What was published, and what had been absent
The quarterly report to disability ministers comes with supplements: the spreadsheets where the scheme’s workings are visible in a way the narrative report does not attempt. In July we reported that several of them had moved from quarterly to six-monthly, and that the utilisation supplement was already absent from the most recent release.
The Q4 2025-26 release settles what that meant in practice. Q3 carried four items. Q4 carries fifteen. The state and territory supplements, participants by service district, access decisions and first plans, and utilisation are all present again. The agency said the first six-monthly release would be Q4, and it was. Reported here because we reported the gap: the data came back, and a reader who saw our July piece should not be left thinking it did not.
The cost is not that the numbers vanished. It is that for two quarters nobody outside the agency could see them, and what follows is the sort of thing that was invisible in the meantime.
Three gradients, one shape
All three tables below cover the same window, 1 October 2025 to 31 March 2026, measured as at 30 June 2026. The NDIA cuts the window off there deliberately, because payments against a recent quarter are still arriving and would understate it. The national figure for that window is 73.8 per cent.
| Support class | Utilisation |
|---|---|
| Core | 79.2% |
| Capital | 76.1% |
| Capacity building | 56.6% |
| National, all classes | 73.8% |
Core supports are the day-to-day ones: assistance with daily living, consumables, transport. Capacity building is the part of a plan meant to increase independence over time, the therapies and skill-building and support coordination. It is the least used category by a margin of more than twenty points, and it is the category whose whole purpose is to change a participant’s trajectory rather than get them through the week.
| Remoteness | Utilisation |
|---|---|
| Major cities | 75.4% |
| Population over 50,000 | 72.3% |
| Population 15,000 to 50,000 | 71.5% |
| Population 5,000 to 15,000 | 69.0% |
| Population under 5,000 | 64.7% |
| Remote | 67.3% |
| Very remote | 57.2% |
That is a near-monotonic decline from the cities outward, with one inversion: remote sits above population-under-5,000. Eighteen percentage points separate major cities from very remote. A plan is a funding entitlement, not a service, and this is the column where that distinction becomes concrete.
| Plan number | Utilisation |
|---|---|
| First plan | 56.6% |
| Second | 63.6% |
| Third | 66.6% |
| Fourth | 69.9% |
| Fifth or later | 79.5% |
This one rises without exception, and it is the cleanest of the three. Participants receiving in-kind supports and plans shorter than 31 days are excluded, so it covers cash-only participants.
Our reading, offered as a view
Three cuts of the same data, measured on three unrelated axes, and each one puts the same group of people at roughly 57 per cent: new participants, remote participants, and capacity building funding. We are not claiming these are the same people, and the tables do not cross-tabulate, so nobody can say from this release whether a first plan in a very remote community is worse again. That would be a real question to put to the agency.
What can be said is that underspending is not spread evenly and does not look random. The plan-number gradient in particular is hard to read as anything other than a learning curve: the same participant, the same scheme, using more of their funding each time round. If capability with the system is what moves the number, then a first plan being barely half used is a description of how hard the system is to start using, not of how much support somebody needed.
The support-class result points the same way. Core supports are the ones with the most established provider markets and the least arranging required. Capacity building generally needs a therapist with capacity, a coordinator, an appointment, sometimes a report. It is the part of the plan that takes the most work to convert into a service, and it is the part least converted.
The wider distribution
The same supplement records the distribution of plans by how much of them gets used, for participants whose initial plan was approved up to 31 December 2025, counting supports provided to 31 March 2026.
| Plan utilisation | Share of plans |
|---|---|
| 0 to 50 per cent | 32.5% |
| 50 to 75 per cent | 22.3% |
| Over 75 per cent | 45.2% |
Nearly one plan in three is less than half used. That is a different statement from the averages above, and a more uncomfortable one, because an average of 74 per cent could describe a scheme where almost everyone lands near three quarters. It does not. The distribution is spread, with the largest single group above 75 per cent and a third of plans down at the bottom.
The one number not to read as a trend
The supplement also carries utilisation by financial year, and the 2025-26 row reads 72.2
per cent against 76.1 for 2024-25. That is not a decline and should not be reported
as one. The table labels 2025-26 Experience still emerging
: payments against
a year keep arriving after it closes, which is visible in the table itself, where the
2024-25 figure moved from 75.9 to 76.1 per cent between the March and June extracts. On the
settled years the number is close to flat: 76.6 in 2021-22, 76.3 in 2022-23, 77.9 in
2023-24, 76.1 in 2024-25.
Methodology
Every figure here was read by us from the NDIA’s own spreadsheets, not from the narrative report or a summary. Supplement E (National) for Q4 2025-26 was downloaded on 17 August 2026 and the values come from its Table E.112 (utilisation by financial year), E.113 (by plan number), E.114 (by support class), E.115 (by remoteness) and E.45 (distribution of plans by utilisation). Percentages are rounded to one decimal place from the ratios in the file; no figure is estimated or interpolated by us.
The three gradient tables all cover 1 October 2025 to 31 March 2026 as at 30 June 2026, which is the window the NDIA publishes precisely because more recent experience is incomplete. The distribution table covers a different population and window, stated above, and the two should not be combined.
On reaching the data. ndis.gov.au returned HTTP 403 to a
plain request and served an interstitial challenge to a browser, so we did not obtain the
report from there. The supplements are published on the NDIA’s separate data site,
dataresearch.ndis.gov.au, which serves them without obstruction, and that is
where these files came from.
What we have not established. The tables do not cross-tabulate, so we cannot say whether the gradients compound. We have not asked the NDIA about the capacity building figure or the remoteness gradient, and we would publish a response.
Primary sources
- NDIA, Quarterly report supplements (read 17 August 2026): that the Q4 2025-26 release carries fifteen items against four for Q3, the list of supplements moving to a six-monthly cycle, and the statement that the first six-monthly release was scheduled for Q4 2025-26.
- NDIA, Supplement E National 2025-26 Q4 (XLSX, downloaded 17 August 2026 from the page above): Tables E.45, E.112, E.113, E.114 and E.115, the source of every utilisation figure in this piece.
- The Scheme Brief, The NDIA has halved how often it publishes some of the scheme’s most useful numbers (31 July 2026): our earlier report of the cadence change and the absent utilisation supplement.
Something here wrong or out of date? Tell us and we will check it against the sources above and log the outcome.