healthmedicinet

Data source

Since 2009, the Gates Foundation, in collaboration with the Chinese Ministry of Health/China
CDC, has been implementing an innovative program on TB/multidrug-resistant TB (MDR-TB)
control and prevention in four Chinese cities 25]. In 2013, the second phase of this program was initiated. The program aims to use
innovative tools and delivery approaches to establish comprehensive TB/MDR-TB control
models which can be scaled up over time by the National TB Prevention and Control
Program. The data used for this study were derived from the baseline studies conducted
for this second phase.

Study sites

The baseline studies were conducted in three cities (Zhenjiang City, Jiangsu Province;
Yichang City, Hubei Province; and Hanzhong City, Shaanxi Province), which are geographically
located in the eastern, central and western regions of China. Three counties or districts
(one from each category of high, middle and low GDP per capita) were then selected
as study sites in each city (Dantu, Yangzhong and Jurong in Zhenjiang; Zhijiang, Yidu
and Wufeng in Yichang; and Chenggu, Mian and Zhenba in Hanzhong). Figure 1 shows the location of these sites. TB dispensaries and designated hospitals were
the institutions authorized to provide TB diagnosis, treatment and case management.
All newly diagnosed TB cases were required to be registered in the local dispensary
or designated hospital and reported to upper level health authorities.

Fig. 1. Location of nine study sites in China

Study participants

A cluster sampling method based on townships/streets was used to recruit TB patients.
Using selected key indicators (financial burden of TB care and treatment adherence),
the minimum required sample size per city was calculated to be 264 cases. Three townships/streets
were selected in each county/district using probability proportional to size (PPS)
sampling and 30 TB cases were randomly sampled in each, using the list of registered
cases and adopting the criteria that each patient had completed or stopped treatment
during 2012. The sample included both new patients and those who had been previously
treated for TB but where treatment had failed or the patient had defaulted or relapsed.

A total of 797 TB patients were interviewed, of whom 50 failed to report out-of-pocket
expenditures or household income. Thus, 747 cases with complete data were included
in the analysis.

Data collection

The baseline survey was conducted between April and May 2013. All of the participants
were interviewed after treatment completion/interruption. The TB patients were interviewed
face-to-face at local TB dispensaries or designated hospitals using a standardized
survey questionnaire which included personal demographic and socio-economic information
(age, sex, education, household income/expenditure, etc.), direct health service expenditures,
and non-medical expenses (transportation and accommodation costs, etc.). The household
income and expenditure (e.g. food expenditure) were ascertained through direct questions.

The patient survey was conducted by university/college students from Huazhong University
of Science and Technology (Yichang), Xi’an Jiaotong University (Hanzhong) and Nanjing
Medical University (Zhejiang). A technical assistance team (TA) from the Duke Global
Health Institute, USA, the Institute of Development Studies, UK and Shandong University,
China also took part in the baseline survey. The interviewers received training on
interview skills and the contents of the questionnaire before the survey.

Measuring CHE and definitions

CHE is usually assessed by incidence (headcount) and intensity indicators, which were
described in detail in Wagstaff et al. 26], 27]. Measuring CHE requires specification of the thresholds for household income or capacity
to pay (non-food expenditure), which household health expenditure should not exceed.
There is no consensus on these thresholds, but the two most commonly used are: 10 %
of total household income; and 40 % of household capacity to pay 28]. In this study, CHE for TB care was estimated using both criteria.

Out-of-pocket (OOP) payments for TB care include direct health expenditures on diagnosis
and treatment (consultation fees, laboratory tests, X-rays, drugs, and hospital care)
and associated non-medical expenses (transport and accommodation costs for patient
and companions, nutrition supplement costs), but not income loss. The headcount indicator
(H) is the percentage of households whose OOP payments (numerator) as a proportion
of household income or non-food expenditure (denominator) exceeds a given threshold.
The mean gap indicator (G) is the average amount by which payments, as a proportion
of household income (or non-food expenditure), exceed this threshold 26], 27]. Finally, the mean positive gap (MPG) is equal to G/H, the excess expenditure per
household experiencing CHE.

Data management and analysis

In order to ensure quality, the double entry data verification approach was adopted
using EPI Data 6.04. The statistical package SPSS 13.0 was used to analyze the data.
Household income, household expenditure, OOP and capacity to pay were presented as
means (Standard deviation, SD) and medians (percentile 25 and percentile 75, p25 and
p75) to allow for the expected positively skewed distributions. Sensitivity analysis
of the incidence and intensity of CHE for TB care was applied using different thresholds
across different household income groups. Chi-square tests were used to identify factors
associated with CHE. Multivariate logistic regression analysis was employed to assess
the effects of explanatory variables in a multivariate model. Two multivariate models
were run respectively using both of the above-mentioned definitions. Statistical significance
was set at 5 %.

Quality assurance

The questionnaire and survey procedures were tested in a pilot study carried out in
Zhenjiang city in March 2013 and then refined before use in the three project cities.
Key researchers from the above four Chinese universities acted as survey supervisors
to check the consistency and quality of the data collected.

Ethical consideration

The Ethical Committee of China CDC reviewed and approved the study protocols and instruments.
Informed consents were obtained from all study participants.