Improving detection and notification of tuberculosis cases in students in Shaanxi province, China: an intervention study

In China, almost every university has its own clinic, which provides medical care for common diseases, including small surgery and pharmacy services. Before initiating the intervention a survey was conducted to describe the baseline situation regarding TB control in the university hospitals. After the implementation of a cooperation mechanism between university hospital clinics and the local CDCs, the effects on the referral of TB suspects from university hospitals to the CDCs and tracing of TB suspects by the CDCs were assessed.

Inclusion criteria for universities were: having an in-patient clinic at the campus, having ≥ 5000 students, and willingness to participate as assessed before the start of the study. All but two of the 72 universities in Shaanxi province that fulfilled the inclusion criteria agreed to participate in this study: 47 universities in the provincial capital Xi'an, and 23 in other districts.

All students of the selected 10 universities in Shaanxi province that were registered at the universities between Jan.1, 2007 and Jan. 31, 2008 were eligible for inclusion in our study. Pulmonary tuberculosis suspects were defined as persons with cough, sputum production or haemoptysis for 2-3 weeks.

The five rates related to notification of university student TB suspects were referral rate, referral arrival rate, tracing rate, tracing arrival rate and total arrival rate. Table 1 shows the definitions and calculation methods of these rates.

We wanted to be able to minimally detect an increase from 50% (P1) to 80% (P2) in any rate after initiation of the intervention. Using a level of significance of 95% (α = 1.96) and a power of 90% (β = 1.28) the required number of TB suspects was calculated using the formula: N = (1.96+1.28)² * (((P1*(1-P1))+(P2*(1-P2)))/(P1-P2)². The required sample size was 48 suspects for the control and intervention group. An additional 20% was added to account for missing data, for a total of 58 suspects. We estimated that in half a year in five universities clinics, at least 58 TB suspects would present themselves.

First, three out of 47 universities were randomly sampled from Xi'an, and two out of 23 from other districts in the province. Thereafter, five universities were selected to match the sampled universities according to the location, type and size of the university. Of the matched pairs one was randomly allocated to the intervention group and the other to the control group.

The study period was divided into two periods, the baseline period (January-July 2007), and the intervention period (August 2007-January 2008). During the baseline period, university hospital directors and doctors, and CDC directors and supervisors of the TB control departments within the CDC were interviewed on the current status of TB control activities within and between these institutes and were asked for recommendations on how to improve TB control in universities by means of a structured interview. Also, baseline referral, tracing, and arrival rates of TB suspects among university students were assessed. In China, pulmonary tuberculosis suspects are defined as persons with cough, sputum production or haemoptysis for 2-3 weeks. Based on findings from the baseline survey, an intervention plan was developed and implemented in August-September 2007 in the intervention group: a TB control program was set up in the university hospitals, and staff members were assigned specifically to TB control. The provincial CDC provided training to the hospital staff members on TB control (diagnosis, reporting, referral, management of TB cases, and health education to students). Hospital doctors received a small incentive of 20 RMB (≈2.5 USD) when they referred a TB suspect to the CDC, and this suspect was diagnosed with TB at the CDC. At the end of the baseline and intervention period, referral and arrival rates were calculated using data from the paper and electronic medical records at the university hospital and the CDC.

Data were entered twice, by different persons, in EpiData v3.1 (Epidata Association, Denmark) and checked for discrepancies. The statistical software package SPSS version 13.0 (SPSS Inc, Chicago, USA) was used for data analysis. Differences between referral, tracing and arrival rates before and after the intervention were tested by means of the χ2-test.

The selected universities did not know which other universities were included in the study. During the study period, no meetings were held for the staff of participating universities. Also, the participating universities were not very closely located to each other. Therefore, we consider it unlikely that the control universities were aware of the intervention that was implemented and as such we do not expect spillover effects from the intervention universities to other universities.

Written informed consent was obtained from all TB suspects whom presented at the included university hospitals. The study was approved by the national ethical committee of operational research at the National Center for TB control. This study complied with the requirements of the Declaration of Helsinki.

All universities had a hospital at the main campus while five also had outpatient clinic(s) at one or more of their other campuses which were managed by the hospital. None of the hospitals could perform sputum smear examination but the hospitals in four universities had X-ray available for screening (Table 2). Five universities had campuses in different districts. Overall, the campuses were distributed among seven districts, 3 of which were in Xi'An city. Each district has its own CDC. Of the seven CDCs, five had a separate room for diagnosis of TB, a sputum smear test room, and X-ray room; and two cooperated with a hospital for TB diagnosis and treatment. In total there were 43 staff members for TB control in these seven CDCs (3-12 staff members per CDC).

Through interviews with university hospital and local CDC staff we identified some factors that impeded effective TB control for university students being implemented. CDC directors indicated that university hospital doctors were not trained on TB control due to limited human resources at the CDCs. Most hospitals had not assigned staff specifically for TB control tasks. No guidelines for patient registration, reporting, and referral were in place. University hospital staff was not familiar with NTP guidelines. Because of fear of spread and student panic in the campuses, the university usually asked university TB patients to suspend schooling and go home or to a hospital for treatment. These patients were never reported to the local CDC.

During the baseline period, not all conditions for good quality TB control were in place (Table 3). In the intervention group, the situation had improved after the intervention period while in the control group the situation was still similar. Six out of seven university hospitals (4/5 universities) in the intervention group had set up a department of TB control, and all the hospitals had assigned a physician for TB control; had formulated guidelines on reporting and referral of TB suspects; had established a cooperation with the CDC; and had arranged health education activities on TB for students.

Before the intervention, there were no significant differences in referral rate, referral arrival rate, tracing rate, tracing arrival rate, and total arrival rate between the control group and the intervention group. After the intervention, all rates were significantly higher in the intervention group compared to the control group, except for the tracing arrival rate (69.2% vs. 45.8%, p = 0.31) (Table 4). When comparing the control group before and after the intervention, there were no differences in these five rates, while the intervention group showed significant increases in all rates except in the tracing arrival rate compared to before the intervention (37.0% vs. 69.2%, p = 0.06, data not shown). During the baseline period, 48.9% (43/88) and 45.7% (42/92) (p = 0.67) of university student TB suspects in the control and intervention group, respectively, arrived and were managed at the CDC. After the intervention, 46.2% (30/65) and 93.2% (68/73) (p < 0.001) of TB suspects were managed by the CDC, respectively (Table 4). Among all 176 university student TB suspects who arrived at the CDC during the study period, 16 (9.1%) (data not shown) were diagnosed with smear-positive TB.

Non-arrival at the CDC in the control group and before the intervention in the intervention group was mainly due to TB suspects being referred to a general or TB hospital (72/130 = 55%) or sent home for treatment (26/130 = 20%) (Table 5). After the intervention, only 5 out of 73 TB suspects did not arrive at the CDC. Four of them were referred to a TB hospital and one went home.