Geospatial AI targeting of tuberculosis active case finding in Pune, India: a quasi-experimental evaluation

BACKGROUND: Active case finding (ACF) in India screens large populations to find few cases. AI-NETRA predicts neighbourhood-level tuberculosis (TB) risk from routine notification data and generates field microplans. OBJECTIVE: To compare the yield and efficiency of AI-guided and standard ACF, and frontline acceptability. DESIGN: Prospective quasi-experimental evaluation in four TB Units of Pune district, comparing two-week AI-guided sweeps (April to May 2026) with the routine campaign in the same units (December 2024 to January 2025). Primary endpoint: bacteriologically confirmed pulmonary TB per persons screened; secondary: all-forms yield, presumptive yield, referral precision, number needed to screen (NNS) and cost per case. Acceptability: 16 focus groups (n=100) and a survey of 101 ASHAs. RESULTS: AI-guided teams screened 66,095 people and confirmed 10 pulmonary cases, against 9 among 177,322 in the standard campaign (rate ratio 2.98, 95% CI 1.09 to 8.29); NNS was 6,610 versus 19,702 and cost per case 66% lower. All-forms detection did not differ (15 versus 32; RR 1.26, 95% CI 0.63 to 2.39); the standard campaign had mobile X-ray. 96% of ASHAs reported finding the same presumptive cases while screening fewer people. CONCLUSION: Geospatial targeting raised confirmed yield per person and lowered NNS; total detection did not differ. A concurrent, like-for-like trial is warranted.

Authors

Institutions

Publication Details

Journal
medRxiv
Published
2026-11-12
DOI
https://doi.org/10.64898/2026.11.11.26367599
Primary Topic
Tuberculosis Research and Epidemiology
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Geospatial AI targeting of tuberculosis active case finding in Pune, India: a quasi-experimental evaluation

Himanshu Vashistha, Mrinalini Das, Archana Beri, Radhakishan Pawar et al.
medRxiv
Tuberculosis Research and Epidemiology
preprint

Geospatial AI targeting of tuberculosis active case finding in Pune, India: a quasi-experimental evaluation

Himanshu Vashistha, Mrinalini Das, Archana Beri, Radhakishan Pawar, Ujjawal Rao, Sarabjit Singh Chadha
preprint en

Abstract

BACKGROUND: Active case finding (ACF) in India screens large populations to find few cases. AI-NETRA predicts neighbourhood-level tuberculosis (TB) risk from routine notification data and generates field microplans. OBJECTIVE: To compare the yield and efficiency of AI-guided and standard ACF, and frontline acceptability. DESIGN: Prospective quasi-experimental evaluation in four TB Units of Pune district, comparing two-week AI-guided sweeps (April to May 2026) with the routine campaign in the same units (December 2024 to January 2025). Primary endpoint: bacteriologically confirmed pulmonary TB per persons screened; secondary: all-forms yield, presumptive yield, referral precision, number needed to screen (NNS) and cost per case. Acceptability: 16 focus groups (n=100) and a survey of 101 ASHAs. RESULTS: AI-guided teams screened 66,095 people and confirmed 10 pulmonary cases, against 9 among 177,322 in the standard campaign (rate ratio 2.98, 95% CI 1.09 to 8.29); NNS was 6,610 versus 19,702 and cost per case 66% lower. All-forms detection did not differ (15 versus 32; RR 1.26, 95% CI 0.63 to 2.39); the standard campaign had mobile X-ray. 96% of ASHAs reported finding the same presumptive cases while screening fewer people. CONCLUSION: Geospatial targeting raised confirmed yield per person and lowered NNS; total detection did not differ. A concurrent, like-for-like trial is warranted.

medRxiv
FIND (CH), Government of Maharashtra (IN)
Tuberculosis Research and Epidemiology
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon