Comparison of clinical and cost-effectiveness of two strategies using mobile digital x-ray to detect pulmonary tuberculosis in rural India

A total of 10.0 million cases of Tuberculosis (TB) were estimated in 2017, of which only 6.4 million (64%) were diagnosed and notified to national programs [1]. In order to achieve the World Health Organization (WHO) goal of ending the TB epidemic by 2035, we need to identify these missing TB cases and provide appropriate care to them. One of the strategies to identify these missing cases is early diagnosis and systematic screening of contacts and high-risk groups, also known as active case finding (ACF). ACF is defined as the “systematic identification and screening of people with presumptive TB, using tests, examinations or other procedures that can be applied rapidly” [2]. Some of the ACF interventions include screening of household contacts of index TB patients, screening of other high risk groups such as people living with HIV, prisoners, slum dwellers etc. and mass community screening of asymptomatic individuals.

India is one of the 20 high TB burden countries with estimated 2.47 million new TB cases and 421,000 deaths in 2017 [1]. Of the estimated new TB cases, only 65% were notified to the national TB program (NTP) in 2017 [1]. To reduce this gap, the National Strategic Plan 2017–2025 has recommended community-based ACF which includes community mobilization, symptom screening and referral [3].

Few randomized trials have assesses the impact of these ACF interventions on TB burden with conflicting findings. The DETECTB study, a cluster randomized trial in Zimbabwe evaluated an intervention involving screening of 100,000 adults with 2 weeks of cough, using either mobile vans or door-to-door visits. The study demonstrated a decline in disease prevalence from 6.5 to 3.7 per 1000 adults in a before-after study design [4]. The Zambia South Africa Tuberculosis and AIDS Reduction (ZAMSTAR) community trial showed no effect of the enhanced case finding (ECF) intervention on the incidence of TB [5]. A large door-to-door ACF activity under Project Axshya covering ~ 20 million people across 300 districts in India resulted in the detection of a large number of persons with presumptive pulmonary TB and smear-positive TB [6]. Different ACF strategies should be evaluated for their effectiveness to establish the most efficient and effective intervention and inform policy.

Medanta - The Medicity, one of India’s largest multi-super specialty corporate hospital located in Gurugram, Haryana launched a “TB-Free Haryana” Campaign in collaboration with the state Government of Haryana, India. Under this public–private partnership (PPP), a mobile van was equipped with a digital CXR machine with the support from other private players. The van was sent to a government health facility every week. The van used to screen patients with presumptive TB using a “health camp” approach. Two strategies of case finding were used. In strategy 1, adult patients (18 years and above) who, according to the Revised National Tuberculosis Control Programme (RNTCP) diagnostic algorithm were eligible but not able to get a CXR (chest symptomatic with sputum smear negative) underwent CXR. In strategy 2, all chest symptomatic patients (cough > 2 weeks) were called for chest x-ray, regardless of their smear status. Patients with abnormal x-ray findings were recalled for sputum testing. This strategy also involved active information education communication (IEC) and community mobilization one week prior to the case finding activity.

In this study, we aimed to assess the (1) yield and cost analysis of two strategies using mobile digital x-ray to detect PTB in rural Haryana, and (2) impact of the case finding project on overall case notifications in district Mewat, Haryana.

The study findings highlight the feasibility of a public-private partnership in the provision of digital chest x-ray at places where the facility is not available to identify the missing TB cases in the community. In this model, a mobile van equipped with digital x-ray was provided by a corporate hospital with scheduled visits to peripheral health institutes within a not-for-profit partnership. It worked in close liaison with local healthcare providers and followed the RNTCP diagnostic and treatment algorithm.

Diagnosis of smear negative TB often involves a longer health service delay because of the algorithm which needs 11–34 days to establish a diagnosis under the most optimistic scenarios [10, 11]. Poor access to chest radiography and trained staff at peripheral health facilities and eventual loss-to-follow-up when patients are referred to the higher level for x-ray are some of the reasons for under-diagnosis of smear negative cases. These delays result in individual morbidity as well as continued transmission in the community. The sputum negative patient with Pulmonary TB can infect 4 patients per year [12]. Thus, with the recent revised operational guidelines recommending x-ray as a screening tool alongside sputum examination in the diagnostic algorithm, MXU model looks feasible and scalable at least for the high-risk groups, if not all. This model provides digital x-ray in remote locations along with the trained staff (radiographer and chest physician at the district level) without having the patient to visit higher levels of facility for x-ray.

A previous study by Story et al. concluded that cases of active pulmonary disease identified by the MXU were less likely to be sputum smear positive on diagnosis than passively identified cases from the same populations [13]. This is probably because the MXU identifies people earlier in the course of disease, before they become infectious, thereby limiting the transmission of the disease in the community. Another explanation could be the access to x-ray facilities which renders diagnosis of smear negative disease which otherwise would have been missed. The present study showed a significant increase in the Case Notification Rate (CNR) in smear negative TB in the study district during the period of intervention. Strategy 1 involved diagnosis of smear negative TB by providing x-ray facility to smear negative presumptive TB cases. In strategy 2, among 67 patients who had both x-ray and sputum examination done, 11 were smear positive whereas 56 were diagnosed with smear negative TB which substantiates the above stated fact.

The cost per TB case diagnosed and started on treatment was much higher in other studies evaluating a mobile screening unit compared to our screening strategy, although the population and the algorithm of screening in those previous studies vary [14]. One of the reasons for this difference in cost might be due to the presence of culture in the diagnostic algorithm in addition to sputum smear and x-ray examination.

One of the key findings of the study is the cost difference in screening and identifying a case of TB between both the strategies. One possible reason for this difference could be the fact that many people attended the screening camps under strategy 2 compared to strategy 1 which is probably due to the additional IEC component under strategy 2. A structured IEC campaign involving multiple relevant stakeholders was planned prior to each camp in and around the local area. This mobilized the entire community leading to higher attendance at the camps. Thus, IEC should be an integral component of any similar screening camps in order to increase its yield and cost-effectiveness.

Mobile x-ray unit (MXU) has been shown to be an effective tool for active case finding especially in difficult-to-reach populations which are known to have a high burden of disease. This is because of its ability to identify cases at an early stage of the disease when they are less infectious, thereby cutting the spread of the disease. This strategy of improving access to x-ray using a mobile unit has also been demonstrated to be cost-effective in such populations in previous studies [13, 15].

The study strengths are that it was conducted within the routine health system engaging a PPM model and thus operationally relevant and can be easily applied to other settings, thus allowing scale-up elsewhere. This operational research also responds to an identified national research priority.

There were few limitations in this study. First, in strategy 2, nearly one-third of the patients with x-ray suggestive of TB did not undergo sputum examination. This project is now being scaled-up to all the districts in Haryana with more rigor. Second, there was no information on impact of the intervention on the delay in diagnosis and treatment and their treatment outcomes compared to those detected by passive case finding. Third, cost analysis was done in terms of cost per screening a presumptive TB case or TB case diagnosed which rules out comparison with interventions in other subjects or sectors. Fourth, overdiagnosis of TB, especially smear negative ones as seen in this intervention might overestimate the cost effectiveness of this model.

The study reports a new initiative within a PPM model to improve the diagnosis of PTB by filling the gap in the current diagnostic infrastructure. The results of our study are encouraging and point towards chest x-ray screening as a viable tool in TB case finding algorithm strategy, both operationally and also in economic terms. Although we believe there is potential for replication of strategy 2 model in other states, we still need more substantial evidence to support our claims. Strategy 2 has now been scaled up in other states as part of the same project and assessment of the same will provide us with the necessary evidence. TB case detection could be improved by adding Xpert machines to the mobile van. However, this needs to be piloted in future studies and scaled-up.