physicians found that 55% used MPTs in 1994 and 91% allowed parent reading of the test.4 Of those using Mantoux tests 19% allowed parent reading of the test.4 Some skin tests done by allergists are also read by parents. Studies in adult clinics found that self-report of skin tests with an illustrated card or pen or palpation technique was reasonably accurate in a subgroup of patients.5-8 However, there are few published data regarding the accuracy of self-report results and compliance with 48- to 72-h reading in a population of poor and medically indigent children. This study was undertaken to determine whether parents of children who are at high risk for tuberculosis can read Mantoux test induration (either induration present or induration absent) with a satisfactory degree of accuracy. Methods. Families attending a food bank (food giveaway) were solicited for participation. Children were included if they were ages 9 months to 17 years, had no previous positive tuberculin skin test or previous antituberculosis medication and had no Mantoux test in the past year. Only one child from each family was allowed to enter the study. Parental informed consent was obtained. Parents were interviewed regarding demographic information and their child's tuberculosis risk factors. Demographics included maternal education level, child's date of birth, health insurance, birth place, race and history of tuberculin induration. Tuberculosis risk factors included information on contacts with persons from countries of high tuberculosis prevalence, homeless, users of street drugs, incarcerated persons or those infected with HIV. We also asked about the parent's level of concern about tuberculosis for their child. Parents were informed about the importance of the test, the need to read test results and how to read induration (induration present or induration absent) at 48 h using an illustrated card. The 10-min educational program was given verbally in English or Spanish, as appropriate, with a standard script. Parents were shown examples of test induration on illustrated cards and then were asked to read sample cards. All written information was in English and Spanish and included the date and time for test reading and return. Children received Mantoux (purified protein derivative, 5 tuberculin units) and Candida intradermal skin tests placed on different arms by a physician. The skin tests were circled with permanent marker. The parent's hand was stamped with the reading date and a hospital bracelet with the reading date was placed on the child's wrist. Parents were instructed to read their child's skin tests just before their return reading visit at 48 to 72 h and to record results on a card. On return one of two investigators (blinded to the parent reading) measured test induration using the pen technique.9 All parents also reread card examples used in the educational program. A $20 food voucher was given on completion of the study. Agreement between parent and physician readings (the standard) was calculated to determine the degree of concordance. Sensitivity, specificity and predictive values were calculated for the Mantoux and Candida tests separately and together. Tests of association or for group differences were used to determine the relationship between the adherence and accuracy of parent reading with demographics and tuberculosis risk factors. Results. Of 37 families tested 36 returned for reading. Children participating were between the ages of 1 and 16 years with a mean of 7.6 years. Sixty-four percent were born in the United States. Seventy-eight percent were Hispanic, 14% African-American and 8% Caucasian. Parents of 41% of the children had less than a high school education; 48% were high school graduates. Regarding contact with people at high risk for tuberculosis, 78% had contact with people born in high prevalence countries, 11% with street drug users, 9% with homeless people, 6% with HIV-infected persons and 14% with incarcerated individuals. Twenty-four children (67%) had induration of ≥2 mm on one of the two skin tests and five children (14%) had induration of ≥10 mm. Of these five, three children (8%) had Mantoux induration of ≥10 mm, and two (6%) had Candida induration of ≥10 mm. Three (8%) families failed to bring back the card, seven (19%) returned an incomplete card and eight (22%) read the test before 48 h. Predictive value of parents readings compared with investigator readings of ≥ 2, 5, and 10 mm indurations are shown in Tables 1 to 3. When asked to reread induration on illustrated cards used in the educational program, nine parents (27%) read no induration on an illustrated example with a 12-mm bump. Only eight families (22%) documented test results and read skin test and illustrated cards accurately. There were no significant differences between families able and not able to read induration with regard to demographics, tuberculosis risk factors or adherence to reading time or documentation. Also those who did not adhere to test reading did not differ in demographics and tuberculosis risk factors compared with those who adhered. Discussion. Self-report of delayed hypersensitivity skin tests has been common practice in tuberculosis and allergy testing despite little study of patients' accuracy. No other screening test used in pediatrics is interpreted routinely by nonprofessionals. With renewed attention to tuberculosis testing and recommendations to eliminate parental self-report, systematic study of self-report information is critical. Although control skin tests to assess anergy are indicated only in specific cases, we used Candida controls to assess parent reading of skin test induration.2 To our knowledge this is the only study that tests prospectively an intervention to teach parents how to read induration or no induration in a high risk population. Acceptance of self-report of test results requires the successful completion of several steps: (1) the parent must remember to check the test at the right time; (2) parents must document the results in a way that can be efficiently transmitted to the health care provider; (3) the test must be read accurately; (4) the parent must act appropriately on the test results. This study tested the first three steps of the self-report process. Remembering to read the skin test result on time presented a problem for some parents. Despite our efforts to remind them, 22% read the test at the wrong time. This occurred even though the reading time was discussed and written on the instruction information, the date was stamped on the parent's hand and on the child's hospital bracelet and the test sites were circled in permanent marker as a reminder to check the test. We suspect that this percentage may be spuriously low because of the Hawthorne effect.10 This effect suggests that the simple act of observation may change behavior. Because this study emphasized adherence and offered incentives for participation, the fact that 78% of parents did read the test at the appropriate time probably overestimates the true degree of adherence, raising further concern about the utility of parental reading. The second step in self-report involves documentation of test results. If test results cannot be documented, providers cannot be sure that tests were read or assume that tests were negative. Although many practitioners using MPT asked parents to return a postcard with test results, most studies have shown variable rates of postcard return.11-13 In our study 8% forgot to bring the card or misplaced it and 19% brought the card but reading time or reading results were not recorded. Again under study conditions our results indicate a “best case scenario” because there were incentives to participate in the study. Finally we found that the accuracy of parental report of induration or no induration was not adequate for screening purposes. Many parents felt no induration when physicians felt induration, thus reducing the sensitivity and negative predictive value of parental reading. When parents were asked to reread illustrated cards used in the educational program, 27% felt no induration on the example that showed a 12-mm bump. This finding raised additional concern about the ability of parents to detect induration. Although this study involved a high risk population and results may not be generalizable to other groups, the findings raise serious concerns about the ability of some parents to follow through the steps required for self-report of results. In this high risk population parent interpretation and documentation of skin test induration was not adequate for screening purposes with only 22% documenting and reading tests appropriately. This supports the recommendation that delayed hypersensitivity skin tests in this population be read by qualified medical personnel. Acknowledgments. The authors thank the families and the food bank for their participation in this study. This project was supported by the Ambulatory Pediatrics Association. Tina L. Cheng, M.D., M.P.H.; Mary Ottolini, M.D., M.P.H.; Pamela Getson, Ph.D.; Muriel D. Wolf, M.D.; Peter C. Scheidt, M.D, M.P.H.