Health insurance fraud detection that starts with the documents.
YIRA MedSense helps health insurers and TPAs catch suspicious claims early. It reads every medical document in a claim, compares it with earlier records, checks that the admitted patient matches the ID photo, and flags matches for your investigators, before the claim is paid.
Medically reviewed by Dr. N. S. Santosh Varma, MBBS, FIIM, FIDM, FICCM · Medical Officer, YIRA MedSense · Updated 12 October 2026
{
"document": "medical_history.pdf",
"prior_record_match": "16 of 17 fields",
"prior_matches": 4,
"identity_check": "ID photo vs hospital photo, confidence score",
"status": "sent to investigator"
}Sample output. A field match is a review signal, not a fraud decision.
Where health claim fraud hides.
Most fraud, waste and abuse (FWA) in health insurance does not look unusual on the claim form. It sits inside the documents.
Recycled documents
The same lab report, discharge summary or medical history reused across claims, patients or policies.
Identity substitution
The person admitted to hospital is not the insured person.
Inflated bills
Charges that do not match the hospital's agreed tariff or room category.
Inconsistent records
Lab values, diagnoses and treatment that do not fit together.
Manual reviewers see one file at a time, so patterns across thousands of claims stay invisible.
Four signals your investigators can act on.
Prior-record matching
Every document is parsed into structured fields and compared with earlier parsing results. MedSense shows how many fields match and which ones, for example 16 of 17 fields with 4 prior matches.
Facial verification
The ID photo is compared with the photo taken at the hospital, with a confidence score, so the team can confirm the patient is the policyholder.
Clinical consistency
Lab values flagged against the report's own range, plus diagnoses, ICD-10 codes and drug names, so medical officers can see whether the story adds up.
Bill vs tariff
Hospital tariffs as procedure-level rate tables with contract rules by room category; missing rates stay visible as NA.
A field match is not a fraud determination, and a confidence score is not validated accuracy. MedSense surfaces signals; your investigators decide.
Manual review vs MedSense.
| Fraud check | Manual review | With MedSense |
|---|---|---|
| Reused documents | Caught only if a reviewer remembers an earlier file | Every new document compared with prior parsing results automatically |
| Wrong patient | Visual check of an ID copy, if at all | ID photo vs hospital photo with a confidence score |
| Clinical red flags | Doctor reads every page | Lab values flagged against reference ranges; diagnoses and drugs extracted |
| Overbilling | Manual tariff lookup per hospital | Bill lines checked against structured tariffs by room category |
| Evidence for SIU | Scattered across files | Matched fields and prior records shown side by side |
One queue, four teams.
Fraud / SIU teams
A queue of flagged cases with the evidence attached.
Medical officers
Structured clinical data instead of scans.
Claims heads
Fewer leaks without slowing honest claims.
Network teams
Tariff deviations by hospital.
5,000+ fraud cases flagged for review.
MedSense runs in production at Health Assist Insurance TPA Pvt Ltd and has parsed more than 1 lakh medical documents across 100+ insurance operations. It sits between incoming documents and your claims system: documents arrive by API, mobile app or bulk scanner portal; flags and structured data go back to your TPA portal, insurance system or analytics database. Client-managed deployment is available, and rollout takes about 4–5 weeks.
Questions from claims teams
What is health insurance fraud detection software?
Software that helps insurers and TPAs find claims that may involve fraud, waste or abuse, such as reused documents, identity substitution or inflated bills, so they can be investigated before payment.
What is FWA in health insurance?
Fraud, waste and abuse. Fraud is intentional deception, such as a reused or false document; waste is unnecessary cost; abuse is billing or treatment practice that is inconsistent with accepted standards.
How does MedSense detect duplicate medical claims?
It compares the fields extracted from each new document with earlier parsing results and flags records that match, showing how many fields match and how many prior records were found.
Does MedSense reject claims automatically?
No. It flags cases for human review. Every decision stays with your investigators and medical officers.
Can MedSense verify the patient's identity?
Yes. It compares the ID photo with the hospital photo and gives a confidence score.
Can MedSense detect forged or edited documents and images?
Tampering and forgery checks on documents and images can be built for your requirements as a scoped custom module. They are not part of the standard product today; the standard product flags reused documents through prior-record matching and verifies identity through face matching.
Does it work for TPAs as well as insurers?
Yes. Health Assist Insurance TPA uses MedSense in production.
How quickly can we start?
About 4–5 weeks from sandbox to go-live, starting with a paid pilot of defined scope.
Book your demo today.
See MedSense on the documents and fields that matter to your claims team.