Finds likely trial patients inside a hospital's records, without exporting them
TL;DR Screenfirst is software installed inside a hospital that reads its lab results, prescriptions and diagnosis codes against a trial's entry criteria and gives the site's coordinator a shortlist to approach. Nothing leaves the building except counts. Sponsors and CROs pay to fill studies faster at fewer, better-chosen sites.
Indian sites often promise enrolment numbers at feasibility that they can't meet, and patients slip away before the site is ready, a pattern one Indian consultant calls recruitment drift[1]. When patients are found, many fail screening on lab values. In a five-year audit of 19 studies at one Indian centre, 59% of screen failures came from abnormal lab results[2].
The silly part? The hospital already had that information sitting in its lab system.
Coordinators still find candidates by asking busy doctors to remember who might fit, and among India-only phase III trials on CTRI, the share completed fell from 56.2% to 27.7% between 2007–2014 and 2015–2025[3].
What it does, in regulatory terms: research pre-screening support only. Screenfirst does not diagnose, recommend treatment or decide eligibility. It produces a candidate list, the investigator confirms eligibility, and the patient's consent under the approved protocol comes before any trial step. Each hospital's ethics committee must approve the use of records for pre-screening.
The software runs on a server in the hospital. A coordinator loads a study's criteria and Screenfirst turns them into queries over the lab, pharmacy and outpatient records, with a language model to read free-text notes where codes are missing. Sponsors only see how many patients match at each hospital, which lets them pick sites on evidence rather than promises.
The seed proposed federated learning over genomic data, and I've reframed it. Few Indian patients outside cancer care have genomic results, and a model trained across hospitals is more than the job needs. A query that runs in place gives you the privacy benefit with far less machinery.
The hospital gets it free or cheap, because its research office earns more when studies enrol. The payer is the sponsor or CRO, through per-study feasibility reports and a fee per pre-screened candidate referred. No public Indian price exists for this. Network-based feasibility data is sold to pharma abroad under private contracts[4].
Indian hospitals are digitising labs and pharmacy records, which makes structured search possible in the larger private chains. CRO registration became mandatory on 1 April 2025, raising the bar on documented site selection[5]. The DPDP Rules, notified on 14 November 2025, turn keeping patient data inside the hospital into a selling point rather than a nicety[6].
There's no public figure for what India spends on trial-recruitment technology. The anchors - 120,027 trials registered on CTRI to date[7], and 1,385 phase III trials registered between 2007 and mid-2025, 84.7% of them industry-funded[3].
The adjacent figure is India's clinical trials market at US$2.05 billion in 2024 (about ₹19,800 Cr), per a trade publication[5]. Only the feasibility and recruitment slice of that is your market.
One global research-data network runs 14,200+ clinical sites in 20+ countries and says data stays inside partner systems[4]. It proves sponsors will pay for site-level patient counts, but its Indian hospital coverage isn't public. Indian CTMS vendors sell trial management and data capture, not hospital-side pre-screening[8]. The gap is a tool that copes with messy Indian hospital records.
| Difficulty | Hard (56/100) |
| Starting capital | ₹1.5–2.8 Cr |
| Time to launch | 12–18 months |
Every hospital's systems are different, and many keep key facts in scanned PDFs or free text. Getting a hospital's IT, legal and ethics committee to approve an install takes months, and that's if things go well. You also need sponsors to trust counts they can't inspect.
No product licence for a pure search tool, but every install needs ethics approval.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't trust the model on eligibility. Messy records produce false matches, so the investigator confirms every candidate and consent always comes first.
Want to work out the AI layer for Screenfirst? Talk to Yash →
Here's what can go south -
I'd say it's worth trying if you have hospital IT integration experience and a contact inside one large private chain. If you're a pure software team that expects clean data then skip it. Prove first that, for one live study, Screenfirst's shortlist yields at least twice the enrolled patients per coordinator-hour of the usual method.
Plan for roughly ₹1.5–2.8 Cr, because each hospital integration is slow and custom.
No product licence for a pure eligibility search. Each hospital's ethics committee must approve the use of records, and you need DPDP processor terms with each hospital.
Yes. Global research-data networks already run this kind of network across 20+ countries. Check your country's health data and research ethics rules before the first install.
The free Screenfirst validation kit has the line-by-line capital breakdown, the effort scoring, the ethics and data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for sites and sponsors, and 5 assumptions to test. Download the kit →
Thinking about building Screenfirst or something close to it? The hard part is the first hospital install, with IT, legal and ethics approval all at once, and that's the kind of thing I help founders plan for. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.