Objective. To determine the most effective strategy to encourage adherence with tuberculosis test reading in a high-risk population.
Design. Prospective randomized controlled trial.
Setting/Participants. Consecutive sample of 627 children ages 1 to 12 years due for a tuberculosis (TB) test in an urban children's hospital outpatient department. One child per family was enrolled.
Intervention. All families received education regarding the importance of skin testing for TB and the need for follow-up, and written and verbal instructions regarding test reading. Families were randomly assigned to one of five strategies for follow-up TB test reading at 48 to 72 hours: 1) routine verbal and written instructions, 2) reminder phone call, 3) transportation tokens and toy on return, 4) withholding of school forms until time of reading and need to repeat TB test if not timely read, 5) parents taught to read induration with nurse home visit. Those who did not have tests read at 48 to 72 hours by a trained professional were phoned 1 week later.
Results. The five groups did not differ with regard to TB risk factor score, maternal education, transportation source, or perceived importance of TB testing. Before the study the follow-up rate of TB test reading by a trained professional was 45%. Reading rates in this study were 58%, 70%, 67%, 70%, and 72% for groups 1 to 5, respectively. In group 4, only 39% had school forms to be completed and their adherence rate was 84% (53/63). Compared to group 1, the only statistically significant improvement was in group 4, especially for those who needed school forms completed, and in group 5. Those not adhering in groups 1 to 4 did not differ from returnees with regard to TB risk factors, maternal education, transportation, or perceived importance of testing. The most common reasons for failing to return included forgetfulness, transportation, and time constraints. Group 5 was stopped early because of difficulty with nurse visits (N = 98). When told of the nurse visit, 9% (9/98) families could not find a time for the visit. Seventeen percent (17/98) were visited but the child was not home, and 7% (7/98) were not visited because of a nurse scheduling problem.
Conclusions. In a high-risk population, adherence with TB test reading is poor. However, education and return of school forms at reading time can significantly improve adherence. Although requiring larger investment in resources, visiting nurses may also aid in test reading.
Despite great strides in preventing, diagnosing, and treating childhood disease, there remain large populations that have not benefited fully from these advances. Immunization and screening for treatable conditions are areas where the gap between potential and performance is particularly notable. As we enter a new era of preventive health care, effective screening policies are critical to expanded prevention initiatives for underserved communities. Relatively little effort has been devoted to examining their effectiveness (real world usefulness), as opposed to their efficacy (usefulness under controlled conditions). In order to guide public health efforts, this study examines the effectiveness of health supervision tuberculosis (TB) screening policies in a poor, urban population.
After consistent decline in TB incidence and predictions of eradication in the United States, this past decade has witnessed a resurgence. This resurgence has disproportionately affected those in urban areas, minorities, and children. Between 1985 and 1992 there was a 20% rise in TB cases overall and a 35% rise in children.1-5 The emergence of drug resistance and coinfection with the human immunodeficiency virus have hindered efforts for effective diagnosis and treatment.6,7 Immigration from countries with high prevalence and a general decline in public health services and access to care in the United States have also contributed to the problem. Though the number of new cases has declined in the last 3 years,8 TB continues to be of great concern in the United States and is responsible for great morbidity and mortality worldwide. Renewed emphasis has been placed on identifying cases of infection and disease and on TB skin testing.
In January 1994, the American Academy of Pediatrics (AAP) issued a policy statement9 that recommended annual Mantoux tuberculin tests for high-risk children including “poor and medically indigent city dwellers” and periodic testing in low-risk children in high-prevalence areas. It also stated that the use of multiple-puncture tests should be “severely restricted, if not eliminated” and Mantoux tests should be “read by qualified medical personnel.” An update in February 1996 reaffirmed the use of Mantoux tests with more selective screening.10
Questions regarding the feasibility of implementing these recommendations have been raised especially concerning adherence in requiring a health care professional read the Mantoux test at 48 to 72 hours. Even after the 1994 recommendations, many providers continue to use multiple-puncture tests with concerns about adherence with Mantoux test reading.11 Published reports regarding return rates for clinician assessment of the test vary widely from 18% to 72% depending on the population.12-14 Little has been published on adherence issues in TB testing. This study was a prospective, randomized trial designed to determine the most effective strategy to encourage adherence with the AAP recommendations for TB test reading in a high-risk population. This study also provided an opportunity to identify the reasons for failure to return for test reading.
A consecutive sample of healthy children due for a TB test were recruited for the study. Children were ages 1 to 12 years and had no recent history of TB contact. Only one child per family was enrolled. The study took place at an urban children's hospital outpatient department. The study was approved by the Institutional Review Board of the Children's National Medical Center, and written informed consent was obtained.
Parents were interviewed by their primary care physician regarding demographics and TB risk factors. They were also asked to rate the importance of TB testing for their children, the level of difficulty in returning for test reading, and their transportation source.
A research assistant coordinated the educational intervention, which included discussion of the importance of TB testing and the need to return to read results. Instructions were given to return to the clinic in 48 to 72 hours. Reading hours were from 8 am to 7:30pm daily and 9 am to noon on weekends. Reminders to have the test read included a written information sheet with the times to return. Skin tests were circled in permanent marker and the date of return was stamped on the mother's and child's hands.
Consecutive children were randomized by day of the week to one of five strategies. Group 1 received verbal and written instructions for follow-up as did all other groups. Group 2 also received a reminder phone call. Group 3, the positive reinforcement group, was offered transportation tokens and a toy on return. Group 4, the negative reinforcement group was (a) asked to leave school forms until they returned for test reading, and (b) told that the test would be repeated if not read on time. This group was oversampled since not all children had school forms to be completed. Group 5 parents were trained to read the Mantoux TB test for induration or no induration and a nurse home visit was scheduled to verify results. Those who did not have tests read were called 1 week after the test was administered.
Demographics and Risk Factors
Six hundred twenty-seven children were enrolled and randomized into one of five groups. Twelve families, or 2%, refused to participate. Ninety-one percent of participants were African-American and 74% were on Medicaid. Twenty-two percent of mothers had less than a high school education, 37% had graduated from high school, and 41% had more than a high school education.
We inquired about six risk factors for TB. In our population, less than 1% of children were born in a country with a high prevalence of TB. Eighteen percent had contact with people from high-prevalence areas. Eight percent had contact with people who were homeless, 14% with street drug users, 17% with incarcerated persons, and 8% with human immunodeficiency virus-infected individuals. A TB risk factor score was created which included each of these six risk factors. Fifty-five percent of participants had none of these six risk factors while 45% had one or more.
Parents were asked about their usual source of transportation to the clinic. Forty-six percent used a family car, 8% used a friend's car, 33% used the bus or subway, 10% relied on a cab, and 4% walked. When asked how easy or hard it is to come back to the clinic in 2 to 3 days, 57% stated it was easy, 20% were neutral, and 23% felt it was difficult. When asked about the importance of TB testing for their child, the vast majority felt it was very important. There were few patients with purified protein derivative ≥10 mm (Table1) and no known cases of TB in the clinic during the study period.
Participants in the five groups did not differ with regard to maternal education, race, TB risk factor score, transportation source, or perceived importance of TB testing (Table 1).
Our clinic had been using multiple-puncture tests with parent reading until 4 months prior to the start of the study, when the policy was changed to Mantoux tests with return reading. In the period before the study, the adherence rate for return for test reading was 45% (N = 742). Adherence rates for all groups increased after we began the study (Table 2). Adherence rates for the groups were 58% for group 1, routine verbal and written instruction; 70% for group 2 with reminder phone call; 67% for group 3 positive reinforcement with the transportation tokens and toy; and 70% for group 4 negative reinforcement with withholding of school forms until test was read and emphasis on repeating the test if not read within 48 to 72 hours. Those in Group 4 who had school forms and left them had an 84% return rate. Those who did not have school forms had a return rate of 62%. Compared to group 1, return for test reading was improved in group 4 among those who left school forms (P = .002) and in Group 5 (P = .037).
Group 2 members received reminder phone calls to return for the test reading. A parent was reached by phone in 55% of calls, messages were left in 37%, there was no answer or a busy signal in 6%, and 2% of calls were disconnected or wrong numbers.
Group 5 was terminated after only 98 patients because of scheduling difficulties with the visiting nurses. Group 5 had a reading rate of 72%. For those randomized to group 5, 9% of families (9/98) could not find time for the home visit and were not scheduled for a visit. Seventeen percent (17/98) were visited but the child was not home, and 7% (7/98) were not visited mainly because of scheduling problems in arranging escorts to go into dangerous neighborhoods.
Parents in Group 5 had been taught how to read for presence or absence of induration and were instructed to check the arm and return a card documenting the results. The visiting nurse rechecked the results at 48 to 72 hours. Of the children who were seen by a visiting nurse, 63 (71%) of 89 parents had read their child's test at the appropriate time and 64 (72%) of 89 returned a completed card documenting results. Three parents noted induration which was not confirmed by nurse reading. Five parents denied any induration on their child's test, although the nurse noted induration of ≥2 mm.
We analyzed a subpopulation to understand which days and at what times families preferred to return for reading. For those having the option of weekday or weekend return (ie, the test was placed on Wednesday and they could return on Friday or Saturday), approximately two thirds chose the weekday and one third chose the weekend day. On weekdays, approximately two thirds came during day hours and one third chose night hours.
Reasons for Poor Adherence
All families who did not return for reading were called 1 week later. After one to three calls, 63% were contacted by phone. 10% had disconnected or wrong phone numbers. When asked for their reasons for failure to return, 25% stated that they did not have time, 18% had transportation or money problems, 15% forgot, 11% had family health problems, 7% stated that they read the test themselves, and 23% had other reasons. Those not returning for reading did not differ from returnees with regard to TB risk factor score, maternal education, transportation source, or perceived importance of testing.
In this high-risk population, adherence with TB test reading was poor. Despite our best efforts, 30% to 40% did not have their test read by a trained individual. Without accurate test interpretation TB testing has little value. With American Academy of Pediatrics recommendations urging Mantoux testing, return visits for test reading, and less frequent and more targeted skin testing, the issue of adherence with test reading is especially critical.
This study has several limitations. It was conducted in an inner-city urban clinic. Results may or may not be generalizable to other high-risk populations. Also, our clinic had limited hours for test reading especially on weekends which may have lowered our return rate. This was a randomized, controlled trial with randomization of patients by group by day of the week. Although there exists the potential for bias in randomization, we found no differences between the groups for the characteristics presented in Table 1.
When comparing adherence rates of Group 1 with the other groups, only Group 4 and Group 5 were significantly improved, although Groups 2 and 3 had increased adherence rates. The study sample size had 80% power to detect a 15% difference at α = .05. It is possible that the study had insufficient power to demonstrate statistically significant increases in adherence in the other groups. Perhaps with a larger sample, the increased adherence would reach statistical significance.
Although adherence with test reading significantly increased over 20% after the commencement of our educational intervention, it is likely that the Hawthorne effect played a role. The Hawthorne effect postulates that the act of observation and study may change behavior, independent of our educational intervention.
Group 5 used visiting nurses to assess the ability of parents to read for presence or absence of induration. Although the majority were able to read presence or absence of induration correctly, few patients had induration. Therefore, while it appeared that parents could read no induration, we were unable to establish whether parents can consistently identify the presence of induration. Other studies have more rigorously addressed this question.15
The effectiveness of screening guidelines rely on adherence by physicians, patients, and families. Physicians have expressed concern about adherence in returning for Mantoux test reading, which has likely influenced their TB testing policies. One survey found that after the 1994 AAP recommendations, the number of practitioners using multipuncture tests declined; however, 55% continued using multipuncture tests.11 Our study affirms practitioner concern about follow-up reading of Mantoux tests in a high-risk population.
We did find that in practice, education and return of school forms at reading time can improve adherence. Although visiting nurses may aid in test reading, this study showed their use to be impractical in this situation.
Further study is needed in other populations and on alternative strategies for screening and test reading. Training community workers like daycare providers and school personnel to read tests and targeting at-risk populations outside of child health supervision are options to be explored. Understanding adherence issues in screening is crucial to guide practice options and determine optimal strategies for screening in different populations.
This project was supported by the Ambulatory Pediatrics Association and presented at the annual meeting in May 1995 in San Diego, CA.
We wish to thank the residents, practitioners, nurses, and patients who participated in the study.
- Received August 5, 1996.
- Accepted January 17, 1997.
Reprint requests to (T.L.C.) Department of General Pediatrics, Children's National Medical Center, 111 Michigan Avenue, NW, Washington, DC 20010.
- TB =
- tuberculosis •
- AAP =
- American Academy of Pediatrics
- ↵Centers for Disease Control and Prevention. Initial therapy for tuberculosis in the era of multi-drug resistance: recommendations of the Advisory Council for the Elimination of Tuberculosis. MMWR. 1993;42(RR7):1–8
- Inselman LS
- American Academy of Pediatrics, Committee on Infectious Diseases
- American Academy of Pediatrics, Committee on Infectious Diseases
- Maqbool S,
- Asnes RS,
- Grebin B
- Copyright © 1997 American Academy of Pediatrics