Tuberculosis (TB) is a curable disease, yet 8.8 million persons worldwide are newly infected and 1.45 million persons die each year from TB [1]. Adherence to a standard course of treatment is paramount to achieving an effective cure of TB disease. Failure to cure persons with infectious TB disease promotes drug-resistant strains, increases patient risk for morbidity and mortality, and facilitates continued community transmission [2-5]. However, despite widespread adoption of the directly-observed therapy short-course (DOTS) strategy for over a decade, treatment adherence and default from treatment continue to be a primary challenge in TB control.

Home visits and community outreach are often employed as methods to optimize treatment adherence among TB patients. Previous research has consistently demonstrated the effectiveness of such patient tracing activities in returning patients to care and improving final treatment outcomes [6-11]. However, the ability to conduct these activities may be constrained by limitations in human and logistic resources, particularly in resource-limited settings with a high TB burden.

In South Africa, where the TB burden is considered the 3rd highest in the world [1], DOTS coverage is reported to be 100% [12]; yet, treatment success rates remain below global targets and the burden of drug-resistant TB strains continues to escalate [1].

In 2008–2009, the South African National TB Program (NTP) implemented a national pilot project, the TB Tracer Project, aiming to decrease default rates and improve patient outcomes, by creating teams dedicated to tracing TB patients [13]. The intervention has been described previously, but in brief, teams of health care workers were employed at health facilities to trace TB patients who had interrupted treatment or had missed scheduled follow up appointments for monitoring sputum status [14]. The project was implemented as a NTP programmatic intervention, therefore tracer team activities varied by sub-district. A recent evaluation utilizing data from routine national TB surveillance demonstrated that sub-districts that participated in the project (tracer) had significantly greater decreases in patient default rates and significantly greater increases in treatment success rates over the project period, compared to sub-districts that were not part of the project (non-tracer) (default rate decline 2.8% tracer vs. 0.7% non-tracer; success rate increase 2.6% tracer vs. 0.8% non-tracer)[13,14]. However, further examination is needed to better understand activities utilized for tracing patients and identify reasons that may help explain the differences between tracer and non-tracer sites.

The current study aimed to describe the knowledge, attitudes, and practices of health care and TB program personnel involved with tracing activities and to evaluate the differences between health facilities that participated in the TB Tracer Project (tracer) and health facilities that provided standard NTP patient services during the Project period (non-tracer).

Of 560 total questionnaires distributed, 270 were completed and returned (response rate 48%; Figure 1). Of the 270 returned, 10 were from TB managers at the district TB office, 27 were from sub-district TB personnel, 120 from tracer TB managers and 113 from non-tracer TB managers. Response rates were similar among tracer and non-tracer TB managers at health facilities (48 vs. 46%, respectively). An additional 25 questionnaires were returned from individual TB tracer team leaders, yielding a 35% response rate (25/72). A total of 295 questionnaires were available for the current study analysis.

This study examined the knowledge, attitudes and practices utilized to trace TB patients in South Africa, both among personnel involved with the TB Tracer Project and those that manage TB patients and may conduct tracing as part of standard care within the NTP. This study elucidated specific concerns and aspects that should be considered when integrating tracing activities into routine TB care. Overall, study participants demonstrated a high level of knowledge on core questions regarding the etiology, diagnosis, treatment, and prevention of TB. There was a similar level of overall knowledge between tracer and non-tracer facilities, suggesting that basic TB knowledge did not factor into previously reported differences on the impact of tracing activities [13,14].

The tracer teams were regarded as an effective method for returning patients to treatment, even with the challenges reported by the tracer teams; these findings are similar to previous research [16]. Practices employed by tracer teams focused on returning patients to treatment, patient education and the development of future treatment adherence plans. Non-tracer facilities also reported that they conduct some tracing activities, however, we were not able to assess the frequency of patient home and workplace visits by tracer facilities teams compared to non-tracers. A greater proportion of non-tracer health facilities reported using phone calls to trace patients and non-tracer facilities were more likely to contact the family members of patients than tracer facilities teams. Non-tracer facilities identified lack of human resources as the top reason why they do not routinely trace patients.

While respondents averaged 9.8 years of experience working in TB, it is of concern that just over half of each respondent group (range 50–59.3%) was able to correctly identify the four components of a DOT encounter, namely, verifying the medication, watching the patient take pills, checking for side effects, and documenting the visit. However, the lack of understanding of DOT may be due to only a fraction of respondents reporting having formal training on DOTS (range 53.1-90%). Our findings are similar to those reported from a survey of private practitioners in Ethiopia, wherein only 39.3% had satisfactory knowledge of DOTS [17]. The study also reported that practitioners who had received training on DOTS within the previous 2 years were over four times more likely to have satisfactory knowledge on TB treatment. These findings underscore the critical need to ensure that all personnel involved with TB treatment and monitoring receive comprehensive training on TB care and treatment.

When asked to identify reasons patients fail to adhere to treatment, over two-thirds of respondents cited that patients no longer feel sick, and over a quarter mentioned patient personal or cultural beliefs, side effects of treatment, fear of stigma, lack of knowledge and lack of access to health care. A review of research on patient adherence revealed similar factors leading to non-adherence and default, citing interpretations of illness and wellness as a major factor wherein patients who feel better no longer believe that they are ill and stop taking their medication [16]. Investigators also pointed out that knowledge and attitudes affect adherence, similar to responses expressed by TB personnel in the present study of patient personal and cultural beliefs creating a barrier to adherence. The findings from the current study, in concert with previous research, highlight the need to include patient education emphasizing the importance of completing treatment to effectively cure TB disease as part of standard practice for treating and tracing TB patients.

Activities conducted upon return to the clinic are also important in patient adherence and reuptake of treatment after default. When a patient returns to the clinic, personnel at tracer facilities were more likely to discuss alternative DOTS arrangements with patients than non-tracer facilities. It is possible that these encounters assessed individual barriers to treatment adherence and identified mutually-agreeable methods for providing TB treatment to these patients, which would favor treatment adherence and successful outcomes [10].

Of the tracer team leaders that returned questionnaires, the majority viewed their relationship with health facility staff as very good or excellent, but they also expressed some major concerns. Their concerns included lack of logistical support and supervision and direction from the district or provincial TB managers. They also expressed concerns over their personal safety, both in terms of physical protection in the areas where they were tracing patients and contracting TB from patients. These negative experiences help to identify areas for improvement of future tracing activities in South Africa and other high burden countries.

Some limitations and biases exist in this study. This study was based solely on qualitative and quantitative data from questionnaires. The questionnaires were sent retrospectively with an extended delay from the time of program implementation making this study subject to recall bias. It is possible that some respondents were no longer in the same positions, and those that were may have trouble recollecting activities conducted during the TB Tracer Project. Reported activities may more accurately represent current activities in the program, which may have changed since the project period. However, it is unlikely that the ability to recall information was substantially different between the different groups of respondents. The overall response rate was very low, subjecting our results to bias if respondents were distinctly different than persons that did not complete and return the questionnaires. Our findings may therefore not be representative of the total target population. Our efforts to improve response rates included calling and emailing districts and individuals to remind them to return the questionnaires. Many of the current district staff were not involved at the time of the tracer project and therefore didn’t send feedback and were not included in the study. However, there was little variation in response rates across the provinces that participated in the evaluation. Additionally, it was not possible to evaluate the success among tracer teams to get patients to return to treatment or final patient outcomes, as no information was collected prospectively to track patients identified for tracing individually and the outcome of tracing efforts. The differences revealed in this study between tracer and non-tracer sites are unlikely to have made any substantial contribution to the greater improvements in treatment outcomes at tracer sites compared to non-tracer sites during the TB Tracer Project period [13,14].

Despite limitations, the current study presents a useful assessment of core TB knowledge among TB program and health facility staff and of practices utilized to increase adherence among TB patients in South Africa. Future studies and health interventions using TB patient tracer teams should include ongoing TB practitioner and tracer trainings. These should stress the need for increased patient education that emphasizes the critical importance of completing treatment irrespective of clinical symptoms. A clear evaluation strategy prior to implementation is also essential in order to accurately understand the relationship of these factors to changes in default and cure rates. Ongoing monitoring and documentation of the frequency of tracing activities will also help to describe differences observed in patient treatment outcomes. These components are vital to identify patient and programmatic factors that may be instrumental in optimizing the success of TB tracing activities.

The study also identified factors that must be considered to optimize the success of tracing activities, including the need for ongoing TB staff education and provisions to ensure physical and personal safety during tracing. The challenges that tracer teams experienced including poor logistical support and fear for safety need to be carefully considered when planning for tracer outreach teams.

TB: Tuberculosis; NTP: National TB Program; WHO: World Health Organization; MDR: Multi-drug resistant; DOTS: Directly observed treatment short-course.

The authors declare that they have no competing interests.

LJP, LEB, MV, AP and LDM conceived of the study, and participated in its design and coordination. NB participated in the design of the study and follow up to improve response rates. CCB, LJP and LEB developed the analytic plan, and CCB carried out the data cleaning and analysis with support from LJP and LEB. CCB, LJP, and LEB were responsible for the interpretation of results. CCBwas the primary author for writing the manuscript; LJP and LEB also contributed to the writing. All authors read and approved the final manuscript.

The pre-publication history for this paper can be accessed here:

http://www.biomedcentral.com/1471-2458/13/801/prepub