Finding Hidden Hepatitis C Through Better Health Data
Hepatitis C can remain silent for decades while gradually damaging the liver. Many people in Australia do not know they have the virus until a routine blood test, a hospital admission, or symptoms of advanced liver disease reveal the infection. Earlier diagnosis creates a clear path to direct-acting antiviral treatment, which can cure most cases and reduce the risk of cirrhosis and liver cancer.
Using big data to identify undiagnosed cases of hepatitis C means bringing together information that already exists across health services, laboratories and public health programs. When handled responsibly, linked data can help clinicians recognise patterns of missed testing, offer screening at the right moment and reach people who may face barriers to regular primary care.
Why hidden infections remain a health priority
Hepatitis C is commonly transmitted through blood-to-blood contact. Past injecting drug use, unsterile tattooing or piercing, medical procedures performed before effective screening, and sharing injecting equipment can all be relevant exposures. Some people may not consider an event from many years ago important when speaking with a doctor.
Australia has made substantial progress in treatment access, yet undiagnosed infection persists in urban, regional and remote communities. A person living in Brisbane may attend several different services without anyone seeing the complete pattern. Someone in a regional Queensland town may rely on a hospital emergency department because a regular GP is difficult to access. These fragmented encounters can hide an opportunity for testing.
Data analysis can identify people who have repeated liver function abnormalities, consultations associated with blood-borne viruses, or pathology results that suggest further investigation. It should support a respectful clinical conversation rather than label a person or make a diagnosis without their involvement.
What health data can reveal
A secure analytic system may connect pathology records, hospital discharge data, general practice encounters, pharmacy information and hepatitis notifications. Signals could include persistently raised alanine aminotransferase levels, unexplained thrombocytopenia, previous hepatitis B testing, opioid substitution therapy, or a pattern of missed follow-up appointments. Each signal is incomplete, but several together may indicate that a hepatitis C antibody test is appropriate.
Machine learning can rank patients for review, while simpler rules may be easier for clinicians to explain and audit. The safest model is usually one that places a trained professional between the algorithm and the patient. A flagged record should prompt checking, consent and clinical judgement, not an automatic message that may cause anxiety or stigma.
This approach resembles work in other areas of precision health, where information from different settings must be interpreted alongside clinical expertise. The genomic research collaboration described by Brisbane Diamantina illustrates how research institutes and clinics can work together to turn complex data into useful care pathways.
Designing analysis for Australian communities
Australian data is collected across state and territory systems, which can make linkage technically and legally complex. In Queensland, a project may need to work with Queensland Health services, pathology providers, primary care organisations and Aboriginal Community Controlled Health Services. The design should account for different records, coding practices and access arrangements rather than assume that one database represents the whole population.
Privacy protections are essential. The Privacy Act 1988, the Australian Privacy Principles and relevant Queensland Health policies shape how personal information can be collected, linked, stored and shared. Research projects may require ethics review, data custodian approval and clear rules about re-identification. My Health Record information also has specific access expectations and should never be treated as an unrestricted research dataset.
Cultural safety must sit alongside technical safeguards. Aboriginal and Torres Strait Islander communities should be involved in decisions about data governance, interpretation and communication. Programs should avoid framing infection as a personal failure and should recognise the role of discrimination, housing instability, incarceration and limited access to culturally safe healthcare.
Turning a prediction into a care pathway
A useful alert connects a person with testing, treatment and follow-up. A GP practice in Logan might receive a prompt when a patient has repeated abnormal liver results but no recorded hepatitis C antibody test. A hospital in Cairns could use a similar process for people who have frequent admissions and no documented viral hepatitis status. The response should include a private conversation, plain-language information and an easy route to confirmatory RNA testing.
Treatment access also depends on practical details. Direct-acting antivirals are available through Australia’s Pharmaceutical Benefits Scheme, and authorised prescribers can support treatment in primary care and other settings. Clinicians need to explain that a positive antibody result indicates exposure, while an RNA test determines whether current infection remains. After cure, reinfection is still possible, so harm-reduction advice and ongoing support matter.
Digital prompts should fit everyday clinical work. An alert that appears too often will be ignored; one that lacks relevant context may waste appointment time. Flexible options such as nurse-led outreach, telehealth for regional patients, mobile services and testing in alcohol and other drug programs can make a data-informed pathway more realistic.
Measuring benefit without creating harm
Success should be measured across the whole pathway, from a data signal to confirmed diagnosis and completed treatment. Useful outcomes include the number of people offered testing, time from flag to test, confirmed infections, treatment initiation, cure rates and follow-up after treatment. Projects should also examine whether benefits reach people in rural areas, people who use drugs, people leaving prison and communities experiencing poorer health outcomes.
Researchers should test models for bias before deployment and at regular intervals. Missing records can make a population appear lower risk, while over-representation of certain services can cause repeated alerts for the same groups. An independent review process should examine false positives, false negatives, patient complaints and whether clinicians can challenge an algorithmic recommendation.
The Brisbane Diamantina network provides a useful setting for conversations among universities, research institutes, health services and communities. Collaboration can help translate a promising analytics project into a governed, clinically useful program that improves outcomes without weakening trust.
Signals and safeguards for responsible use
Potential signals for clinical review:
- Repeated unexplained liver enzyme abnormalities
- Previous blood-borne virus testing or exposure indicators
- Hospital and primary care records showing missed follow-up
- Treatment or service contacts associated with higher transmission risk
Essential safeguards for implementation:
- Consent, secure linkage and strict access controls
- Aboriginal and Torres Strait Islander data governance
- Clinician review before any patient notification
- Public reporting of benefits, limitations and unintended effects
Big data cannot replace a conversation with a trusted healthcare professional. Its value lies in helping that conversation happen sooner, especially for people whose care is spread across emergency departments, community services, pharmacies and specialist clinics. Health partners across Queensland can use linked evidence to find missed opportunities, make testing routine and connect more people with curative treatment.
Support responsible hepatitis C data research by building partnerships between clinicians, researchers, data custodians and affected communities, then turn validated insights into accessible testing and treatment across Australia.
