UN SDG-3 · Capstone · ISDRC / NGL

Navigating the Silent Pandemic.
Humans lead. AI assists.

An AI-enhanced strategic framework to close India's 84.5% mental-health treatment gap across rural and urban ecosystems — anchored in ASHAs, faith spaces, Panchayats, and a back-office AI layer governed by dignity, consent, and human oversight.

"Invisible does not mean imaginary — dignity is the foundation of healing."
84.5%
Treatment gap
NMHS 2015–16
0.75
Psychiatrists / 100k
WHO benchmark > 3
150M
Indians needing care
< 30M receive it
4 : 1
WHO ROI
Every $1 invested
Key findings

Why a human-led, AI-assisted hybrid — not an app-first model.

01

Population-scale need, structural mismatch

10.56% current prevalence of mental disorders; 84.5% treatment gap. India spends ~0.05% of its health budget on mental health; most states < 1%.

02

Workforce math never closes without task-sharing

Shortage of ~27,000 psychiatrists. At 1,000 entering yearly, ~27 years to reach the WHO norm. 70% practise in urban areas; entire districts have zero.

03

Rural digital infrastructure can't carry frontline AI

Only 53.6% rural internet-proficient. 4.4% own a computer. 24% have internet. Persistent gender device gap. App-first models exclude the poorest.

04

Human-led models have the strongest causal evidence

VISHRAM raised depression coverage 6×. Atmiyata: aOR for recovery 3.0 at 8 months. Healthy Activity Program: 64% remission at ~$6 per BDI-II point.

05

Policy scaffolding exists — implementation lags

Mental Healthcare Act 2017 + Tele-MANAS (29.8L+ calls in 20 languages) + 1.64L Ayushman Arogya Mandirs. But DMHP underspend ~62%.

06

Faith & gathering spaces are proven channels

Polio eradication mobilised Ulema, Friday sermons, Social Mobilization Network. Kumbh Mela deployed IT disease surveillance for millions.

07

AI's legitimate role is back-end and supervised

Triage, decision-support, translation, documentation, supervision — never autonomous diagnosis or unsupervised therapy in low-literacy rural settings.

The Hybrid Model

Humans hold the frontline. AI runs the back office.

A five-tier architecture inside existing DMHP, Ayushman Arogya Mandirs, Tele-MANAS and ABHA — not a parallel system. Every patient touch-point is human; every AI output is clinician-overridable.

Tier 0Human
Community

Faith spaces, Melas, Panchayats, ASHAs, Champions. Stigma reduction & case-finding.

AI role
Back-end SMS/IVR nudges, population dashboards.
Tier 1Human
Sub-centre / AAM

ASHA + CHO. PHQ-9 / GHQ-12 screening, psychoeducation, first-line counselling.

AI role
Offline decision-support, auto-documentation, translation, risk flags.
Tier 2Human
PHC / CHC

Medical Officer mhGAP management, medication, supervision of Tiers 0–1.

AI role
Triage prioritisation, supervision analytics, teleconsult scheduling.
Tier 3Human
District / DMHP

Specialist care; Mental Health Review Board compliance.

AI role
Caseload analytics, surveillance, training content.
Tier 4Human
Tele-MANAS / Tertiary

Tele-referral, crisis routing — integrates with '112'.

AI role
Call routing support, multilingual transcription. Clinician override mandatory.
SafeMind

Back-end orchestration: triage, surveillance, mass-gathering MH (Kumbh IHIP precedent).

MindWeave + Sidhana–Pattojoshi Protocols

Clinical task-sharing pathways — mhGAP / HAP lineage.

PANKHUDI Portal

ABHA-linked data + two-way referral backbone across the tiers.

UN-policy mapping

Every component traceable to a binding instrument.

SDG 3.4 / 3.4.2
Reduce premature NCD mortality; promote mental health & well-being; suicide mortality rate.
WHO Action Plan 2013–2030
Strengthen community-based care; reshape environments; deepen commitment.
WMHR 2022
Inverse-care law correction; task-sharing as the central reform.
MHCA 2017
Rights-based care, advance directives, MHRBs, decriminalised suicide — embedded in 'Healing Justice'.
UNESCO AI Ethics 2021
Human oversight, proportionality, fairness, accountability — embedded in AI back-end-only design.
DPDP Act 2023
Multilingual consent, withdrawal/erasure, data localisation, DPO for significant fiduciaries.
Financial sustainability

Blended finance — the money exists.

Government

DMHP / NMHP (₹522 cr 2015–21, ~38% utilised — absorptive capacity is the bottleneck). Tele-MANAS ₹230+ cr. Ayushman Bharat.

CSR

₹34,908.75 crore spent on CSR in FY2023–24. Mental health is a permissible Schedule VII purpose.

Multilateral

World Bank, ADB, Global Fund-style blended finance windows aligned to SDG 3.4.

Return

WHO: every US$1 invested in scaled depression & anxiety treatment returns US$4 in better health and productivity.