26 OCT 2026 — Indonesia's national health insurer handles about two million claims every day and has roughly 1,000 people to check them. BPJS Kesehatan now wants hospitals to send those claims as structured electronic medical records instead of hard copies and PDFs, so that its systems can compare what a facility says it did with what it bills. Setiaji, the agency's Director of Information Technology, set out the plan in an interview published on 26 October, and said artificial intelligence could help verifiers spot claims that need a second look.
Setiaji described the gap bluntly. Many hospitals already keep electronic records in their own management systems, yet still submit claims to BPJS Kesehatan as PDFs. In his words, "something has become disconnected."
What the July circular set in motion
The policy behind the interview was signed three months ago. On Thursday 30 July 2026, the Ministry of Health, the Corruption Eradication Commission (KPK), the Ministry of Home Affairs, BPJS Kesehatan and the National Cyber and Crypto Agency (BSSN) signed a joint circular on speeding up the use of electronic medical records linked to SATUSEHAT, the ministry's national health data platform, in managing claims under the National Health Insurance scheme (JKN). A ministry release says Azhar Jaya, Director General of Advanced Health Services, signed for the ministry.
The release presents the circular as national guidance for hospitals using records tied to a patient's national identity number (NIK) as supporting documents for JKN claims. The ministry's Secretary-General, Kunta Wibawa Dasa Nugraha, framed it partly as corruption prevention. A digital system leaves a data trail that supervisors can follow, the release says, and that trail is meant to make irregularities in services and claims harder to hide.
How the records would flag fraud
Setiaji said verification draws on 18 types of data, among them visits, diagnoses, medication, radiology and hospital stays. When those records arrive in a consistent format, BPJS Kesehatan can set the services a facility recorded against the claim it filed and look for mismatches. He gave an example of the kind of mismatch he means: "For example, a diagnosis that does not match the patient's condition. Adding a diagnosis to increase the claim value can be an indication of potential fraud."
AI enters at the review stage. "With the ability to read a document in less than a second, AI is expected to help verifiers identify claims that do not meet the required criteria," he said. He was careful about what a flag means. Each finding, he said, still has to be investigated before anyone can conclude that a claim is wrong.
The agency had already said as much in public. During a visit to a central government hospital in Makassar on 31 July, Setiaji said BPJS Kesehatan was developing AI to process the roughly two million daily claims faster and more precisely, according to a ministry account of the visit. He told the hospital it was inefficient for a digital system to export documents to PDF only for staff to upload them again by hand.
Why fraud is the stated target
The scheme has a record that explains the emphasis. In July 2024 officials from the KPK, the health ministry, the state audit agency BPKP and BPJS Kesehatan disclosed suspected phantom billing and diagnosis manipulation at three private hospitals in North Sumatra and Central Java. KPK deputy Pahala Nainggolan said the hospitals had claimed 4,341 physiotherapy cases, of which only 1,071 had medical records, leaving 3,269 suspected to be fictitious, the health ministry reported. In a sample of 39 cataract patients, only 14 matched their diagnosis.
The ministry's account did not name the hospitals. The figures came from a fraud prevention team, not from a court, and BPKP said it would validate any state loss later. That case turned on claims with no matching medical record behind them, exactly the gap structured records are supposed to close.
Standardisation is the hard part
BPJS Kesehatan works with about 20,000 healthcare facilities that run different systems at very different levels of readiness. "The main challenge is standardisation," Setiaji said. His answer is Fast Healthcare Interoperability Resources (FHIR), the data-exchange standard, applied to diagnosis codes, medication, radiology and laboratory tests so a record means the same thing whichever system reads it. Without that, any mismatch flag is only as good as the data under it.
Some facilities cannot connect at all. Setiaji said around 700 of the country's 10,200 puskesmas have neither internet access nor electricity. The ministry runs a programme supplying Starlink satellite internet to poorly connected areas, and BPJS Kesehatan offers regional IT teams, a helpdesk, a sandbox with simulated data and a three-day "Connectathon" where facilities test whether their systems can exchange data using FHIR.
What patients would notice
Integration also means a doctor could see a patient's history from other providers. "Just enter the Population Identification Number (NIK), and the Master Patient Index appears immediately," Setiaji said. Patients would still register at each facility, and under the new rules they can consent to their treating doctors seeing their records. How useful the exchange proves will depend on how complete those records are and how well access is controlled.
What has not been shown yet
None of this has been measured yet. The rollout is recent, and GovInsider notes that its effectiveness has not been evaluated. The interview and the ministry's releases give no figure for how many claims now arrive as structured records, how many have been flagged, or whether any flag has led to a confirmed case. Those numbers will show whether two million claims a day can be checked by machine before a person looks at them.
