Pulse ·
72% of dementia diagnoses first made in hospital — what needs to change
A NSW Agency for Clinical Innovation study of 91,888 dementia diagnoses found 72.4% were first recorded in hospital — not general practice. Even regular GP attendees had dementia first identified in hospital in 63.1% of cases, usually during an unrelated admission. Comprehensive cognitive assessment takes around 90 minutes and is not adequately funded under current Medicare, which is the structural driver of this gap.
Proposed reforms include disease-specific Medicare items, memory clinics linked to general practice, and dementia navigator roles. These have been named but not yet funded.
Nearly three-quarters of Australians with dementia have it diagnosed for the first time not by their GP, but during a hospital admission.
A NSW Agency for Clinical Innovation study of 91,888 dementia diagnoses between 2014 and 2024 found that 72.4% of first diagnoses were made in hospital. Of those diagnosed in hospital, approximately 66% had been admitted for an entirely unrelated reason — a fall, an acute illness, a surgical procedure — and dementia was identified incidentally.
Among people who regularly attended a GP, 63.1% still had their dementia first diagnosed during a hospital admission. That is not a failure of individual GPs. It is a structural problem, and the study names the structure that failed.
The structural problem
The core barrier is time. Comprehensive dementia assessment — a thorough cognitive history, formal cognitive testing, physical examination, collateral history from a family member, and documentation — requires approximately 90 minutes. Under current Medicare arrangements, that is not economically feasible in a standard general practice appointment.
Professor Perminder Sachdev, co-director of the Centre for Healthy Brain Ageing at UNSW, is direct: GPs need training, GPs need incentives, and GPs need support. All three are currently insufficient.
Additional barriers named in the study: competing health priorities in older patients with complex comorbidities, where dementia screening is deprioritised against acute problems; and specialist memory clinic capacity — some areas face waiting lists exceeding six months, meaning even when a GP suspects cognitive impairment, the referral pathway delivers assessment too late for meaningful early intervention.
The downstream consequences of late diagnosis are significant. Only 4% of the 91,888 patients were living in residential care before diagnosis; 31% were in residential care within five years of diagnosis. The transition between community living and residential care is often precipitous — and frequently happens without a dementia-specific care plan having been in place.
The AIHW reports more than 400,000 Australians are living with dementia, a figure projected to grow as the population ages. Delayed diagnosis is not a benign outcome — the evidence supports better quality of life and stronger carer support when diagnosis and management are structured and early.
What the proposed fixes are
The study and the clinicians responding to it are specific about what would help. Three structural reforms are named:
Disease-specific Medicare items for dementia assessment. A longer, funded consultation item specifically for cognitive assessment would allow GPs to provide what the current rebate structure does not support. The 90-minute assessment that currently falls outside normal billing should have a home in the item schedule.
A national network of memory clinics linked to general practice. The specialist access problem cannot be solved by GP training alone. Memory clinics that connect directly with general practice — and that GPs can refer into without multi-month waits — would shift the diagnostic locus earlier in the disease course.
Dementia navigators. Modelled on the breast cancer nurse coordinator role that transformed care coordination in oncology, dementia navigators would support patients and families through the diagnostic and management journey — care planning, carer education, and coordination between services.
Both sides of this
The data is clear and the proposed fixes are evidence-informed. The concern is that they have been named before, in various forms, and the Medicare structure has not yet changed.
The 2024–2025 federal budget did not fund disease-specific dementia assessment items. Advocacy for a dedicated dementia care coordinator role has not been adopted as policy. Recognising the funding gap on the record is a necessary step; it is not the same as closing it.
What the ACI study adds is harder data than previous advocacy has had: the 72.4% hospital diagnosis rate is a number that quantifies the systemic failure concisely. If the gap between problem identification and funding response has been a political one, sharper data helps make the case.
The maybe verdict here reflects that the fix is structural and unfunded, not that the problem is uncertain.
What this means in general practice now
While systemic reform is pending, the tools available for dementia assessment in general practice include the GPCOG (General Practitioner Assessment of Cognition) and the Mini-Cog, both validated for use in Australian general practice. A longer dedicated appointment — billed as a time-based item or under the health assessment MBS items for older patients — can support more thorough assessment than a standard consultation allows.
Early referral to a geriatrician, neurologist, or old age psychiatrist remains the established clinical path for patients in whom cognitive impairment is suspected. Where waiting lists are long, initiating a referral early — before function is significantly impaired — is the practical response to an imperfect system.
Documentation of cognitive concerns, even where a formal diagnosis cannot be made in general practice, creates a longitudinal record that informs the eventual diagnostic workup. It also tells the patient: I noticed. I’m watching. You are not being dismissed.
Verdict: maybe — the data is stark, the proposed reforms are well-specified, but they are unfunded. Watch for federal budget announcements.
Sources cited
Frequently asked questions
-
Why are most dementia diagnoses happening in hospital rather than general practice?
The NSW ACI data points to structural barriers: comprehensive dementia assessment takes around 90 minutes — not economically feasible in standard general practice appointments — and competing health priorities in older patients mean cognitive decline can be overlooked. Specialist memory clinic waiting lists also extend beyond 6 months in many areas, limiting early referral pathways.
-
What can GPs do now while systemic reforms are pending?
Brief validated screening tools — the GPCOG (General Practitioner Assessment of Cognition) and the Mini-Cog — are practical for general practice use. Longer cognitive assessment appointments can be structured as time-based or health assessment items under current Medicare. Early referral to a geriatrician, neurologist, or old age psychiatrist remains the evidence-based path for confirmed or suspected dementia where specialist access allows.