The tuberculin (TB) skin test with purified protein derivative is the recognized method of choice for diagnosing infection with Mycobacterium tuberculosis1, but TB skin tests are not easy to interpret. In areas with a low prevalence of tuberculosis infection, providers may not have the opportunity to read many positive TB skin tests, and this can result in inaccurate interpretations. Both false positive and false negative TB skin test readings can lead to inappropriate treatment decisions. Therefore it is important to determine how well health care providers interpret TB skin tests and also to identify any measurement techniques that can increase the frequency of correct readings.
There are several methods of interpreting TB skin tests, with the most widely accepted ones being the palpation method and the ballpoint pen technique. The American Thoracic Society recommends using palpation to assess induration but also states that the ballpoint pen technique might decrease interobserver variability. 1 The 2000 Red Book suggests using the ballpoint pen technique and does not mention the palpation technique. 2 Kendig et al. 3 reported that 30% of pediatricians, nurses and residents attending a pediatric conference read a 15-mm TB skin test reaction as <10 mm, but only 2 of the 107 participants used the pen technique. Several investigators have reported that use of the pen technique results in less inter- and intraobserver variability compared with using palpation alone. 4, 5 Others who compared the pen technique directly to the palpation method found them comparable. 6–8 In these studies a small number of providers read the TB skin tests, and the majority of the skin test reactions were <10 mm. The pen technique has not been evaluated in a setting of a large number of providers inexperienced in reading TB skin tests. We wished to determine how well a group of pediatric care providers with a wide spectrum of training and experience interpreted a positive TB skin test. In addition we wanted to identify any particular method that could decrease misreadings.
SUBJECTS AND METHODS
Participants and procedures
We asked health care providers, nurses, staff pediatricians, residents and medical students who were working in the pediatrics clinic to participate in the study. We obtained verbal but not written informed consent, and the investigators made it clear that participation in the study was purely voluntary. Fifty-seven care workers agreed to participate, and none refused. The participants were divided into 4 groups: 16 staff pediatricians; 13 pediatric and family practice residents; 20 nurses [5 RNs and 15 licensed practical nurses (LPNs)]; and 8 medical students (4 third year and 4 fourth year). We included LPNs because at our institution LPNs interpret TB skin test reactions. One of the investigators, a white man and a known purified protein derivative converter, had a standard 5-unit tuberculin dose placed subcutaneously on the volar aspect of his left arm. Approximately 48 h later, both the Chief of Pediatric Infectious Diseases and the Chief of Allergy and Immunology independently measured the reaction. They used the ballpoint pen technique (their choice); this consisted of running the tip of the ballpoint pen to the edge of induration to demarcate the site of induration. The widest diameter of induration provided the value for the skin test reaction. One of the “official readers” measured the reaction to be 15 mm and the other 14–16 mm, so we chose 15 mm as the valid measurement. On that same day, during a 6-h period, we asked the study participants to interpret the skin test reaction. We told each provider to use any method they desired, and we provided a ruler, tape measure and pen, but only on request. After each measurement we carefully cleansed the skin test site with alcohol before moving on to the next interpreter. At the end of the study, the skin test site was remeasured to determine whether the size of the induration had changed from the beginning to the end of the study, and it was still 15 mm.
Outcome measures and statistical analysis
The primary outcome measure was the percentage of providers who read the TB skin test reaction as ≥10 mm. We considered this a correct interpretation and any value <10 mm as a misreading. We chose this cutoff because we felt that anyone who read a 15-mm skin test as <10 mm had definitely misread it. Secondary measures included the number of providers who interpreted the reaction as negative (defined as ≤5 mm) and the number of providers who used the ballpoint pen technique. Only providers who used the ballpoint pen to appropriately mark and then measure the diameter of the induration were included in the ballpoint pen technique group. We also analyzed differences in these outcome measures between those participants who used the pen technique and those who did not. We used the Fisher’s exact and chi square tests to assess intergroup differences among the four groups (nurses, medical students, residents and staff physicians).
Fifty-seven care workers participated in the study, and all but two of them knew to measure the size of the induration rather than the erythema. Measurement techniques ranged from palpation to simple visual inspection, and 32% of the participants used the pen technique. All but one participant used a ruler or tape measure to determine the actual size of the reaction. Less than one-half of the participants remembered receiving formal training on how to read a TB skin test or having actually seen a positive TB skin test reaction.
There were no significant differences between the 4 groups of providers with respect to how well they interpreted the skin test reaction (Table 1). Seventy-seven percent (44 of 57) of the participants interpreted the reaction as ≥10 mm, but this meant that 23% failed to correctly interpret the skin test reaction. In addition 10 participants (18%) read the reaction as ≤5 mm (Table 1). Ninety-four percent of the participants who used the pen technique measured the skin test reaction as ≥10 mm compared with 69% of those who used other methods (P = 0.04;Table 2). In addition no one who used the pen technique read the skin test reaction as ≤5 mm compared with 26% of those who used other methods (P = 0.02;Table 2).
In this study we identified several problems with the interpretation of a positive TB skin test reaction. First we found that many providers, whether nurses, residents, medical students or staff physicians, could not remember having received formal training in skin test interpretation. Many of them had not even seen a positive TB skin test. Second we found that approximately one-fifth of the providers interpreted a 15-mm TB skin test reaction as <10 mm whereas one in six actually read it as negative (≤5 mm). Finally use of the pen technique was associated with a decrease in the number of incorrect readings.
Interpretation of a TB skin test is not easy, and errors in reading it can result in patients receiving inappropriate treatment. False positive readings, which are more likely when the prevalence of TB infection is low, can lead to overprescribing antituberculous medications, whereas false negative readings may delay appropriate treatment. In our study 23% of providers would have inappropriately withheld treatment based on their misreading of a positive TB skin test as <10 mm. This is an unacceptable error rate and underscores the necessity of finding ways to improve the accuracy of skin test interpretation.
Other groups of providers might perform better with respect to reading TB skin tests than the participants in our study, but we believe that our study participants were fairly representative of the medical population at large. The 4 groups in our study differed widely in their training and experience, yet they all performed comparably. Furthermore in a similarly designed study by Kendig et al., 3 33% of 107 pediatric health care providers misread a 15-mm TB skin test reaction as <10 mm. Only 2 providers in that study opted to use the pen technique, so that method could not be evaluated.
We found that the participants in our study who used the pen technique had fewer false negative skin test readings. This suggests that the pen technique can help providers interpret a positive TB skin test more accurately. In studies that have compared the pen technique to palpation for both intraobserver and interobserver variability, the pen technique fared either better than or the same as palpation. 5, 6 In these studies the persons reading the skin tests were quite experienced in skin test interpretation. Bouros et al. 8 found that a single inexperienced reader interpreted 539 skin tests accurately with both the palpation method and the pen technique. However, in this study the “inexperienced” provider rapidly gained experience with both methods of interpretation as hundreds of skin test reactions were read. In our study multiple providers interpreted a single test, so there was no chance for our participants to gain experience during the study. We found that the pen technique definitely improved the accuracy of TB skin test interpretation in a group of providers who were not experienced in reading positive skin tests. Thus the pen technique may not be necessary for persons experienced in interpreting skin tests but can be helpful in enabling those who are inexperienced to measure skin test reactions correctly.
Our study had several limitations. We had only one skin test reaction for participants to read, and this was positive at 15 mm. Thus we were not able to evaluate false positive rates. It would have been useful to have had a truly negative skin test reaction with some erythema and no induration to be able to assess for the accuracy of negative readings. In addition we had only 10 to 15 participants in each of our subgroups. With more participants we might have been able to more clearly identify intergroup differences as well as to more precisely determine the benefit of using the pen technique. Also we did not prospectively assess the utility of the pen technique or compare it directly with the palpation method. We were only able to evaluate the pen technique because enough of the participants opted to use it.
In conclusion we found that ∼20% of medical health care providers in a pediatric clinic misread a 15-mm TB skin test reaction as <10 mm, a false negative rate similar to that reported by other investigators. Providers who used the pen technique had fewer incorrect readings. Thus the ballpoint pen technique may help reduce incorrect readings. We suggest that medical and nursing schools as well as residency programs institute formal training in TB skin test interpretation and consider teaching the ballpoint pen technique.
1. American Thoracic Society. Diagnostic standards and classification of tuberculosis in adults and children: official ATS statement. Am J Respir Crit Care Med 2000; 161: 1374–95.
2. American Academy of Pediatrics. In: Pickering LK, ed. 2000 Red Book: Report of the Committee on Infectious Diseases. 25th ed. Elk Grove Village, IL: American Academy of Pediatrics, 2000:596–8.
3. Kendig EL, Kirkpatrick BV, Carter WH, Hill FA, Caldwell K, Entwistle M. Underreading of the tuberculin skin test reaction. Chest 1998; 113: 1175–7.
4. Sokal JE. Measurement of delayed skin-test responses. N Engl J Med 1975; 293: 501–2.
5. Bouros D, Zeros G, Panaretos C, Vassilatos C, Siafakas N. Palpation vs.
pen method for the measurement of skin tuberculin reaction (Mantoux test). Chest 1991; 99: 416–9.
6. Longfield JN, Margileth AM, Golden SM, Lazoritz D, Bohan JS, Cruess DF. Interobserver and method variability in tuberculin skin testing. Pediatr Infect Dis 1984; 3: 323–6.
7. Jordan TJ, Sunderam G, Thomas L, Reichman LB. Tuberculin reaction size measurement by the pen method compared to traditional palpation. Chest 1987; 92: 234–6.
8. Bouros D, Maltezakis G, Tzanakis N, Tzortzaki E, Siafakas N. The role of inexperience in measuring tuberculin skin reaction (Mantoux test) by the pen or palpation technique. Respir Med 1992; 86: 219–23.
9. Presented at the International Conference of the American Thoracic Society, Toronto, Canada, May, 2000.