Is the public-private mix approach increasing tuberculosis case notification in Azad Jammu and Kashmir , Pakistan ? A cross-sectional study

Pakistan ranks among high tuberculosis (TB) burden countries, Background: with about 200,000 missing TB cases. Despite significant increase in case detection and treatment outcomes through public-private mix (PPM) globally, the contribution of the private sector may vary in different parts of Pakistan This was a cross sectional study conducted in state of Azad Jammu Methods: and Kashmir (AJK), Pakistan. The study was a retrospective record review of routine TB notification and treatment outcomes for 2015 to 2016 in districts with and without a PPM approach. The study was conducted in all districts of AJK, including all public and private health facilities. Intra-district comparison in PPM supported districts was also carried out. Total notified TB cases during 2015-16 were 11479. Districts with Results: PPM support has notified 5882 (51.24%) of TB cases while districts without PPM support contributed 5597 (48.87%). Pulmonary clinical diagnosis was higher in PPM supported districts (45.43%) as compared to districts without PPM support (31.11%). Notification of extra pulmonary TB cases was lower in PPM supported districts (1256; 21.35%) as compared to districts without PPM support (1877; 33.54%). Unfavourable treatment outcomes (treatment failure, died, lost to follow-up) was higher in PPM facilities (5.84%). The study shows minimal increase in TB case finding through the Conclusion: PPM approach. While this is an important aspect in END TB strategy, this needs more careful evaluation.

This article is included in the gateway. TDR

Introduction
Tuberculosis (TB) remains a major public health problem, as it is the leading cause of death from a single infectious agent. Globally, about 10 million people developed TB disease with 1.3 million TB deaths annually 1 . TB case notification is challenging as there was a 3.4 million gap globally between notified and incident cases 2 . To ensure universal access to quality TB services is a major challenge due to lack of systematic engagement of all health care providers, especially in the private sector 1 .
The public sector plays a major role in control of TB in Pakistan. More than 70% of total case notification was from public sector in 2017 1 . Similarly, a study from India showed that the public sector has contributed 84% in TB case notification 3 . In addition, the role of the public sector in management of TB was assessed in Thailand and it was concluded that 90% of TB cases were diagnosed and treated in public health facilities, while the private sector contributed only 10% in case notification and treatment 4 .
The public-private mix (PPM) approach has been suggested by the World Health Organization to engage all health care providers 5 . PPM approach was tested in 23 countries from 1999 to 2000 and there was an increase in case detection rate and treatment outcome remained above 85% 6 . Evidence suggests that PPM has the capacity to achieve increased case notification, increased treatment outcome and more importantly it also improves patient satisfaction 7 .
Several studies have been conducted in the region for estimation of cost effectiveness and the impact of PPM on case detection and TB management. These studies suggested that PPM is a promising model with increased case notification and improved treatment outcome 8-10 . However, PPM intervention can vary in terms of accessibility, cost effectiveness, acceptability and quality 11 . About 80% of TB patients in Pakistan attend private sector facilities for their initial diagnosis and management and most of them are not reported to the National TB Control Program 7 . Pakistan lists among countries with the highest number of unnotified TB cases 12 . In 2006, Pakistan adopted PPM model which has been gradually scaled-up 13 . The contribution of PPM in all case notification in Pakistan has been reported to be up to 18% 14 . Studies from Pakistan have revealed that PPM can play a substantial role in achieving the targets of treatment success rate and case detection 7,15,16 . Majority of TB cases which are managed by private sector, are not notified to National TB Control Program. Although the referral of presumptive TB cases is higher (70.9%), but only 29.1% are treated 17 .
All studies conducted in Pakistan focused the role of private sector but there is limited data available for inter-district and intra-district comparison. There is a need to compare the PPM districts with public sector districts, which are in close vicinity and with minimal geographical differences. Therefore, this study provided an inter-district as well as intra-district comparison in intervention districts in terms of case notification and treatment outcomes. This study also provided the overall contribution of public and private sector in terms of TB case notification and treatment outcomes.

Study design
This was a cross-sectional study based on the retrospective review of routine data of State TB control program from 2015 to 2016.

Study setting
Azad Jammu and Kashmir (AJK) is a self-governing state in Pakistan with a population of 4.045million. Administratively it is divided into three divisions and ten districts. The total area of AJK is 5134 square miles (13,297 square km). The topography of the area is mainly hilly with high mountains, valleys and stretches of plains. The rural urban ratio is 88:12 18 . State TB Control Program is actively participating in the control of TB. PPM approach was adopted in 2011 in three districts and then scaled up to four districts in 2015. PPM approach in AJK is carried out by Sub-Recipient with Global Fund's Principle Recipient.
PPM approach in four districts (Bagh, Kotli, Mirpur and Bhimber) comprises almost half of the population of AJK ( Figure 1). In these four districts, the General Practitioner (GP) model has been adopted and currently 73 GPs and 14 diagnostic laboratories work actively in these districts.

Data variables
TB case notification including pulmonary and extra pulmonary cases was assessed. TB treatment outcomes (cured, treatment completed, lost to follow-up, treatment failure, died and not evaluated) was compared between districts.

Data sources and collection
Data was collected from district reports. TB case notification report and TB treatment outcomes report was used for retrospective data analysis. Data collected at district level was verified and validated with TB registers available at each TB facility in all districts.

Data analysis
Data was collected from district reports and then entered into Epi-Data software (version 3.1 EpiData Association, Odense, Denmark). Data was analysed and percentages were obtained. Case notification of districts with and without PPM support was calculated. The proportion of both groups was assessed in terms of case notification and treatment outcomes.

Ethical consideration
Written approval (849/DHS-CDC/18) was obtained from the TB program manager of AJK-Pakistan. As secondary data was used, there was no direct involvement of patients/human subjects in this study.  The analysis of TB treatment outcomes are shown in Table 2. This shows that patients lost to follow-up was less in PPM supported districts with a percent change of -2.91%. Total number of TB cases under category of "not evaluated" was lower in PPM supported districts with percent change of -1.75%.

Total
Intra-district comparison of PPM supported districts is shown in Table 3. The PPM TB health facilities contributed 15% of TB case notification with a higher proportion of clinical diagnosis and lower proportion of extra pulmonary TB cases. Public health TB facilities contributed 24% of extra pulmonary cases, while PPM TB facilities contributed only 6.41%. Total TB case notification by PPM supported districts was higher in Mirpur, followed by Bhimber, Bagh and Kotli. TB treatment outcomes among PPM supported districts were analysed. Unfavourable treatment outcomes (treatment failure, died, lost to follow-up) was higher in PPM facilities (5.84%).

Discussion
It is evident from the study that the public sector remained the major contributor in TB case notification; the public sector notified 85% TB cases while the private sector contributed 15% only. In our study, inter-district comparison showed that case notification in districts with and without PPM support was almost the same with minimal difference. Pakistan has prioritized the PPM approach and there was a 30% increase in case detection rate in 2017 1 . Data from Pakistan suggested that TB  case notification has been increased due to involvement of private health care providers 14 . Similar studies from Pakistan also suggested that TB case notification has been improved due to involvement of private health care providers 6,7,19 . All these studies were conducted in PPM supported districts, but there is limited data available to compare PPM supported districts with non-PPM districts (public sector only).
In this study, pulmonary TB clinical diagnosis was higher in PPM support districts with a percent change of 14.32%. This showed that the private sector is relying more on X-ray based diagnosis or clinical diagnosis without referring the presumptive TB cases for laboratory investigation. A prevalence survey in Pakistan also showed that the majority of pulmonary TB cases are diagnosed by radiography in the private sector 20 . A study conducted in Karachi-Pakistan also concluded that the numbers of TB cases from the private sector are clinically diagnosed and there is a need for strengthening reliance on TB laboratories for screening of presumptive TB cases 21 . Therefore, the private sector is relying more on X-ray based diagnosis rather lab based diagnosis 22 . Laboratory based diagnosis in the private sector is limited and more than half of all presumptive TB cases in the private sector is referred to the public sector for diagnosis 23 .
In this study extra-pulmonary TB case notification in districts with PPM support is lower with a percent change -12.82%. Extra pulmonary TB case notification is usually difficult and needs extra investigation with high quality tests. GP model needs more TB treatment outcomes are one of the major indicators for assessing a successful TB control program. A study showed that AJK had reported a 95% treatment success rate. The proportion of unfavourable TB treatment outcomes, died, TB treatment failure and lost to follow-up, were 3.8%, 0.1% and 0.2% respectively 14 . Our study also showed similar results with minor difference in public and private TB health facilities. Overall TB treatment success rate was above 95% in public TB health facilities. The proportion of unsuccessful TB treatment outcomes was higher in PPM facilities (5.84%).
Our study showed that TB case notification through PPM was only 15% in PPM supported districts, but there was no such significant difference when PPM supported districts were compared with non-PPM districts (public sector only). A comparative performance of public and private health sector in low and middle income countries also support our study that the public sector is more efficient, accountable and medically effective than the private sector 24 . In Pakistan there is need of more stringent selection criteria for GP selection to improve their involvement in TB control 7 . PPM is an important approach to achieve global TB targets; however it could be affected by contextual characteristics in different areas 25 .
The strength of the study is that it compares PPM supported districts with other districts (non-PPM supported). Previous studies in Pakistan focus on few districts while this study covers a whole region of AJK, and all districts and health facilities were included in this study.
The limitation of our study is the reliance on program data. Accuracy and completeness of the data cannot be assured. There could be many reasons for deaths, and it is important to know that deaths occur due to TB rather that to another reason.

Conclusions
The study showed minimal increase in TB case finding using the PPM approach. While this is an important aspect in END TB strategy, this needs more careful evaluation. The public sector is contributing effectively to TB case notification and has better TB treatment outcomes. Therefore, there is a need to strengthen the public sector.

Recommendations
Although the PPM approach is promising and is included in the END TB strategy, our study showed that the public sector is more efficient than PPM in terms of case notification and treatment outcomes. Public sector of AJK is comparable with PPM supported districts, with more good indicators. It could be further improved by monitoring and evaluation. PPM approach needs stringent selection criteria so that it could perform more efficiently. Contextual characteristics need to be addressed while implementing PPM. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.