**GeneXpert and Community Health Workers Supported Patient Tracing for Tuberculosis Diagnosis in Conflict-A**ff**ected Border Areas in India**

**Mrinalini Das 1,\* , Dileep Pasupuleti <sup>2</sup> , Srinivasa Rao <sup>3</sup> , Stacy Sloan <sup>2</sup> , Homa Mansoor <sup>1</sup> , Stobdan Kalon <sup>1</sup> , Farah Naz Hossain <sup>1</sup> , Gabriella Ferlazzo <sup>4</sup> and Petros Isaakidis <sup>4</sup>**


Received: 19 November 2019; Accepted: 11 December 2019; Published: 21 December 2019 -

**Abstract:** Médecins Sans Frontières (MSF) has been providing diagnosis and treatment for patients with tuberculosis (TB) via mobile clinics in conflict-affected border areas of Chhattisgarh, India since 2009. The study objectives were to determine the proportion of patients diagnosed with TB and those who were lost-to-follow-up (LTFU) prior to treatment initiation among patients with presumptive TB between April 2015 and August 2018. The study also compared bacteriological confirmation and pretreatment LTFU during two time periods: a) April 2015–August 2016 and b) April 2017–August 2018 (before and after the introduction of GeneXpert as a first diagnostic test). Community health workers (CHW) supported patient tracing. This study was a retrospective analysis of routine program data. Among 1042 patients with presumptive TB, 376 (36%) were diagnosed with TB. Of presumptive TB patients, the pretreatment LTFU was 7%. Upon comparing the two time-periods, bacteriological confirmation increased from 20% to 33%, while pretreatment LTFU decreased from 11% to 4%. TB diagnosis with GeneXpert as the first diagnostic test and CHW-supported patient tracing in a mobile-clinic model of care shows feasibility for replication in similar conflict-affected, hard to reach areas.

**Keywords:** sputum; health promotion; operational research; indigenous population

#### **1. Introduction**

The management of tuberculosis (TB) is challenging for patients residing in remote and inaccessible areas. The scale of the challenge escalates when inaccessibility to healthcare increases due to conflict. Patients in these hard-to-reach areas need special attention from TB programmes and implementing partners [1].

India is a high-burden TB country, contributing to approximately a quarter of global incident TB cases. In 2018, the estimated number of TB cases in the country was 2,790,000 [2]. The border areas of central India (including four states, i.e., Chhattisgarh, Odisha, Telangana, Andhra Pradesh) have been affected by a long-standing, low-intensity, chronic conflict [3]. The majority of the population residing in these areas belong to various tribes and have limited access to healthcare services, including access to TB diagnosis and treatment facilities [4].

The Revised National TB Control Programme (RNTCP) has been providing TB care to remote and tribal populations [5]; however, these services in conflict-affected areas are often interrupted due to frequent instances of minor clashes. Basic healthcare services are provided at selected primary healthcare centers, but patients need to travel more than 50–100 km to access tertiary care services in district hospitals.

Médecins Sans Frontières (MSF), a nongovernmental, medical humanitarian organization, has been providing primary healthcare services, including diagnosis and treatment for patients with TB, via mobile clinics in the chronic conflict-affected border areas of Chhattisgarh, India since 2009 [6]. A unique model of care in collaboration with RNTCP has been implemented, aiming at offering improved delivery of TB diagnosis and treatment services. GeneXpert in the nearby government hospital (Bhadrachalam district hospital) has been utilized as the first diagnostic test for TB diagnosis since January 2017. Community Health Workers (CHW) are trained in patient tracing (that is, in the follow up of patients) in order to minimize pretreatment loss-to-follow-up (LTFU).

To date, there has been no documentation of this TB model of care in India. The aim of this study is to contribute to the body of evidence related to TB diagnostic delivery in conflict-affected and tribal areas, and to help policy makers and implementers to develop tailor-made, diagnostic strategies for such conflict-affected, hard-to-reach populations.

The specific objectives of the study included determining the number and proportions of (1) patients diagnosed with TB, (2) pretreatment lost-to-follow-up patients (from first presentation for diagnosis up to the date of receipt of TB diagnosis results, and (3) to compare the proportion of bacteriological confirmation and pretreatment LTFU between two time periods: (a) April 2015–August 2016 and (b) April 2017 August 2018 (before and during the utilization of GeneXpert as a first diagnostic test for TB diagnosis).

#### **2. Materials and Methods**

#### *2.1. Study Design*

This was a retrospective analysis of routinely collected clinical and programmatic data.

#### *2.2. Setting*

The state of Chhattisgarh in central India has a population of 26 million [7], including conflict-affected zones in the Sukma, Dantewada, and Bastar districs. Accurate information on the populations residing in the conflict-affected zones is not available [8]. The total number of notified TB cases in the Sukma district (where MSF TB Programme is operational) was 335 in 2018 [2].

#### TB Model of Care Description

MSF has been providing routine primary healthcare services, including TB care in the conflictaffected border areas with an estimated population of 90,000 since 2009 [6,9,10]. TB diagnosis and treatment is offered by a multidisciplinary team including doctors, nurses, counselors, health promoters, and CHWs. A doctor and nurse are TB focal points for the TB program. The nurses provide support with sample collections for TB diagnosis. The counselors provide information to patients and family members about TB signs/symptoms, treatment regimen, routes of TB transmission, and infection control. Health promoters carry out community sensitization sessions in villages every month on TB, malaria, diarrheal diseases, general hygiene, and sanitation.

A group of local CHWs are trained the identify the symptoms/signs of TB and support the tracing of patients. In case patients miss appointments for two weeks or more, a CHW visits the patient at their residence in the villages. Repeated sensitization of CHWs (once every three months) is carried out by the TB focal points and health promoters. The CHWs are paid a fixed stipend every month.

Since 2017, GeneXpert in the nearest district hospital in Bhadrachalam, Telangana has been utilized as the first diagnostic test for TB diagnosis in patients with presumptive TB. In India, studies have shown that GeneXpert has a sensitivity and specificity of 100% each for pulmonary TB samples and a sensitivity and specificity of 90.7% and 99.6% respectively for extra-pulmonary TB samples, in comparison with composite reference standards [11]. The patients with presumptive TB are requested to provide a spot sample in order to avoid the need to travel 10–15 kms to visit a clinic. Patients are given sputum containers for morning samples to be submitted on the next mobile clinic day (3 days later). The GeneXpert results (using spot samples) are given to patients on the next mobile van visit (the same visit when the morning sample is submitted); in cases whereby the GeneXpert result on the spot sample is negative, microscopy is performed on the morning sample. Patients are referred to the district hospital for a biopsy or chest X-ray, as required. For those with negative TB diagnostic results, a clinical decision is taken by the medical team. As laboratory results become available and are reported to the patient, pretreatment counseling is provided and treatment is initiated.

#### *2.3. Study Site and Population*

All patients with presumptive TB who received care in the MSF TB Programme in border areas of Chhattisgarh between 01 April 2015 and 31 August 2018 were included.

#### *2.4. Data Variables and Sources, Data Analysis*

The demographic (age, sex) and clinical characteristics (presence of cough, history of previous TB) of patients with presumptive TB, date of presentation, date of sputum collection, and diagnostic results (sputum, GeneXpert, Biopsy, Xray) were extracted from an electronic database and imported into STATA (version 11, StataCorp, College Station, Texas, USA) for analysis. TB diagnosis and pretreatment loss-to-follow up were summarized using frequency and proportions. Continuous variables such as age were summarized using median and inter-quartile range (IQR). Categorical variables (sex, previous history of TB, site of TB) were summarized as frequency and proportions. Associations between demographic and clinical characteristics and diagnosis of TB were assessed using a chi-square test and unadjusted relative risks (RR) with 95% Confidence Intervals (95% CI). A p value of less than 0.05 was considered statistically significant.

#### *2.5. Operational Definitions*

1. Presumptive TB: Presumptive TB refers to a patient who presented with symptoms or signs suggestive of TB [12]

2. Bacteriological confirmation: Presence of MTB+ in GeneXpert results; smear microscopy or culture evaluation was considered as a means of bacteriological confirmation

3. Clinically-diagnosed TB case: Patient diagnosed with active TB by a clinician on the basis of X-ray abnormalities and/or clinical evaluation. This includes smear-negative pulmonary TB and extra-pulmonary TB cases without laboratory confirmation.

4. Confirmed TB case: Patients with bacteriological confirmation or clinically-diagnosed TB

5. Error: Failure to test for diagnosis of TB was termed as error, which included poor-quality of sputum, technical error of equipment, machine malfunction, etc.

6. Prediagnosis loss-to-follow-up: If the patients with presumptive TB, after the first consultation, did not visit the clinic to provide a sample for TB diagnosis, it was considered a prediagnosis loss-to-follow-up.

7. Diagnosed TB loss-to-follow-up: If the patients were diagnosed with TB but did not visit the clinic for receipt of results within 1 month of THE initial consultation date, it was termed as diagnosed TB loss-to-follow-up.

8. Pretreatment loss-to-follow-up: The prediagnosis loss-to-follow-up and diagnosed TB loss-to-follow-up patients were together termed as pretreatment loss-to-follow-up.

#### *2.6. Ethics*

This research fulfilled the exemption criteria set by the Médecins Sans Frontières Ethics Review Board for a posteriori analyses of routinely-collected clinical data. and thus, did not require MSF ERB review. It was conducted with permission from Medical Director, Operational Centre Brussels, Médecins Sans Frontières. Since it is a record-based study, we obtained a waiver from obtaining informed consent. Permission for conducting the study was sought from the National TB Programme of India (RNTCP).

#### **3. Results**

Among 1042 patients with presumptive TB identified in the program during April 2015 to August 2018, 376 (36%) were diagnosed with TB. The demographic and clinical characteristics of the patients with presumptive TB and those diagnosed with TB are shown in Table 1. The proportion of patients diagnosed with TB was largest (44.8%) in children aged 0–14 years compared to other age groups; it was similar in males (37.7%, 216/573) and females (34.1%, 160/469). Diagnosis of Pulmonary TB (PTB) was much more common than extra-pulmonary TB (82.9% versus 17.1%); however, a larger proportion of extra-pulmonary presumptive TB patients had confirmation of TB diagnosis than pulmonary TB patients (60.0% versus 32.4%). The younger age group [(0–14 years RR (95% CI): 1.5 (1.2–1.9); 15–24 years: 1.4 (1.1–1.8)] and extra-pulmonary TB [1.9 (1.6–2.1)] had a higher risk of developing TB.



\* Column percentage, \*\* Row percentages, Unadjusted RR: Unadjusted Relative Risk; CI: Confidence Intervals.

A total of 60 (5.8%) out of 1042 patients with presumptive TB were prediagnosis LTFU, while one died before providing sample for TB diagnosis. Of those who were diagnosed with TB (with bacteriological or clinical confirmation, n = 376), nine (2.4%) did not come back to receive the test results (termed as 'diagnosed TB LTFU') and did not initiate treatment during the study period. Thus, 69 (6.6%) of 1042 patients were pretreatment LTFU. Of those for whom a diagnosis of TB was made, 217 (57.7%), 49 (13%), and 110 (29.2%) had sputum-positive pulmonary TB, sputum-negative pulmonary TB, and extra-pulmonary TB, respectively (nontabulated).

Upon comparing the "before–after" time periods (Figure 1), bacteriological confirmation increased from 20% (67/342) to 33% (109/335). Errors in TB diagnoses decreased from 9% (34/376) to 0.1% (1/336). The pretreatment LTFU decreased from 11% (45/417) to 4% (13/346) during the study period.

SPPTB: Smear-positive pulmonary tuberculosis; SNPTB: Smear negative pulmonary tuberculosis; EPTB: Extra-pulmonary tuberculosis

**Figure 1.** Bacteriological confirmation and pretreatment loss-to-follow-up during two time periods: 1) Apr. 2015–Aug. 2016 (Before GeneXpert was used as first diagnostic tool for TB-diagnosis), 2) Apr. 2017–Aug. 2018 (GeneXpert used for TB-diagnosis) in conflict-affected border areas in India. **Figure 1.** Bacteriological confirmation and pretreatment loss-to-follow-up during two time periods: (1) Apr. 2015–Aug. 2016 (Before GeneXpert was used as first diagnostic tool for TB-diagnosis), (2) Apr. 2017–Aug. 2018 (GeneXpert used for TB-diagnosis) in conflict-affected border areas in India.

#### **4. Discussion 4. Discussion**

A model for TB diagnosis with the use of GeneXpert as a first test and CHW-supported patient tracing resulted in a 66% reduction of pretreatment loss-to-follow-up during 2015–2018 in a conflict-affected tribal area in India. A model for TB diagnosis with the use of GeneXpert as a first test and CHW-supported patient tracing resulted in a 66% reduction of pretreatment loss-to-follow-up during 2015–2018 in a conflict-affected tribal area in India.

More than one-third of patients identified with presumptive TB were diagnosed with active TB; this is higher than other studies reported in similar tribal areas (6.5% in the Bharia tribe of Madhya Pradesh, 12.5% in Maharashtra, and 21% in the Sahariya tribe of central India) [13–15]. This could be due to the availability of a trained medical team for supporting TB diagnosis. Further, the availability of GeneXpert as a first diagnostic test for presumptive TB patients in government district hospitals likely contributed to the increased number of detected cases [16,17], some of which may have been missed earlier. Investments for the continuous operation of GeneXpert must be considered by TB programs. The availability and accessibility of diagnostic tools like GeneXpert [17] and the implementation by a trained team [18] help in early and appropriate TB diagnoses in these hard-to-reach areas. More than one-third of patients identified with presumptive TB were diagnosed with active TB; this is higher than other studies reported in similar tribal areas (6.5% in the Bharia tribe of Madhya Pradesh, 12.5% in Maharashtra, and 21% in the Sahariya tribe of central India) [13–15]. This could be due to the availability of a trained medical team for supporting TB diagnosis. Further, the availability of GeneXpert as a first diagnostic test for presumptive TB patients in government district hospitals likely contributed to the increased number of detected cases [16,17], some of which may have been missed earlier. Investments for the continuous operation of GeneXpert must be considered by TB programs. The availability and accessibility of diagnostic tools like GeneXpert [17] and the implementation by a trained team [18] help in early and appropriate TB diagnoses in these hard-to-reach areas.

Studies in conflict areas across the globe have reported multiple challenges of access to healthcare services [19–21]. Few TB programs have been successful in delivering treatment in conflict areas by adapting to local needs [22]. Other than direct medical care under the National TB Programme, intersectoral measures such as access to a public distribution system, nutritional support, social welfare schemes, and security measures at the central and state levels will help to Studies in conflict areas across the globe have reported multiple challenges of access to healthcare services [19–21]. Few TB programs have been successful in delivering treatment in conflict areas by adapting to local needs [22]. Other than direct medical care under the National TB Programme, intersectoral measures such as access to a public distribution system, nutritional support, social welfare schemes, and security measures at the central and state levels will help to minimize the TB burden.

minimize the TB burden. The proportion of children and young adults was low in this group of patients with presumptive TB; however, the proportion of diagnosed TB was high compared to other age groups. As children are considered proxy indicators of TB transmission [23], it may be noted that the burden of TB is high in this tribal population, as it is in other tribal populations in the country [15]. The The proportion of children and young adults was low in this group of patients with presumptive TB; however, the proportion of diagnosed TB was high compared to other age groups. As children are considered proxy indicators of TB transmission [23], it may be noted that the burden of TB is high in this tribal population, as it is in other tribal populations in the country [15]. The national TB program in country, with the involvement of other NGOs and stakeholders, must devise tailored approaches for the provision of improved diagnoses and treatment in children [24,25].

The proportion of diagnosed EPTB cases among presumptive EPTB cases was higher than expected. This may hint at the late arrival of EPTB cases for diagnosis [26] and lower awareness about extrapulmonary signs of TB among the community [27]. EPTB is often considered low priority by TB programs, as it does not lead to the transmission of infection [28]. Strategies must be proposed to improve access to diagnostics for EPTB (fine needle aspiration cytology, biopsy) in the nearest healthcare facilities [29].

The proportion of pretreatment loss-to-follow-up is lowerthan in other studies in the country [16,30]. This could be due to the dedicated and trained CHW personnel who were responsible for tracing the patients, in case they missed mobile clinic appointments. The CHW were from the same communities, and therefore, were likely to be more accepted by the population [31,32]. However, the uncertain security situation and the movements among the population to nearby cities during nonharvest seasons (for work) posed major challenges in tracing patients.

The study had several limitations. The mobile clinics often faced limitations in routine activities in case of security issues. It was difficult for CHWs to trace patients from far off villages, deep in forest areas. The study results may not be generalized to other tribal areas, as the study is based on a resource-intensive TB program of a medical humanitarian NGO, working in the same area for about a decade. Most of the authors of the study were employees of the NGO, though the research team was not involved in the implementation of the program in the field. We believe that this might have moderated the potential bias in reporting on the implementation of the program. Despite the limitations, this is one of the first studies from a conflict-affected, tribal area in India, describing a unique diagnostic TB model of care based on a routine, mobile clinic-based TB program of an NGO, and thereby documenting the reality on the ground.

#### **5. Conclusions**

A mobile-clinic model of care for TB shows feasibility for replication in similar 'hard to reach' (namely conflict-affected and tribal) areas for improved access to quality TB diagnosis and care. Improved diagnostics such GeneXpert and utilizing CHWs from the communities for tracing and following the patients would be beneficial for early TB diagnoses.

**Author Contributions:** Conceptualization and protocol writing: M.D., H.M., S.K., G.F., P.I.; Data collection: M.D., D.P.; Data analysis: M.D., D.P., S.R., S.S.; Drafting the paper: M.D., D.P., S.R., S.S., S.K.; Critical review and approval for submission: M.D., D.P., S.R., S.S., H.M., S.K., F.N.H., G.F., P.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** We would like to acknowledge the time and efforts of patients with TB and their families, health care providers and project team involved in providing care to patients with TB.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **The Impact of Funding on Childhood TB Case Detection in Pakistan**

**Amyn A. Malik 1,2,3,\* , Hamidah Hussain <sup>2</sup> , Jacob Creswell <sup>4</sup> , Sara Siddiqui <sup>1</sup> , Junaid F. Ahmed <sup>1</sup> , Falak Madhani <sup>1</sup> , Ali Habib <sup>5</sup> , Aamir J. Khan <sup>2</sup> and Farhana Amanullah <sup>6</sup>**


Received: 15 November 2019; Accepted: 13 December 2019; Published: 15 December 2019 -

**Abstract:** This study is a review of routine programmatically collected data to describe the 5-year trend in childhood case notification in Jamshoro district, Pakistan from January 2013 to June 2018 and review of financial data for the two active case finding projects implemented during this period. The average case notification in the district was 86 per quarter before the start of active case finding project in October 2014. The average case notification rose to 322 per quarter during the implementation period (October 2014 to March 2016) and plateaued at 245 per quarter during the post-implementation period (April 2016 to June 2018). In a specialized chest center located in the district, where active case finding was re-introduced during the post implementation period (October 2016), the average case notification was 218 per quarter in the implementation period and 172 per quarter in the post implementation period. In the rest of the district, the average case notification was 160 per quarter in the implementation period and 78 during the post implementation period. The cost per additional child with TB found ranged from USD 28 to USD 42 during the interventions. A continuous stream of resources is necessary to sustain high notifications of childhood TB.

**Keywords:** pediatric TB; verbal screening; contact tracing; resources

#### **1. Introduction**

Childhood TB diagnosis can be difficult and hence many children who develop TB are missed. Of the 10 million people who develop TB each year, 10% or 1 million are children. While national TB programs (NTP) do not report 34% of all incident cases, more than half of children with TB are believed to be missed, resulting in 233,000 deaths each year [1,2]. Modeling studies suggest that 96% of these deaths occur in children who do not access TB treatment [3].

Children with TB are often not diagnosed and reported because of limited capacity of frontline health providers [4,5], lack of dedicated child health services with experienced and appropriately trained clinicians [4], non-specific symptoms overlapping with other common childhood diseases [6], complex diagnostic algorithms [6,7], lack of a sensitive point of care test and technical resources [8], and minimal contact tracing activities [2].

In Pakistan, the estimated TB incidence in 2018 was 265 cases per 100,000 population with approximately 62,000 cases in children. Of the estimated childhood cases, about one in four cases were not notified to the national program [1].

Successful interventions to improve case detection among children have included systematic screening at outpatient departments of hospitals and general practitioners with studies from Pakistan showing that this can increase the case notification among children between 2.5 and 7 times [9,10].

In many high TB burden countries, the response to the epidemic is highly donor dependent. Periodic funding of targeted interventions can lead to increases in diagnosis and notification [11] with a hope that the increase will be sustained given the strengthened health system and capacity building. Recently, the United Nations held a High-Level Meeting on Ending TB (UNHLM), where heads of states committed to mobilize at least 13 billion dollars annually by 2022 for the sufficient and sustainable financing of the global TB response, and to diagnose and treat 3.5 million children with TB between 2018 and 2022 [12].

Our objective is to describe the 5 years trend in childhood case notification in a rural district in Sindh province of Pakistan before, during and after focused active case finding and contact tracing efforts with injection of resources. We sought to understand the impact and cost of finding a child with TB during periodic funding from external sources.

#### **2. Materials and Methods**

#### *2.1. Setting and Study Design*

This study is a review of programmatically collected case notification data to describe the 5 year trend in childhood case notification in Jamshoro district, Pakistan from January 2013 to June 2018.

As part of the Stop TB Partnership's TB REACH wave 4 funding, an active case finding and contact tracing project in a district in rural Sindh was conducted between October 2014 and March 2016. The detailed methodology of this project and results are reported elsewhere [10]. Briefly, the intervention systematically screened all children in outpatient departments of four large public sector hospitals in Jamshoro district for symptoms of TB and conducted household contact tracing of adults and children diagnosed with TB at these facilities. Three of these four hospitals had pediatric TB specialists as part of their medical staff and were already reporting pediatric TB cases. No other center reported pediatric TB cases regularly. One of the four hospitals is a specialized chest treatment center, which treats both drug-susceptible and drug-resistant TB. Community health workers were recruited from the catchment area and trained to administer questionnaires to assess TB symptoms using a custom-built mHealth data collection application with decision support. All individuals with a high likelihood of TB disease were referred to a TB medical officer for free evaluation and testing. Adults with TB and guardians of children diagnosed with TB were also asked to bring their family members to the health facility for TB screening.

At the specialized chest center, active case finding was re-started through a Global Fund initiative in October 2016 by adding one doctor and one nurse and providing support for data collection. The nurse was trained to administer questionnaires to assess TB symptoms in children in outpatient department using a custom-built mHealth data collection application with decision support. All children with a high likelihood of TB were referred to the medical officer for further evaluation and free testing.

There were no other notable changes in the district during the five-year period being analyzed.

#### *2.2. Data Collection*

Age-disaggregated TB case notification data were extracted from the registers of the provincial TB program (PTP) from quarter 1, 2013 to quarter 2, 2018.

Financial data from the TB REACH project was extracted from the accounting system maintained by the finance department. We calculated the operational cost per child verbally screened and cost per TB patient diagnosed through active case finding at the specialized chest center during the implementation period. It included human resources, design, deployment and maintenance of electronic data collection systems and the laboratory tests. We employed two community health workers and one field supervisor exclusively for the intervention and a government employed doctor was incentivized to screen and treat additional children found through the project. The project bore the costs of chest X-rays, Xpert MTB/RIF, Acid Fast Bacilli (AFB) smear and other laboratory and radiological tests as required. Costs were incurred in Pakistani Rupees (PKR) and were converted to US dollars (USD) using the average 2015 exchange rate of 1 USD to 103.1 PKR. *Trop. Med. Infect. Dis.* **2020**, *5*, x FOR PEER REVIEW 3 of 9 electronic data collection systems and the laboratory tests. We employed two community health workers and one field supervisor exclusively for the intervention and a government employed doctor was incentivized to screen and treat additional children found through the project. The project bore the costs of chest X-rays, Xpert MTB/RIF, Acid Fast Bacilli (AFB) smear and other laboratory and

Financial data from the Global Fund project for the support provided to the specialized chest center was extracted from the accounting system from October 2016 to June 2018. During this period, the facility employed one doctor and one nurse. A dedicated doctor was only employed for half of the time period. Chest x-rays, Xpert MTB/RIF, AFB smear and other laboratory and radiological tests as required were done free of charge for the patients. Costs were incurred in Pakistani Rupees (PKR) and were converted to US dollars (USD) using the average 2017 exchange rate of 1 USD to 105.3 PKR. radiological tests as required. Costs were incurred in Pakistani Rupees (PKR) and were converted to US dollars (USD) using the average 2015 exchange rate of 1 USD to 103.1 PKR. Financial data from the Global Fund project for the support provided to the specialized chest center was extracted from the accounting system from October 2016 to June 2018. During this period, the facility employed one doctor and one nurse. A dedicated doctor was only employed for half of the time period. Chest x-rays, Xpert MTB/RIF, AFB smear and other laboratory and radiological tests as required were done free of charge for the patients. Costs were incurred in Pakistani Rupees (PKR)

and were converted to US dollars (USD) using the average 2017 exchange rate of 1 USD to 105.3 PKR.

#### *2.3. Analysis*

We analyzed the changes in quarterly notifications of childhood TB in the district through three periods: (1) a baseline period when no resources for active TB case finding and contact tracing interventions were in place with only passive case finding with no questionnaire-based screening and contact tracing happening (January 2013 to September 2014); (2) an implementation period when active TB case finding and contact tracing interventions were deployed (October 2014 to March 2016); and (3) a post-implementation period when the project ended and additional resources for active case finding and contact tracing were withdrawn (April 2016 to June 2018) (Figure 1). We adjusted for trend in our analysis extrapolating from the baseline period. We used linear regression to calculate the effect of intervention period, a proxy for additional resources, on case notification adjusting for time. *2.3. Analysis*  We analyzed the changes in quarterly notifications of childhood TB in the district through three periods: (1) a baseline period when no resources for active TB case finding and contact tracing interventions were in place with only passive case finding with no questionnaire-based screening and contact tracing happening (January 2013 to September 2014); (2) an implementation period when active TB case finding and contact tracing interventions were deployed (October 2014 to March 2016); and (3) a post-implementation period when the project ended and additional resources for active case finding and contact tracing were withdrawn (April 2016 to June 2018) (Figure 1). We adjusted for trend in our analysis extrapolating from the baseline period. We used linear regression to calculate the effect of intervention period, a proxy for additional resources, on case notification adjusting for time.

**Figure 1.** Figure enumerating the details of different intervention periods Jamshoro District, Sindh, Pakistan between Q1 2013 and Q2 2018. **Figure 1.** Figure enumerating the details of different intervention periods Jamshoro District, Sindh, Pakistan between Q1 2013 and Q2 2018.

We analyzed the changes at the specialized chest center through the baseline January 2013 to March 2015 (27 months) and implementation May 2015 to March 2016 (11 months) periods for the specialized chest center as we had phased in the implementation of the TB REACH project. Because active case finding was re-started in October 2016 at this center through Global Fund resources, the results from this center includes a post-implementation period of two quarters (Q2, 2016–Q3, 2016), and subsequent intervention quarters (Q4, 2016–Q2, 2018) that we refer to as 'New Active Case Finding Project'. We analyzed the changes at the specialized chest center through the baseline January 2013 to March 2015 (27 months) and implementation May 2015 to March 2016 (11 months) periods for the specialized chest center as we had phased in the implementation of the TB REACH project. Because active case finding was re-started in October 2016 at this center through Global Fund resources, the results from this center includes a post-implementation period of two quarters (Q2, 2016–Q3, 2016), and subsequent intervention quarters (Q4, 2016–Q2, 2018) that we refer to as 'New Active Case Finding Project'.

As the specialized chest center received additional resources in the post-implementation period, a more nuanced approach is required to fully understand the trends in notification in relation to available resources. We stratified the data by center type to analyze the trends, separating the specialized chest center from the other centers in the district. We also compared the proportion of TB patients diagnosed and yield of patients diagnosed per child screened across centers during the implementation period to assess the impact by center type. For financial analysis, we calculated the overall cost of the active case finding during the two different intervention phases at the specialized chest center. This cost did not take into account the As the specialized chest center received additional resources in the post-implementation period, a more nuanced approach is required to fully understand the trends in notification in relation to available resources. We stratified the data by center type to analyze the trends, separating the specialized chest center from the other centers in the district. We also compared the proportion of TB patients diagnosed and yield of patients diagnosed per child screened across centers during the implementation period to assess the impact by center type.

existing government infrastructure in place. We calculated the additional cost per additional patient For financial analysis, we calculated the overall cost of the active case finding during the two different intervention phases at the specialized chest center. This cost did not take into account the existing government infrastructure in place. We calculated the additional cost per additional patient found by dividing our overall cost by the trend-adjusted cumulative increase in the case notifications (additional patients) at the center. All analyses were conducted using Microsoft Excel 2019 and Stata version 15 (StataCorp, College Station, TX, USA). *Trop. Med. Infect. Dis.* **2020**, *5*, x FOR PEER REVIEW 4 of 9 found by dividing our overall cost by the trend-adjusted cumulative increase in the case notifications

#### *2.4. Ethical Approval* (additional patients) at the center. All analyses were conducted using Microsoft Excel 2019 and Stata

As this study used de-identified aggregated numbers from existing data sources, this study was exempted from full-review by the Institutional Review Board (IRB) of Interactive Research and Development (IRD). The TB REACH funded project was approved by the same IRB. version 15 (StataCorp, College Station, TX, USA). *2.4. Ethical Approval*  As this study used de-identified aggregated numbers from existing data sources, this study was

#### **3. Results** exempted from full-review by the Institutional Review Board (IRB) of Interactive Research and Development (IRD). The TB REACH funded project was approved by the same IRB.

The average childhood TB case notification rate in the district was 86 a quarter between quarter 1, 2013 and quarter 3, 2014 (seven quarters). It rose to an average of 322 per quarter during the six intervention quarters (quarter 4, 2014 to quarter 1, 2016), a trend-adjusted increase of 2 times (*p* < 0.01). During the nine post-implementation quarters, the average case notification was 245 per quarter, a trend-adjusted increase of 0.9 times (*p* < 0.01) (Figure 2a). **3. Results**  The average childhood TB case notification rate in the district was 86 a quarter between quarter 1, 2013 and quarter 3, 2014 (seven quarters). It rose to an average of 322 per quarter during the six intervention quarters (quarter 4, 2014 to quarter 1, 2016), a trend-adjusted increase of 2 times (*p* <

At the specialized center, the average case notification was 50 per quarter during the baseline period. In 2014, the center had screened 762 household contacts of all ages through passive contact screening with 21 contacts diagnosed with TB disease. The case notification rose steadily throughout the project implementation at the center reaching a peak of 354 children with TB in the first quarter of 2016 with an average of 218 a quarter during this period, a trend-adjusted increase of 2.6 times (*p* < 0.01). There was a fall in the case notification in quarters 2 and 3 of 2016 when no additional resources were available. Starting from quarter 4, 2016 when the new funding for active case finding started, the case notification rose again and reached 207 children with TB in quarter 1 of 2018 with an average of 172 children with TB notified a quarter during this period, a trend-adjusted increase of 1.4 times (*p* < 0.01) (Figure 2b). 0.01). During the nine post-implementation quarters, the average case notification was 245 per quarter, a trend-adjusted increase of 0.9 times (*p* < 0.01) (Figure 2a). At the specialized center, the average case notification was 50 per quarter during the baseline period. In 2014, the center had screened 762 household contacts of all ages through passive contact screening with 21 contacts diagnosed with TB disease. The case notification rose steadily throughout the project implementation at the center reaching a peak of 354 children with TB in the first quarter of 2016 with an average of 218 a quarter during this period, a trend-adjusted increase of 2.6 times (*p*  < 0.01). There was a fall in the case notification in quarters 2 and 3 of 2016 when no additional resources were available. Starting from quarter 4, 2016 when the new funding for active case finding started, the case notification rose again and reached 207 children with TB in quarter 1 of 2018 with an average of 172 children with TB notified a quarter during this period, a trend-adjusted increase of 1.4

The notifications in the remaining facilities in the district are depicted in Figure 2c. The average case notification was 36 a quarter during the baseline period rising to an average of 160 per quarter during the implementation period, a trend-adjusted increase of 3.9 times (*p* < 0.01). Notifications declined to an average of 78 per quarter during the post-implementation period when additional funding ceased, a trend-adjusted increase of 0.8 times (*p* = 0.07). times (*p* < 0.01) (Figure 2b). The notifications in the remaining facilities in the district are depicted in Figure 2c. The average case notification was 36 a quarter during the baseline period rising to an average of 160 per quarter during the implementation period, a trend-adjusted increase of 3.9 times (*p* < 0.01). Notifications declined to an average of 78 per quarter during the post-implementation period when additional funding ceased, a trend-adjusted increase of 0.8 times (*p* = 0.07).

**Figure 2.** *Cont.*

*Trop. Med. Infect. Dis.* **2020**, *5*, x FOR PEER REVIEW 5 of 9

**Figure 2.** (**a**) Trend in childhood TB case notification before, during and after active case finding and contact tracing implementation in Jamshoro District, Sindh, Pakistan between Q1 2013 and Q2 2018. (**b**) Trend in childhood TB case notification before, during and after active case finding and contact tracing implementation at a specialized chest center in Jamshoro District, Sindh, Pakistan between Q1 2013 and Q2 2018. (**c**) Trend in childhood TB case notification before, during and after active case finding and contact tracing implementation at rest of the centers, in Jamshoro District, Sindh, Pakistan between Q1 2013 and Q2 2018. During the implementation period, a total of 1807 children were diagnosed with TB in the four **Figure 2.** (**a**) Trend in childhood TB case notification before, during and after active case finding and contact tracing implementation in Jamshoro District, Sindh, Pakistan between Q1 2013 and Q2 2018. (**b**) Trend in childhood TB case notification before, during and after active case finding and contact tracing implementation at a specialized chest center in Jamshoro District, Sindh, Pakistan between Q1 2013 and Q2 2018. (**c**) Trend in childhood TB case notification before, during and after active case finding and contact tracing implementation at rest of the centers, in Jamshoro District, Sindh, Pakistan between Q1 2013 and Q2 2018.

hospitals with the specialized chest center contributing 820 (45%) of them including 188 children detected through contact tracing. The other three centers contributed 987 (55%) TB cases including During the implementation period, a total of 1807 children were diagnosed with TB in the four hospitals with the specialized chest center contributing 820 (45%) of them including 188 children detected through contact tracing. The other three centers contributed 987 (55%) TB cases including 202 through contact tracing (Table 1). The yield of TB cases diagnosed per child screened from the specialized chest center was 7.5 times higher as compared to the other three centers.


**Table 1.** (**a**) Yield of Active Case Finding in Children by Center type in Jamshoro District, Sindh, Pakistan between Q4 2014 and Q1 2016. (**b**) Yield from household contact investigation by Center type in Jamshoro District, Sindh, Pakistan between Q4 2014 and Q1 2016.

Table 2 summarizes all costs incurred at the specialized chest center for the active case finding from May 2015 till March 2016 (11 months). The majority of the cost incurred (70%) was for salaries of the medical officer and community health workers hired at the center. The next biggest contributor to the cost was development of clinical decision support system (CDSS) with 18% of the total funds expended being used for it. The additional cost per additional child diagnosed was USD 41.8.

**Table 2.** Cost categories at specialized chest center in Kotri through active case finding (May 2015 to June 2018).


Conversion rate for May 2015–March 2016: 1 USD = 103.1 PKR. Conversion rate for October 2016–June 2018: 1 USD = 105.3 PKR.

Costs incurred from October 2016 to June 2018 at the specialized chest center for active case finding in the post implementation period are also summarized in Table 2. The major cost incurred was CDSS development and maintenance cost (44%) followed by salaries of additional staff hired (35%). The additional cost per additional child diagnosed was USD 27.7.

#### **4. Discussion**

Studies from Pakistan, India, Nepal and Nigeria have shown that intensified case finding can result in large increases in childhood TB case notification [9,13–15]. Our study indicates that injection of new resources through focused active TB case finding and contact tracing efforts can substantially raise the baseline TB case notification among children. The peak case notification was reached when

all aspects of the project, active case finding and contact tracing, were fully functional. Once funding for activities ceased, the district saw a marked decrease in childhood TB case notification although a residual effect of the intervention persisted with somewhat elevated notifications despite the removal of funding for new resources. Once funding for active case finding activities started again, a second increase in childhood TB case notification was observed, but the increase was smaller, likely due to limited resources available. Our findings suggest that activities that go above the routine work of the NTP require additional funding to show impact on childhood TB case diagnosis and notification.

Actively screening children for TB resulted in more than doubling of the case notification in the district during the implementation period [10] with the yield from the specialized chest center being 7.5 times higher as compared to the other three centers in this project. We believe the increased yield was due to the profile of the children presenting at the specialized center being a select population with respiratory symptoms. The cost incurred per additional child with TB found through active case finding was less than USD 42. The cost incurred per additional child with TB found through active case finding in the post implementation period was less than USD 28. The lower costs were due to the absence of active contact tracing involving greater costs from home visits.

The three non-specialized centers saw a slow decline in the case notification after the project ended, returning to baseline after almost two years. This slow decline is likely the combined result of establishment of referral behaviors, transfer-out of health center staff, programmatic and health systems strengthening and a modest communication campaign that the TB REACH project implemented. However, with time the institutional memory eroded, trained health staff left for other jobs and things returned to baseline [16,17].

Although we did not setup the evaluation as a strictly controlled trial, our results strongly point to the role that additional funding and resources play in improving performance of case detection. While not all case finding approaches will have an impact on the numbers of people being notified, a combination of different approaches including strengthening of existing systems are needed in order to improve on the status quo [9,10,18]. Most of the time, additional work will cost more money, but if the interventions are impactful, continued support must be sought [19]. As there were no notable changes in the district and the time frame is relatively short, we do not believe that external factors confound our findings.

Active TB screening in outpatient departments requires resources that NTPs do not always have. The cost of active case finding ranges from USD 72 to 963 per patient found depending on the screening algorithm used and the population being screened [20,21]. However, it will always cost more to increase the number of people offered TB services. It is estimated that in 2020 there will be a shortfall of approximately USD 6 billion globally for TB prevention, diagnosis and treatment services given the current available funding of USD 6.8 billion per year [1,22]. A significant portion of the current funding in low- and middle-income countries with high TB burdens outside of the BRICS countries is through large international donors. For example, in 19 of the 30 high burden countries more than 50% of the TB program-specific budget is through international funding [1].

Recently, the United Nations held a High-Level Meeting on Ending TB (UNHLM) where a political declaration on grounds of human rights with a target of successfully diagnosing and treating 40 million people with tuberculosis including 3.5 million children by 2022 was adopted [22]. To achieve these milestones, a greater commitment by the countries to public policies and practices related to TB will need to be realized including additional domestic funding. Increases in domestic funding has been responsible for the progress in BRICS and other European and Latin American countries in their efforts to end the TB epidemic [23]. India provides a good example where domestic funding for TB increased almost four times between 2015 and 2018 and accounts for 77% of the total TB budget of the country with no funding gap in 2019. India has seen case notifications increase by 24% nationally between 2015 and 2018 [1].

In the context of our case study, the budget for Pakistan's national program for the year 2019 is USD 135 million with only approximately 3% of the budget funded through domestic sources and 67% of the budget remaining unfunded [1]. The proportion of domestic funding for TB Pakistan will need to increase dramatically to meet its baseline provision of services as well as provide TB screening and testing facilities at high patient-volume centers, including contact management and follow-up in all districts.

#### **5. Conclusions**

Children have been a historically neglected population in the TB community as they have not been sources of transmission nor can they be diagnosed with the basic tools promoted in the early years of the TB response. Global leaders, including those of Pakistan, have signed the UNHLM for TB political declaration which mandates that countries hold themselves accountable for reaching their targets. Successful, low cost, interventions such as this one that resulted in finding large numbers of missing children with TB should be scaled up with domestic funding to include health centers where patients with respiratory symptoms seek care.

**Author Contributions:** A.A.M., H.H. and F.A. conceptualized the study. A.A.M., S.S., J.F.A., F.M. and A.H. collected the data. A.A.M., H.H., J.C., F.M., A.H., A.J.K. and F.A. performed the analysis. A.A.M., J.C., H.H. and F.A. wrote the initial draft of the manuscript. All authors helped interpret the findings, read and approved the final version of the manuscript.

**Funding:** Active case finding during October 2014 to March 2016 was supported through Stop TB Partnership's TB REACH initiative. TB REACH is generously supported by Global Affairs Canada. Active case finding during October 2016 to March 2018 was supported through The Global Fund funding.

**Acknowledgments:** The authors would like to acknowledge Manzoor Brohi and Provincial TB Program Sind for providing the routinely collected programmatic data used for the analysis and Salman Khan.

**Conflicts of Interest:** J.C. is employed by Stop TB Partnership, but had no role in the decision to fund the active case finding and contact tracing project described; F.A. is the Chair of the WHO's Child and Adolescent TB Working Group. All other authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Tropical Medicine and Infectious Disease* Editorial Office E-mail: tropicalmed@mdpi.com www.mdpi.com/journal/tropicalmed

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com ISBN 978-3-0365-2550-1