Background
Despite the recent progress of global efforts, tuberculosis (TB) is still one of the leading causes of morbidity and mortality world-wide, and remains as a major public health burden in many developing countries [
1]. Current anti-tuberculosis therapy consists of a cocktail of drugs taken over a period of at least 6 months for new patients and 8 months for retreatment patients. Because of the long duration of the therapy, there is a risk of treatment interruption or default, a phenomenon that contributes to prolonged infectiousness, drug resistance, relapse and death [
2,
3]. The difficulty experienced by patients in following treatment regimens has raised the awareness of adherence as a complex behavioral issue [
4]. Efforts to improve treatment outcomes require a better understanding of particular barriers to and facilitators of patient's adherence [
5].
Studies of socioeconomic and behavioral factors affecting adherence have been conducted previously [
6‐
9]. In Hong Kong, China, a study of 102 defaulters matched to 306 controls indicated that tobacco smoking, a history of prior treatment default or poor adherence, treatment side effects, and subsequent hospitalization were associated with treatment default [
10]. A study in Fujian, China, that combined quantitative and qualitative methods reported that treatment adherence was associated with the intention of patients and the behavior of health service providers, but not with gender, age, career, education level or social stigma [
11]. Another study in Chongqing, China, which involved interviewing patients and health staff, indicated that additional tests and drugs, especially liver protection drugs, may entail considerable financial barriers to starting and continuing treatment [
7]. However, these aspects of treatment adherence have not been studied previously in Jiangsu Province of China. Therefore, to identify the social context, patient characteristics, and health system factors affecting patient's adherence to anti-tuberculosis treatment in Jiangsu Province, we conducted a study using both qualitative and quantitative methods. Our goal was to provide policy-makers with recommendation for more organized TB control program to improve the adherence to anti-tuberculosis treatment.
Methods
Study sites
This study was carried out in Jiangsu Province, which is located along the eastern coast of China and covers an area of 102.6 thousand square kilometers, about 1% of the total area of the country. Jiangsu Province contains 13 municipalities and 106 counties (districts), with a total population of 74 million. The population density is 736 per square kilometer, the highest among all provinces of China. The annual net per capita income of farmers in Jiangsu Province is 6561 Yuan ($964), and the annual employee salary is 27234 Yuan ($4005). The average life expectancy of the local population is 75.3 years, with men at 72.9 and women at 77.9 years, respectively. DOTS (direct observed therapy, short course) strategy for TB was introduced in the 1990s and is now 100% available at the county level. An internet-based surveillance system was set up in 2004, and all newly diagnosed TB cases are required to be registered with the local TB dispensary and reported to upper level health authorities.
Data collection
This study was designed to use both quantitative and qualitative methods in order to gain insights into the factors that could contribute to adherence. A multi-stage sampling strategy was implemented for the quantitative aspects of the study. We selected 13 municipalities as the first sampling unit. In each municipality, one county (district) was randomly selected as the study site. In each site, 60 sputum-smear-positive TB patients registered since 2006 were consecutively selected as study subjects. The total estimated sample size was 780, given a confidence interval = 95%, an estimated non-adherent proportion = 20%, a relative precision = 0.2P, design efficiency = 2, and the number of study sites = 13. After obtaining informed consent, trained local TB dispensary staff interviewed all participants with a structured questionnaire, including basic characteristics, socioeconomic status, treatment history and adherence to anti-tuberculosis treatment. Non-adherent patients were further presented with 16 options for reasons of their non-adherence. In this study, observed treatment was divided into four categories: self-administered (patient took drugs without external observation); observed by family members or others (patient took drugs under the observation by family members or other volunteers); home-based drug delivery (local health workers sent anti-tuberculosis drugs to patients' home regularly, but did not observe them taking each dose of drugs); and directly observed by village doctors (patients took drugs under the direct observation by village doctors each time).
For the qualitative study, 20 TB patients (from Xuanwu and Taichang, 15 men, 5 women; 18 newly treated, 2 previously treated) were invited for in-depth interviews based on a convenience sampling strategy. A semi-schematized guide covering general as well as specific questions was utilized. The in-depth interview focused on patients' health-care seeking history, knowledge and attitudes towards TB, and their adherence to TB treatment. Ten local health workers from Taichang, including 5 village doctors and 5 community hospital doctors, were also invited for in-depth interviews. Themes for the in-depth interviews with these doctors included the current TB control program in the village/community, the main problems with and suggestions for the current TB program, treatment adherence of patients in the corresponding village/community, and the main reasons for non-adherence. Interviews were conducted in local health facilities by trained professionals from Jiangsu Provincial Center for Disease Prevention and Control with Chinese Mandarin. They lasted 20-40 minutes and were tape-recorded with permission.
Data analysis
In this study, patients who had missed 10% or more of the total prescribed dose of TB drugs were deemed as non-adherent. For the quantitative study, data were entered in Epidata (Denmark) and analyzed using STATA 10.0 (College Station, TX, USA). Associations between selected factors and non-adherence were estimated by computing odds ratios (ORs) and their 95% confidence intervals (CIs) from an unconditional logistic regression model. Predictive variables that were independently significantly associated with treatment completion in univariate analysis were included in a multiple logistic regression model to determine their relative contributions in predicting treatment adherence while simultaneously adjusting for each of their effects. The criterion for significance was set at P < 0.05 based on a two-sided test. Continuous variables, including age and income, were converted to dichotomous variables using the median as the cutoff point.
For the qualitative study, content analysis was applied [
12,
13]. Tape-recorded in-depth interviews were firstly transcribed in Chinese characters and then translated into English by two trained staffs in the Jiangsu Provincial Center for Disease Prevention and Control. Codes were then developed based on the original terms used by participants. The transcripts and notes were analyzed thematically by categorizing the interview data under the main topic headings. The codes were then presented, discussed and checked within the research team. Tentative categories and sub-categories were created from the clustered codes, and subsequently main themes emerged based on the patterns and relationship between the categories.
Ethical consideration
This study was approved by Ethics Committee in Center for Disease Control and Prevention. Oral informed consent was obtained from all participants.
Discussion
Early detection of patients and providing effective treatment are the main interventions to prevent the spread of TB. However, current long-term anti-tuberculosis therapy could easily lead to patient non-adherence, which presents an important barrier for TB control programs [
8,
14]. Patients' adherence to their medication regimens has been reported to be influenced by the interaction of number of factors [
8]. These various factors may be grouped as: health-system factors, social and family factors, and personal factors [
15]. The factors that influence patient adherence to TB treatment vary in different populations, and those that emerged in our current study might be similar to, or different from, those reported in other areas of China.
The financial burden on TB patients was one of the key issues we found to be associated with non-adherence in our study. Although the government of China has established a "free TB service policy" with the aim of decreasing the financial burden on patients, this free policy seems to be not as well known or well implemented as planned in some areas [
12,
16]. In a previously reported study in four provinces of China [
17], TB patients were obliged to pay 12-40% of their annual income for TB services, despite the fact that all smear-positive and some severe smear-negative patients received free drugs. It is believed that the heavy financial burden on patients is one of the main reasons that some TB patients fail to access and complete the treatment. The main financial burdens, as evidenced in the present study, are the extra cost for medical examinations, the need to purchase liver protection drugs, and hospitalization costs.
Conflicts of interest may be an underlying reason for the existence of heavy financial burdens on patients in the presence of the free treatment policy. Briefly, with the transformation from a traditional planned economy to a market-oriented economy beginning in the early 1980s, a larger part of rural health facilities in China were either totally or partially privatized [
18]. These rural facilities have developed a variety of means to attract patients, in an effort to generate more revenues by providing more clinical services and selling more drugs [
18]. Even in the state-owned hospitals, a fee for service and bonus-related revenue system has been adopted to encourage medical staff to make more money, a process that can increase the economic burden on patients, elevate health-care costs, and ultimately impede effective TB control in China.
Drugs are frequently prescribed in some countries, and especially in China, to protect the liver during treatment for TB [
19]. A systematic review based on 85 research articles, which evaluated 30 distinct types of putative liver protection compounds, found no reliable evidence to support prescription of drugs or herbs for this purpose to people on TB treatment [
19]. However, a variety of presumed liver protection drugs are still widely used in China and their purchase accounts for a large part of patients' medical expenditures. On one hand, doctors prescribe these drugs with the hope that they will prevent hepatic injury induced by anti-tuberculosis treatment. On the other hand, doctors generate more revenue by selling these drugs, as liver protection drugs are not freely provided under the current TB program. To quote one doctor in our study, "if we only deliver free anti-tuberculosis drugs rather than selling private expense drugs to patients, how can we get the bonus?"
Adverse reactions to current TB treatment, including hepatic injury, also contribute to non-adherence, as has been described in numbers of studies [
8]. Xia reviewed reports published on this topic between 1996-2005 in China and found that the overall incidence of anti-tuberculosis drug induced adverse reactions was 12.6%, with hepatic injury the most common one (mean = 11.90%, median = 15.66%, range 0.75%-71.43%) [
20]. In our present study, adverse drug reaction to TB treatment was also listed by patients as the most common factor (37.8%) associated with non-adherence. Thus, active surveillance of patients' adverse reactions would be far preferable to merely prescribing possibly ineffective liver protection drugs. To minimize the impact of adverse reactions, it is important that health staff provide concise pretreatment counseling to patients and that they manage such side-effects with timely recommendations and services [
21].
TB was originally named "Phthisis" and regarded as a deadly disease in China. With the development of modern medicine, TB is no longer an incurable disease. However, TB patients still experience stigma and are always isolated from the community. Although the Chinese government has carried out massive education programs in recent years, knowledge of TB among the general population is still limited [
16]. Social support can help patients overcome structural barriers as well as personal barriers, and community and family members' attitudes may influence a patient's decision whether to stop or continue TB treatment. In such circumstances, community-based TB treatment programs and stronger involvement of local social networks to support TB patients may be justified [
6].
Direct observation of treatment (DOT) has been recommended by WHO to enhance patients' adherence and is regarded as a main component of the "breakthrough" in TB control programs [
22]. However, DOT is not implemented as well as expected in some areas of China [
7,
23,
24]. A study in Chongqing reported that less than 5% of TB patients had been treated under direct observation by health staff, and 12.5% of patients reported "interrupted treatment" [
7]. In Shandong Province, only 21% of patients took their medicine under the surveillance of village doctors, and this proportion varied widely across the counties, from 0% to 70% [
23]. Our findings from this study also proved that DOT is not well accepted in some rural areas of Jiangsu. Other alternatives need to be studied as possible ways to reinforce adherence in patients who do not accept direct observation treatment in village clinics. In Jiangsu Province, patients usually visit the TB dispensary monthly and get the anti-tuberculosis drugs for the whole month. They swallow the drugs under the direct observation by doctors or family members or supervise by themselves. Due to different reasons, some patients might not attend the regular visits to TB dispensary, resulting in the interruption of treatment. In such a case, local health workers are responsible for delivering drugs to patients' home and perform patient retrieval. Strictly, it is not a real direct observation because health workers don't observe patients swallowing each dose of drugs. But it is still believed to be an alternative way to help increase the adherence, given that DOT is not accepted by the patient. A centerpiece of the DOTS strategy is that it shifts responsibility for completion of therapy and successful clinical outcomes to the healthcare provider rather than to the patients themselves [
25]. It is clear that good health services are necessary, but not sufficient, to ensure treatment success [
15]. Patients still need to choose to take drugs. Building supportive patient-doctor relationships, rather than solely relying on authoritarian supervision, can best improve treatment adherence and TB control [
26].
Several methodological issues should be discussed. Firstly, results of this study come from Jiangsu Province and may not represent the conditions in all of China. Secondly, 14.1% of selected subjects could not be traced and were not involved in the analysis, raising the possibility of some selection bias. Thirdly, data on treatment history and possible explanatory factors in this study were based on self-report of patients. Recall bias and investigator bias were thus unavoidable.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
WX, WL, YZ and LZ conceived the idea and implemented the field study. WX, HS and JW participated in the statistical analysis and wrote the manuscript. All authors read and approved the final manuscript.