This study used data from the project on treatment and effectiveness evaluation of LTBI among college students in Shandong province, China. More information about procedures has been described in detail in our previous publication [21]. The study began in September 2020 and was conducted in 6 cities (Jinan, Qingdao, Yantai, Rizhao, Linyi, and Zibo) using a cluster random sampling method to select samples. We selected students aged ≥18 years with positive purified protein derivative test results (ie, induration diameter ≥10 mm or accompanied by blisters, papules, double circles, and lymphatitis) to participate in our questionnaire survey. We excluded patients previously diagnosed with active pulmonary TB or extrapulmonary TB as well as those with a history of mental illness or epilepsy. Respondents were instructed to fill out the electronic survey questionnaire by trained graduate students to ensure the quality of the questionnaire. In total, 1691 individuals were eligible, and 1631 completed the whole survey.

We used the following formula to calculate the sample size:

In this formula, n, δ, and p are estimations for sample size, allowable error, and population rate π, respectively. Because acceptance of LTBI treatment among college students with LTBI is not clear in China, we used the positive prevalence of the tuberculin skin test (TST) to calculate the sample size. The prevalence of positive TSTs in Chinese college students using 10mm cutoff value was 9%-49% [22,23]. We set P=.09, α=.05 (2-tailed), and δ=0.009. Considering the sampling method and the 20% lost follow-up rate, the TST screening sample was finally determined to include 20,330 people [21].

This study protocol received ethics approval from the Ethical Committee of the School of Public Health in Shandong University (LL20200306). The investigation was conducted after the acquisition of written informed consent from all participants.

First, a descriptive analysis was used to describe participants’ characteristics with mean (SD) or median (IQR) values for continuous variables. Frequency and percentage values were used to describe categorical variables. Second, Pearson chi-square test (for categorical variables) and 2-tailed student t test or Wilcoxon rank-sum test (for continuous variables) were used to compare characteristics of male and female students. Then, since 1547 participants came from 6 cities and 16 colleges, multilevel mixed-effects logistic regression (Multimedia Appendix 3) was performed to examine the moderating role of gender. We added the multiplicative interaction term (perceived stigma toward TB × gender) in models. The association between perceived stigma toward TB and acceptance of LTBI treatment was also assessed by multilevel mixed-effects logistic regression. In model 1, no covariate was included. Model 2 was based on model 1, with additional adjusting for age, type of student, type of residence, father’s level of education, mother’s level of education, household income, smoking, current drinking, physical activity level, exposure history of TB, boarding experiences, and knowledge of TB. Odds ratios (ORs) and 95% CIs were presented as measures of effect. A 2-tailed P<.1 was considered as statistically significant. We conducted all analyses in Stata (version 14.2; Stata Corp).

In this study, only 723 (46.7%) college students diagnosed with LTBI were willing to accept preventive treatment. This figure is lower than what has been previously reported for other students’ acceptance rates of LTBI treatment, ranging between 63.9% and 87% in China [28,29]. The possible reason is that participants in this study were general college students, while previous research focused on students who had close contacts with TB and was related to the screening and management of students’ close contacts with TB [30]. For close contacts, schools actively organized screening and recommend LTBI treatments. Close contacts were more aware of TB than the student and more likely to accept LTBI treatment for fear of developing TB.

Our study observed a higher acceptance rate of LTBI treatment among female students than male students, contrary to previous studies in the United States and Canada [31]. One possible explanation for our findings is that female students have a higher level of TB knowledge than male students. Previous studies found that patients with LTBI who had low TB knowledge scores in student TB contacts or TB-designated hospitals were less likely to accept preventive treatment [28,32]. Students with higher knowledge of TB had some knowledge of the spread, harm, and prevention of TB. They wanted to reduce their risk of developing TB. When health workers recommended LTBI treatment, female students were more likely to accept it.

Our results showed that higher perceived stigma toward TB was positively related to acceptance of preventive treatment among college students with LTBI. This did run counter to our initial hypothesis, wherein we assumed that perceived stigma toward TB may be negatively associated with acceptance of LTBI treatment. No direct evidence has shown this association among college students, but prior studies found that TB-related stigma was negatively associated with LTBI treatment acceptance in other populations (eg, refugees and migrants) [12,13]. On the one hand, we speculate that the difference is due to people’s surroundings. College students live in a trustful and inclusive environment. College students did not score high on perceived stigma toward TB in our study, suggesting that they did not have embedded fears surrounding TB. Of course, it could also be that they did not know much about TB. After all, the TB knowledge score of college students in the study was 19.0 out of 45. However, TB stigma is relatively high in the community. Immigrants may feel singled out and stigmatized for other reasons [33,34]. One study in Eritrea [35] showed that restrained contacts due to fear of getting infected with TB, gossiping, and finger pointing in the community caused enacted and anticipated stigma among patients with TB. These findings indicate that the relationship between perceived stigma toward TB and acceptance of LTBI treatment may be quite opposite in different populations. On the other hand, LTBI is still different from TB infection. Although participants experienced enforced stigma, which could impede treatment initiation, their strong motivation to prevent becoming sick with TB resulted in acceptance of LTBI treatment.

It is worth noting that perceived stigma toward TB was positively associated with acceptance of LTBI treatment in male students, but there was no association between the stigma toward Tb and LTBI treatment acceptance in female students. Similar to previous studies on the impact of gender differences on TB epidemiological characteristics and access to services [15,16], we found gender differences played a part in the association between perceived stigma toward TB and acceptance of LTBI treatment. One possible factor that accounts for the disparity could be the different temperament and personality characteristics of male and female students. Women tend to be cautious, careful, worried, or pessimistic; men are more carefree, relaxed, and upbeat, even when most people are worried [36,37]. Therefore, male students may be more likely to accept LTBI treatment after feeling perceived stigma toward TB, whereas female students are not. Another plausible explanation is that the differences are related to men’s unhealthy behaviors, such as smoking and drinking. Studies found men were overrepresented in various pulmonary TB risk groups, notably alcoholic and tobacco smokers [16,38]. In this study, the proportion of male students who smoke and drink alcohol far exceeded that of female students. Despite higher perceived TB stigma, male students with LTBI were more likely to accept preventive treatment for fear of developing active TB. However, these pathways are speculative. Future research can explore the underlying mechanisms of the relationship between perceived stigma toward TB and LTBI treatment acceptance.

Perceived stigma toward TB was not associated with acceptance of LTBI treatment in female students. The reason may be that perceived stigma is not a key factor in female students accepting preventive treatment. In our study, the proportion of female medical specialty students was higher than that of male students. Moreover, female students have similar stigma scores but higher TB knowledge and LTBI treatment acceptance scores compared to male students. Of the added covariates, we found that exposure history of TB (Yes: OR 3.00, 95% CI 0.98-9.16; P=.05 and Unknown: OR 1.61, 95% CI 1.10-2.37; P=.02) and TB knowledge (OR 1.02, 95% CI 1.00-1.05; P=.08) were positively associated with acceptance of LTBI treatment among female students. In the future, research may focus on the factors influencing the acceptance of preventive treatment in women.

The current findings have some policy implications. Gender may be an important heterogeneous factor influencing the acceptance of LTBI treatment in colleges. It is recommended that colleges regularly and continuously carry out various forms of TB and LTBI knowledge publicity and education activities to raise students’ awareness of TB. The perceived stigma toward TB among male students needs to be addressed in improving their treatment acceptance. It is suggested that colleges provide suitable psychological counseling for male students. Overall, perhaps gender-specific management can improve the acceptability of LTBI treatment among college students.

This study has some limitations. First, the results were based on colleges in Shandong Province. Our findings were not necessarily generalizable to other settings within or outside China. Moreover, cross-sectional data used in this study were impossible to demonstrate the causality between perceived stigma toward TB and acceptance of LTBI treatment. Establishing and analyzing a multiyear longitudinal cohort of colleges covering different geographic regions in future studies could better elucidate the impact of gender on the relationship between perceived stigma toward TB and willingness for LTBI treatment.

In conclusion, we found that the acceptance rate of preventive treatment among college students with LTBI was low. The acceptance rate of female students was higher than that of male students. There were gender differences in the association between perceived stigma toward TB and acceptance of LTBI treatment. Only among male students, those with high perceived stigma toward TB were more likely to accept LTBI treatment. This study implies that gender-specific strategies in colleges may be effective in improving the acceptability of preventive treatment.