A cross-sectional quantitative study was done with drug-susceptible TB (DS-TB) patients and retreatment-TB patients currently undergoing treatment at 163 private clinics in Kachin, Kayin, Yangon and Ayeyarwady region, of Myanmar, selected for their high TB caseloads and concentration of private healthcare providers in these areas [14]. This study was part of a larger assessment which was to conduct the private sector providers’ landscaping and their role in TB care in high burden areas. In the project, the selection of areas was based on existing tuberculosis project implementation in private sector and the selected areas had no or little TB private sector providers.

The private sector providers in Myanmar are essential partners in TB care, with the potential to enhance access to services, improve patient outcomes, and contribute to broader public health efforts against tuberculosis. While the care and management of drug-resistant TB (DR-TB) are mainly operated by the national TB control programme and public health sector, our study specifically focused on drug-susceptible TB (DS-TB) patients seeking care in the private sector.

Our study participants were drug-susceptible TB patients receiving treatment from private GPs, who had partnerships with non-governmental organizations. Specifically, the private GPs collaborated with TB PPM partners like Population Services International (PSI), as well as other TB implementing partners such as Medical Action Myanmar (MAM), the Asian Harm Reduction Network (AHRN), and the Myanmar Anti-TB Association (MATA) which played crucial roles in the TB healthcare landscape within the study regions.

The study utilized a standardized questionnaire from the World Health Organization (WHO) to collect data [15]. This questionnaire gathered information on cost components such as direct medical, direct non-medical, and indirect costs incurred by TB patients. It also aimed to capture details about the coping mechanisms employed by patients.

Participants eligible for the study were 18 years or older, diagnosed with drug-susceptible tuberculosis, and receiving care (Initial or retreatment) from public-private mix (PPM) clinics. Recruitment occurred between December 2021 and April 2022. Both patients and healthcare providers received information about the study and obtained initial consent from potential participants. The research team obtained a weekly list of patients, contacting them to arrange for telephone interviews.

As of January 1st, to April 1st 2022 report, there were 1860 eligible participants for the study from 163 private sector clinics within the four states and regions. 62.6% (1165 TB patients) were being excluded due to completed treatment, age under 18, serious illness, death or loss to follow-up, and incomplete data. Various contact issues such as phone power-off, no answer, wrong number, refusals, and completed TB treatment were the primary reasons for exclusion.

Study samples were selected from TB patient lists provided by five TB implementing partners receiving anti-TB treatment at private clinics from January 1st to April 1st, 2022. The sample size (n = 700) was estimated considering a 15% nonresponse rate and private sector's presumed contribution of 18.9% to total TB cases, as reported in the 2018 National TB program report. A systematic random sampling approach with a design effect of 1.5 was applied by using the following formula [16].

$$ } = \frac \times p\left( \right)}} }}}} \times p\left( \right)}} N}}} \right)}} $$

where e = margin of error (0.03), and z = z-score (1.96), p = proportion of total TB cases contribution by private sector (0.188), and N = population size (136,039) based on the 2018 NTP report [17].

The number of samples required from each state and each IP was determined based on probability proportional to the size (PPS) of the total patient list collected in a defined period. Among the listed patients, 1860 selected patients were approached for telephone interviews. In the end, 695 TB patients (37.4%) were successfully interviewed, while the remaining 1165 TB patients (62.6%) were unable to be reached due to registered phone numbers being logged off, not answered, wrong number, refused, patients passing away, treatment being completed, and interrupted interview. The participation rate was 99.3% anticipated by the survey design (695/700).

Household socio-economic status information such as household assets, housing materials, drinking water source, etc. was collected through a short version of equity tool [18].

Cost analysis, from the patients' perspective, covered various financial aspects. Direct medical costs included payments for consultations, tests, medicines, and medical procedures, while direct non-medical costs included expenses like transportation and accommodation. Costs were calculated separately for different stages of treatment: pre-TB treatment, and post-TB treatment (intensive treatment, and continuation treatment) [15].

Hospitalization costs included all expenditures related to a patient's hospital stay for patients and accompanying family or friends [19]. Indirect costs included productivity loss (the loss of personal income due to TB illness) [19], and coping costs (selling assets and borrowing money).

Indirect costs were assessed using a human capital approach. This method was chosen due to the unreliable household income data reported by patients and the high proportion of patients engaged in informal employment compared to other sectors in Myanmar. Indirect costs were determined by multiplying the reported hours spent seeking and receiving care during the TB episode by the individual’s hourly income. This approach accounted for time lost traveling to health facilities and waiting during healthcare consultations for both patients and their caregivers. The total self-reported time spent on these activities was multiplied by the estimated hourly income per person. Hourly income was estimated from the self-reported income data collected from all survey participants, calculated based on their reported individual income and hours worked.

The food-share-based poverty line was set using the proportion of household total expenditure that a household allocated to food [24]. A household was classified as poor if its total expenditure was less than its calculated subsistence spending. Conversely, a household was considered non-poor if its total expenditure was equal to or greater than its subsistence spending. The poverty line (PL) in this calculation was defined as the subsistence expenditure per equivalent capita, which was calculated by (1) identifying the food expenditure share of households, specifically those in the 45th to 55th percentile range of food expenditure as a share of total household expenditure; (2) calculating the weighted average of food expenditures within this range; and (3) determining the minimum standard of income necessary to meet the basic needs of a household of equivalent size.

Catastrophic heath expenditure occurred when a household’s total out-of-pocket health payments equal or exceed 20% of household’s capacity to pay or non-subsistence spending [20, 21]. This threshold was used to identify households facing financial burden due to TB-related expenses.

Data collection took place in May and June 2022 using a WHO-adapted questionnaire translated into Myanmar [15, 21] and was created in CSPro® (Census Bureau, USA). The telephone survey covered TB patient information on demographic details, economic status, current TB treatment, direct medical and non-medical payments, indirect costs (income loss or time), caregiver costs, coping mechanisms, household assets and income during post-TB treatment phase (e.g., either the intensive phase or the continuation phase). The questionnaire was pretested with TB patients and the PSI/Myanmar staff to test fluency of the questions and then they were modified accordingly. All the duration for completing one interview lasted 45–60 min.

Interviews were done via telephone calls. The study collected data and checked for completeness and consistency using the Stata version 14.2 (StataCorp, College Station, TX, USA). Access to the secure server was limited to field management staff. Moreover, these uploaded electronic data were stored in password-protected devices. The PSI Myanmar research team had access to this information for analysis and research purposes.

For the recruitment, PSI Myanmar initially informed private health care providers and TB implementing partners via phone or email, usually 1–2 weeks in advance, seeking their approval to recruit their clients. The TB patient information was collected by using the client listing form, including mostly data elements from the clinic register or provider records, such as patient ID or name, patient phone number, age, type of TB, residential township, and current treatment status.

Before participants were recruited for the study, all participants were asked for their verbal consent for the interview. The interview session happened after receiving their consent. The contact information for those not consenting was deleted immediately.

The data were exported to Stata version 14.2 (StataCorp, College Station, TX, USA) for cleaning and analysis. The socio-economic status of households, ranging from the poorest to the richest, was determined using a set of questions about household assets [18]. This tool measured the relative wealth of a household by converting its assets into a composite score and applying a cut-off to establish five wealth quintiles [22]. The cut-offs applied were those from national quintiles, therefore, the wealth of studied households represented their relative wealth with reference to national wealth quintiles [23].

All cost data were collected in Myanmar Kyats and converted to USD (1 USD = 1850 Myanmar Kyats as of July 20, 2022). Descriptive statistics, including median, mean, inter quartile range (IQR), and 95% confidence interval (CI) for continuous data, as well as frequencies for categorical data, were used. Continuous variables with normal distribution were presents as mean ± standard deviation (SD); non-normal variables were reported as median (IQR). Costs were analyzed from the patient's perspective, including direct and indirect costs. For each treatment stage, costs per visit were calculated, and total costs were estimated using median costs with ranges. Median values were used due to data skewness.

The patient costs were separately calculated for the stages of treatment pre-TB treatment, and post-TB treatment (intensive treatment, and continuation treatment). For patients with two to six months of treatment, total costs for the continuation phase were computed by summing actual visit costs and estimated costs calculated by multiplying their average treatment cost per visit based on remaining visits. The latter estimate was based on their remaining treatment duration and clinic visit frequency to get the best estimate. For patients in the intensive phase (< 2 months), average continuation phase costs (estimated cost between two and six months) were added to their initial two months for the entire six-month duration.

Statistical significance was defined as P < 0.05. The out-of-pocket payments were calculated as sum of direct and indirect TB treatment expenditures, including productivity and coping costs due to TB. Catastrophic health expenditure (CHE) was defined as total health expenditure equaling or exceeding 20% of the household's capacity to pay [3, 15, 20, 21, 24]. Moreover, the variable threshold levels were considered: 40%, 30% and 10%. Additionally, we explored different outcome measures, such as the human capital and output approaches, to provide a more comprehensive analysis.

Bivariate and multivariable logistic regression analysis were used to identify variables significantly associated with CHE. The variables included patients' age, gender, regions, total monthly household income, national SES quintile, treatment category, history of hospitalization, history of coping strategy use, health-seeking channel, and comorbidities. P-values less than 0.05 and adjusted odd ratios (aOR) with 95% CI were reported.