Treatment outcomes and factors affecting treatment outcomes of new patients with tuberculosis in Busan, South Korea: a retrospective study of a citywide registry, 2014–2015

Despite efforts to eliminate tuberculosis (TB), the incidence and mortality of TB remain high worldwide. In 2016 alone, there were 10.4 million new TB cases globally, and 1.7 million people died from the disease [1]. South Korea is not free from TB, despite having abundant medical resources. In 2017, a total of 36,044 patients in South Korea infected with TB (70.4/100,000 population) were listed on the Korean National TB Surveillance System (NTSS), and 78% of them (28,161 patients; 55.0/100,000 population) were new patients [2]. A total of 2186 patients died from TB in South Korea in 2016 [2]. South Korea has the highest TB incidence and mortality rate of all member countries of the Organization for Economic Co-operation and Development [3], and TB remains a serious public health problem in South Korea.

The “End TB Strategy” of the World Health Organization (WHO) has set a target to increase the TB treatment success rate to 90% by 2025, and to reduce TB deaths by 95% (compared to 2015) by 2035, to end the global TB epidemic [4]. To facilitate these goals and evaluate the effectiveness of the strategy, it is essential to accurately record and report the treatment outcomes TB patients. Treatment outcomes, as measured by a standardised method, are key indicators of national TB program (NTP) effectiveness [5]. In addition to monitoring treatment outcomes, improving treatment success rates is important and depends on identifying vulnerable populations and the risk factors for poor treatment outcomes. This information is important for policymakers in terms of resource planning, prioritisation and distribution.

To achieve the WHO’s TB targets, South Korea has made many changes to its NTP since 2011. The Korea Centers for Disease Control and Prevention (KCDC) outlined the “New 2020 Plan,” which set a target to cut the TB incidence rate in half by 2020 (< 50/100,000 population) [3]. Several programs were implemented to reach this goal, including expanded public-private mix (PPM) collaborations, reinforcement of outbreak investigations, contact investigations and treatment of latent TB infection, forced hospitalisation for non-adherent patients and those with multidrug-resistant (MDR) TB, and making all hospital visits for diagnosis and treatment of TB free-of-charge.

Despite these efforts to improve TB treatment in South Korea, there are limited data regarding recent treatment outcomes of South Korean TB patients and little is known about the effects of the 2011 policy changes. Factors affecting treatment outcomes of South Korean TB patients are not well understood. Most previous studies included only a small number of patients, were conducted in a single institution, or were conducted solely in the private sector [6‐8]; therefore, the overall status of TB in South Korea is poorly understood. In the present study, we investigated treatment outcomes of new patients with TB in Busan, which is one of the largest and most TB-prevalent cities in South Korea, using a citywide registry. We also evaluated the factors affecting treatment outcomes of TB patients.

To our knowledge, this is the first large-scale study to investigate the treatment outcomes, and factors affecting outcomes, of new patients with TB in Busan, South Korea. A major strength of our study was that it included a large cohort of TB patients, drawn from the population of a major city. Another strength is that we were able to include patients who transferred to other hospitals within the city, by using patient identification numbers to determine treatment outcomes. The Korean NTSS successfully registered 94.0% of all TB cases in 2014 [13], so it is likely that our results are representative of the overall TB status of Busan.

During the study period, the treatment success rate of new TB patients in Busan was 83.9%, which was similar to that of South Korea as a whole (84.4% for 2014 and 83.8% for 2015) [14]. However, neither rate met the 90% target set by the WHO. This sub-optimal treatment success rate was mainly due to high mortality (8.0%). One of the main risk factors for death of TB patients is old age. A lower in treatment success rate due to the high incidence of death of elderly TB patients is typical of both low- and high-TB incidence countries [15‐17]. Several factors are associated with death in elderly TB patients. Besides aging per se, high mortality in elderly TB patients may be due to age-related comorbidities, reduced immune function, missed diagnosis and delayed treatment initiation due to atypical presentation of TB, and difficulties in accessing healthcare services [18‐21]. In our cohort, elderly TB patients (≥ 60 years old) accounted for nearly 40% of all of the patients and, not surprisingly, old age was an independent risk factor for death. The number of elderly TB patients is expected to increase as life expectancy continues to improve, which presents major challenges to TB control in South Korea [22]. In addition to old age, male gender and PTB (especially smear-positive PTB) were also independently associated with high mortality in our study, which is consistent with previous reports [17, 23]. Higher death rates in male patients may be due to behaviours such as alcohol consumption, smoking, and non-adherence to treatment [24‐27]. Patients with smear-positive PTB may have advanced disease. Treatment outcomes being reported by a hospital was another risk factor for death in our study, but this could be due to selection bias: a considerable number of TB patients who have chronic critical illness and limited options for treatment are transferred to such hospitals in the post-infectious stage, likely leading to higher acuity of hospital patients. Unexpectedly, Q5 of the regional deprivation index was associated with lower mortality. However, this was attributable to the higher rate of loss to follow-up in the Q5 group; we estimate that a significant proportion of the patients lost to follow-up in the Q5 group were indeed dead.

To reduce death from TB, interventions for elderly patients should be prioritised. Elderly patients are more likely to show an atypical presentation of TB, so actively searching for cases would facilitate early diagnosis and timely treatment, which would in turn likely reduce mortality [28, 29]. The standard sputum AFB smear has low sensitivity in elderly patients, making TB detection more difficult, but molecular tests such as the Xpert MTB/RIF assay could overcome this problem [30]. Administrative support, such as financial support and help with transportation, could improve access to healthcare services for vulnerable elderly patients. TB-related deaths are common in the intensive treatment phase [7, 16], so intensive monitoring early in this treatment period should be considered not only for elderly patients, but also for patients with other risk factors for death (e.g. male, smear-positive PTB, non-adherence to treatment).

The number of elderly TB patients is expected to increase in the future, and efforts to reduce mortality of elderly patients may be stymied by the inherent demographic characteristics of this population. In addition, all new patients with TB had low rates of treatment failure which would be difficult to decrease further. Therefore, to achieve WHO targets, a paradigm shift in TB control strategies may be necessary to reduce non-adherence to treatment among patients in Busan. Non-adherence results in serious problems not only for the patients themselves (including development of complications, acquired drug resistance, and death), but also for communities in terms of the spread of infection [31]. In our study, male gender, foreign nationality, smear-negative PTB and EPTB, and treatment outcome reported by institutions other than public health centres were risk factors for loss to follow-up, which was consistent with previous studies [32‐35]. The number of TB patients of foreign nationality has increased in South Korea (from 849 patients in 2010 to 2045 in 2017) [2]. The higher rate of loss to follow-up in these foreign national populations may be due to difficulties in accessing healthcare services due to legal, employment, economic, or language problems [33]. The higher rate of loss to follow-up in cases of EPTB may be due to excessively long treatment durations, lack of motivation due to lower symptom severity compared to PTB, or reduced attention from TB control programs due to lower infectivity [36, 37].

To improve patient compliance, PPM collaborations should be expanded throughout South Korea, as well as in Busan specifically, to enhance TB control in the private sector. A previous South Korean study showed that PPM collaborations decreased the rate of loss to follow-up in the private sector [38]. For non-adherent patients of foreign nationality, international collaborations to ensure continuity of treatment after departure could be useful, in addition to administrative support to improve access to healthcare services [39]. Patient-centred care is also important for reducing non-adherence to treatment. This may include providing sufficient information on TB to patients, improving communication between patients and healthcare providers, and providing social support to help overcome the stigma of TB. Although studies of the efficacy of directly observed therapy (DOT) have reported inconclusive results, DOT may be considered for non-compliant patients [3, 35, 40, 41].

There were several limitations to this study. First, socioeconomic parameters that could affect patient treatment outcomes were not recorded (e.g. smoking status, alcohol consumption, level of education, degree of employment, and economic status). We used a regional deprivation index as an indirect indicator of socioeconomic status, but it showed inconsistent associations with treatment outcomes. Because the regional deprivation index is based on the data of a given area (town level), not a patient, it might not reflect the socioeconomic status of the individual patient. Also, Busan is one of the most highly developed metropolitan cities in South Korea, so differences in regional deprivation might not be particularly significant. Further, socioeconomic differences may be diluted by free-of-charge hospital visits for TB and relatively easy access to healthcare services for all patients in Busan. Second, we did not include previously treated patients, or those with RR- or MDR-TB in the analyses. These patients are well known as high-risk groups for poor treatment outcomes. Further investigations are needed on these vulnerable populations. Finally, we could not distinguish between TB-related and non-TB related deaths. Although the WHO defines TB death as that occurring during TB treatment, irrespective of cause, knowing the actual cause of death may help to improve TB control programs.

Treatment outcomes of new TB patients were found to be sub-optimal in Busan, South Korea. To reduce mortality and loss to follow-up rates, and therefor increase the treatment success rate, TB control programs should involve close monitoring and provide more socioeconomic support to patients at high risk for poor treatment outcomes.