Dementia funding class was the only cognition class cut in AN-ACC recalibration

Tyler Fisher, Data Scientist, Mirus Australia.
Tyler Fisher, Data Scientist, Mirus Australia.

New analysis of 573 residential aged care services has found the AN-ACC funding model responded more strongly to physical immobility than cognitive load, with Class 8, the classification covering many ambulant residents living with dementia, and the only cognition class to record a funding reduction in last year’s recalibration.

The analysis by Mirus Australia Head of Data and Insights Ty Fisher examined AN-ACC classification, funding and care minutes data across 573 residential aged care services operated by 134 providers, covering 46,696 permanently classified residents. The analysis is based on a point-in-time extract from 22 September 2026. Funding comparisons use the variable component of AN-ACC only and exclude the Base Care Tariff, initial entry adjustment and supplements. 

The cohort closely mirrors the national AN-ACC class distribution, with every permanent class within 0.1 percentage points of the national profile. 


The findings, released during this week being Dementia Action Week, suggest the funding model places considerably greater weight on physical dependency than cognitive and behavioural complexity. Download the full report here.

Mirus found that every additional 10 percentage points of not-mobile residents added $10.26 per resident per day to a home’s variable claim, compared with $5.66 for the same increase in ambulant cognitive load.

The analysis found the two measures had markedly different relationships with average NWAU: the correlation was 0.77 for the proportion of residents who were not mobile, compared with just 0.11 for ambulant cognitive load. 

Class 8 funding falls

The difference is also evident in individual classifications.

A Class 8 resident described in the report as ambulant with low cognition and often among the most behaviourally complex residents in a home attracted $177.38 per day under the current settings.

A Class 11 resident, classified as not mobile with lower function and lower pressure sore risk, attracted $201.04 per day.

That represents a difference of $23.66 per day, or $8,636 a year per resident. 

The October 2025 AN-ACC recalibration further highlighted the disparity.

Class 3 received the largest proportional increase within the cognition branch, rising 13.2%, from $104.50 to $118.26 per day.

However, Class 3 contained just 0.53% of residents.

By comparison, Class 8 funding fell 1.9%, from $180.76 to $177.38 per day. It accounted for 11.3% of residents in the Mirus cohort.

Mirus said Class 8 was the only class within the cognition branch to move backwards in cash terms.  

Dementia isn’t recorded by name

Under AN-ACC, diagnoses such as dementia are not themselves used as cost drivers. Instead, the model classifies the consequences of conditions, including impaired mobility, functional decline, cognitive loss and behavioural disturbance. 

This means dementia does not appear explicitly in funding records.

The Mirus analysis found that the classification system also shifts many ambulant residents with cognitive impairment into the assisted mobility branch.

Across the cohort, just 249 residents, or 0.53%, were in Class 3, while Classes 6, 7 and 8 together accounted for 34.6% of residents. Mirus therefore defined its broader ambulant cognitive load cohort as residents in Classes 3, 6, 7 and 8, representing 16,421 residents or 35.2% of the permanently classified population. 

Mr Fisher said the model’s diagnosis-blind design meant the evidence supporting a resident’s classification increasingly depended on how cognitive and behavioural needs were documented.

“Dementia is now Australia’s leading cause of death, and around 446,500 Australians are living with it, yet our funding model literally cannot hear the word,” he said.

Care minutes tell a different story

Importantly, the analysis did not find that homes with higher cognitive loads were systematically failing to deliver their care-minute targets.

Total care-minute performance was broadly flat across homes as ambulant cognitive load increased, with a correlation of just 0.02.

Homes in the lowest cognitive-load quintile delivered 103% of their total care-minute target, compared with 103.3% in the highest quintile.

RN minutes increased from 109.3% to 113.3% of target across the same range. 

Mirus said this suggests the care-minutes component of the system is absorbing cognitive load because both the funding claim and the care-minute target are derived from the same AN-ACC class mix. 

The report therefore distinguishes between whether homes are meeting their staffing targets and whether the funding attached to residents adequately reflects the workload involved.

A funding design tension

The analysis identifies Classes 7, 8 and 11 as particularly important because they collectively account for 43.4% of residents and sit around the boundary between assisted mobility and the not-mobile branch.

According to Mirus, a resident who wanders, resists care and requires two staff members to redirect them can fall into Class 7 or Class 8 depending on their assessed cognition, creating a $17.73 per day funding difference.

If that resident subsequently becomes sufficiently immobile to move into Class 11, the funding increases by another $23.66 per day, even though the behavioural workload may have reduced. 

Mirus argues the findings warrant closer examination of how the AN-ACC model captures the workforce demands associated with behavioural and cognitive complexity.