This single-center case–control study retrospectively reviewed the data of 2,072 patients aged more than 18 years who were registered and treated for pulmonary TB or extrapulmonary TB (EPTB) at Chonnam National University Hospital from January 2010 to December 2014. TB was diagnosed on the basis of Mycobacterium tuberculosis identification by culture or polymerase chain reaction (PCR) from clinical specimens or clinical, radiological, or histological findings compatible with TB and responses to TB treatment. All TB patients received TB health education as well as monitoring for drug compliance and adverse drug reactions during treatment by specially trained nurses who participated in the Public–Private Mix (PPM) project for TB control in Korea.

TB drugs were self-administered with the support of trained PPM nurses during treatment. At each clinical visit, treatment adherence was assessed by asking patients about their drug intake and checking the remnant tablets of the previously prescribed regimen.

Once TB was diagnosed at our institute or a TB patient was transferred from another clinic, the patient was enrolled in the national TB registry. Physicians reported the patients’ data to the registry, and PPM nurses entered the data into registry database and updated it as necessary. The following patient information was obtained through electronic medical records: baseline clinical data, comorbidities, history and outcome of previous TB treatment, radiologic findings at diagnosis, microbiological characteristics at diagnosis (acid-fast bacilli stain, mycobacterial culture, and PCR in clinical specimens; drug resistance; and culture conversion after treatment), and side effects during TB treatment. Employment was categorized as “regular” if the patient was employed in a year-round position or was a full-time student and “irregular or unemployed” if the patient had a seasonal or day-labor position (12). Education was classified into three groups according to the highest level the patient completed: elementary school, middle or high school, and college or university education was classified as primary, secondary, and post-secondary, respectively. Income was categorized on the basis of the average monthly income of the patient or their household in Korea with reference to national statistics (Korean Statistical Information Service, http://kosis.kr) according to a cut-off corresponding to the fourth quartile. Health insurance was categorized into two groups on the basis of qualification for national health insurance. Public assistance was applicable to people who were unable to support their family. In underlying diseases, chronic pulmonary disease was defined as a non-infectious chronic pulmonary disease such as chronic obstructive pulmonary disease and asthma. Chronic kidney disease was defined as a disease causing a progressive loss in renal function. Chronic cardiac disease was defined as a disease which needed a long term use of cardiac medications such as coronary artery disease, valvular heart disease, cardiomyopathy, and cardiac arrhythmias. Chronic liver disease was defined as a disease casing progressive destruction and regeneration of liver parenchyma such as chronic viral hepatitis, alcoholic liver disease, and liver cirrhosis.

Patients were classified into two groups according to treatment outcome (2). The case group comprised defaulters, i.e., TB patients who did not start treatment or whose treatment was interrupted for 2 consecutive months or more. Meanwhile, the control group comprised non-defaulters who had been cured or completed treatment. Patients with TB-related death or those who were transferred out were excluded from the control group.

To track defaulters, PPM nurses tried to contact all of the listed patients by telephone and ascertained individual clinical and socioeconomic data. For non-defaulters, the data were obtained by one or two in-person interviews by PPM nurses during follow-up at the outpatient department. Questionnaires were administered to collect information about personal details, disease perception (i.e., TB), and causes of default. The questions were selected from among several factors found to be significantly associated with default in previous studies (7891012). Disease stigma was defined as feeling ashamed to have TB, trying to hide from or disclose to others, or attempting to be isolated (89). Incorrect perception or information about the disease was defined as a patient’s lack of adequate knowledge and education about TB or their recent disease status, such as the efficacy of anti-TB treatment, possibility of feeling better or being cured, and treatment duration (78). A low level of satisfaction with the clinic or doctors reflected poor healthcare worker attitude towards patients, long waiting times at the clinic, or lack of opportunity to express feelings (810).

All data are expressed as median (interquartile range [IQR]) or numbers and percentages. Intergroup comparisons were performed using the Mann–Whitney U-test for continuous variables, and Pearson’s χ ² test or Fisher’s exact test for categorical variables. We evaluated trends using the χ ² test for trends. Multivariate analysis was performed to determine the predictors of default from TB treatment. Possible predictors found to be significant in univariate analysis were entered into binary logistic regression. The level of significance was set at P < 0.05.

We planned a study of independent cases and controls with two controls per case. Data of the entire population with TB at our clinic indicate the probability of default (i.e., default rate) was 0.049. If the true odds ratio for disease in exposed subjects relative to unexposed subjects is 5, we would need to study 49 case patients and 98 control patients to be able to reject the null hypothesis that this odds ratio equals 1 with a probability (i.e., power) of 0.8. We tried to contact all 102 defaulted patients using telephone, but we could connect 58 patients. Of these patients, 46 patients agreed to telephone interview. Ultimately, we recruited 46 cases and randomly selected 100 controls who were eligible for personal interview and regularly followed up. The type I error probability associated with this test of this null hypothesis is 0.05.

Permission to review and publish patient records was obtained from the institutional review board of Chonnam National University Hospital (IRB No. CNUH-2014-260). Informed consent was obtained from patients personally, and confidentiality and anonymity were assured. In cases of defaulters who would not or could not visit the hospital, assistants asked for consent to participate in this study by telephone.

Among 2,072 patients with TB treated at Chonnam National University Hospital from January 2010 to December 2014, 102 were defaulted from treatment, giving a default rate of 4.9%. Among them, 46 defaulters were assigned to the case group, and 100 patients were randomly selected as controls.

Regarding sociodemographic characteristics, a larger proportion of defaulters had low income (< 2,000 US dollars per month) than controls. Other factors such as body mass index, smoking, alcohol, employment, marriage, education, insurance, and hospital accessibility did not differ significantly between groups (Table 1).

The median (IQR) time from treatment initiation to default was 101 (35–185) days. Regarding clinical characteristics, patients treated by a non-pulmonologist and patients with EPTB had a significantly higher default rate. Patients with a history of treatment for TB or former default also had a significantly higher default rate (Table 2). Adverse drug reactions predominantly affected the skin (e.g., pruritus, rash, and acne) and gastrointestinal system (e.g., nausea, vomiting, and anorexia). However, there were no differences in adverse drug reactions between groups (Table 3).

The results regarding disease perception are shown in Table 3. Interestingly, most defaulters did not feel stigma about their disease, whereas more than half of non-defaulters did.

The most common cause of treatment default according to patients’ personal opinions was incorrect perception or information about their disease (19/46, 41.3%) and side effects of anti-TB treatment (19/46, 41.3%). Poor treatment tolerability such as work interference with adherence or unwillingness to taking medicine (12/46, 26.1%) and low level of satisfaction with the clinic or doctors (5/46, 10.9%) were less common reasons.

Significant risk factors for default in univariate analysis, including low income (< 2,000 US dollars/month), treatment by a non-pulmonologist, history of previous treatment, former default, combined EPTB, and absence of TB stigma, were included in multivariate analysis. In multivariate analysis, low income (adjusted odd ratio [aOR]: 5.246, 95% confidence interval [CI]: 1.249–22.029, P = 0.024), treatment by a non-pulmonologist (aOR: 14.567, 95% CI: 3.260–65.089, P < 0.001), former defaulters (aOR: 33.226, 95% CI: 2.658–415.309, P = 0.007), and absence of TB stigma (aOR: 46.299, 95% CI: 8.078–265.365, P < 0.001) were independent predictors of default from TB treatment.