How to Write a PHN Needs Assessment: A Step-by-Step Guide with Data Sources

PHN needs assessments are due every three years under the Department of Health's PHN Program Performance and Quality Framework. Most teams find the analysis straightforward. It's the data assembly that takes weeks.

This is a practical guide to working through a needs assessment without losing a fortnight to the wrong correspondence files. It's structured the way an NA is actually structured: what each section needs to answer, where the canonical data lives, and the gotchas that catch first-timers.

If you've written one before, skip to the section you got stuck on last time.

What an NA has to cover

The PHN Program guidelines require an NA to address, at minimum:

1. A demographic and socioeconomic profile of the catchment population.
2. Health status — mortality, morbidity, risk factors, mental health, Aboriginal and Torres Strait Islander health.
3. Service utilisation patterns — what services are being used, by whom, at what rate.
4. Service gaps and unmet need — where the use rate doesn't match the prevalence.
5. Stakeholder consultation findings — qualitative input from clinicians, providers, consumers.
6. Priorities and intervention areas — the conclusions that feed into commissioning.

Most PHNs structure their NA in that order. The data work is sections 1–4. We'll spend most of this post there.

A short note on cadence: the Department wants a comprehensive NA every three years, but most PHNs run a lighter "annual update" in between. Build your data pipeline so the annual update is a re-pull, not a re-do.

Section 1 — Catchment demographic profile

What it should answer

Who lives in the catchment, how is the population changing, where are they geographically, what's the socioeconomic distribution.

Data sources

• ABS Census 2021 — population by age and sex, country of birth, language spoken at home, household composition, tenure, qualifications, employment. Available at SA1.
• ABS SEIFA 2021 — four socioeconomic indices (IRSAD, IRSD, IER, IEO). Available at SA1.
• ABS Estimated Resident Population (ERP) — annual updates of population by single year of age. Available at SA2, LGA. More current than Census between Census years.
• ABS Population Projections — 5-yearly forward estimates. Available at LGA. Use the medium series unless your audience is comfortable with scenario ranges.
• Department of Home Affairs Settlement Data — recent humanitarian and family migration. Available at LGA. Useful when migration is a planning theme.

How to structure it

Headline at catchment level. Distribution at SA2 or SA1 (this is where you show the variance — see Plan at SA1, Not LGA). Trends over time using ERP and the previous Census.

For each demographic dimension that matters to your priorities, show the catchment vs. comparator. Most PHNs pick one of:

• State average
• All-PHN average
• Two or three "similar" PHNs (matched by population size and metro/regional/remote mix)

Be opinionated. "Our catchment looks much like NSW on age but has 1.8x the rate of overseas-born residents" is a useful statement; "our catchment differs from NSW across several measures" is filler.

Gotchas

• ERP vs. Census population. Census is point-in-time; ERP is annual. They will not match exactly. Pick one for the headline and stick to it (most PHNs use ERP).
• SEIFA at very small geographies. SEIFA scores at SA1 are intentionally noisy at the tails. The decile is the planning-friendly summary; the raw score should mostly stay in appendices.

Section 2 — Health status

What it should answer

What are people dying from, what are they living with, what are the leading risk factors, where are the gaps for priority populations (Aboriginal and Torres Strait Islander people, mental health, older people, children).

Data sources

• AIHW METeOR indicators — the canonical population-health indicator library. Includes the National Health Information Indicators (NHIA) used in COAG reporting. Most are available at LGA.
• AIHW National Mortality Database — cause-of-death data. Available at SA3 or LGA, depending on the cause. Available with a lag (typically 2 years).
• ABS Causes of Death — companion dataset to AIHW's, published annually. Available at LGA.
• AIHW Aboriginal and Torres Strait Islander Health Performance Framework — disaggregated metrics for First Nations populations. Some at LGA, most at SA3 or state.
• AIHW MyHospitals — admission rates, separations, ED activity by Local Hospital Network. Maps imperfectly to PHN catchments.
• ABS National Health Survey — risk factor and chronic-condition prevalence. Available at state. Use for catchment comparison against the national distribution; not granular enough for sub-catchment analysis.
• ABS National Aboriginal and Torres Strait Islander Health Survey — companion for First Nations health status. State-level.
• State health departments — chronic disease registries, screening data. Coverage varies by state. NSW Health, Victorian Department of Health, Queensland Health, WA Health, and SA Health all publish dashboards at varying granularity.
• National Cancer Screening Register — bowel, breast and cervical screening participation. Available at LGA via AIHW Cancer Data.

How to structure it

Cover the big mortality categories first: cardiovascular, cancer, diabetes, respiratory, mental health and suicide, injury. Then a section on chronic-disease morbidity (prevalence and management). Then a section per priority population.

For each indicator, the test is: does our catchment differ meaningfully from the comparator? "Meaningfully" needs a definition — most NAs use a 10% relative difference or a confidence-interval check.

Gotchas

• Indicator coverage at LGA. Not every AIHW indicator is published at LGA. Some are state-only. Build your indicator inventory before you write — discovering that the indicator you wanted doesn't exist at your geography is mid-draft is painful.
• Lag periods. Mortality data lags 18–24 months. Risk-factor data from NHS lags 12–18 months. Be explicit about reference periods in every chart caption.
• First Nations data caveats. Small numbers in some catchments make Aboriginal and Torres Strait Islander-specific indicators noisy at LGA. Consider rolling up to SA4 or state for the catchment-comparison narrative, and reserve LGA-level data for trend analysis.

Section 3 — Service utilisation

What it should answer

What primary-care and related services are being used in the catchment, by whom, at what rate, and how that compares.

Data sources

• Medicare Benefits Schedule (MBS) data — Services Australia publishes claims data by demographic and provider type. Available at SA3 or LGA via PHN Data Portal and ABS-published extracts.
• Pharmaceutical Benefits Scheme (PBS) data — script counts and ATC categories. Available at SA3 or LGA.
• Practice Incentives Program (PIP) data — GP practice participation. Available at LGA.
• Australian Immunisation Register — childhood and adult immunisation coverage. Available at PHN catchment level via AIHW.
• AIHW Mental Health Services in Australia — Medicare-subsidised mental health services. Available at PHN catchment level.
• State emergency-department activity — typically published quarterly by state health departments at LHN/hospital level. Joins to PHN catchment via patient postcode.
• Aged Care Provider Data — My Aged Care service availability. Available at LGA.
• Workforce data — GP, allied health, specialist workforce by area. Available via the National Health Workforce Dataset (NHWDS) and Health Workforce Locator.

How to structure it

A service-use overview first (GP attendance per capita, MBS spend per capita, PBS scripts per capita). Then by service line: chronic disease management (PIP-CDM, CDM Medicare items), mental health (MBS BAP items, Better Access), preventive care (immunisation, screening). Then a workforce section.

For each, the planning question is: is the use rate matched to the prevalence? If the catchment has a chronic-disease prevalence 1.5x the state average and an MBS-CDM-item rate at 0.9x the state average, that's a gap. That's where commissioning goes.

Gotchas

• MBS by postcode. Medicare publishes claims by patient postcode of residence. To map to a PHN catchment, you need POA → LGA → PHN. Population-weighted overlaps, with the postcode gotchas covered in the catchment-mapping guide.
• Provider postcode vs. patient postcode. Most service-use data is patient-residence-based, which is the right denominator for "did residents of our catchment get the service." Provider-postcode data tells you a different thing — "did people get this service from a provider in our catchment", which includes patients from outside. Don't confuse them.
• Workforce supply vs. demand. The Health Workforce Locator gives you supply (FTE per 100,000). Demand-side indicators (waiting times, declined-referrals) are much harder to pin down and usually require state-specific data requests.

Section 4 — Identifying service gaps

What it should answer

Where the catchment's pattern of need does not match its pattern of service use.

This is the analytical heart of the NA. Sections 1–3 are inputs; this is where you say "we have a gap".

Method

A simple framework that holds up:

1. For each priority area (e.g. chronic disease management, mental health, immunisation, screening), pull a need indicator and a service-use indicator at the smallest available geography.
2. Standardise both to a comparable measure — most commonly z-scores against the state average.
3. Map the difference. Areas where need is high and service use is low are the gaps.
4. Validate against stakeholder consultation. The data shows you where to look; the consultation confirms whether the gap reflects access, awareness, workforce, cultural fit, or some combination.

Tools

For a serious gap analysis, you want hotspot maps at SA2 and SA1. Tools like QGIS, R {tmap} and Python geopandas handle this. Most PHN data tools assume LGA presentation — fine for board narrative, less useful for hotspotting.

Domato PHN generates SA1-level need-vs-use overlays out of the box for the common indicators; we built it specifically because hotspotting was the most time-expensive part of NAs at scale.

Section 5 — Stakeholder consultation

Not a data section, but worth a paragraph. The Department wants evidence of consultation with:

• Clinical Council members
• Community Advisory Committee
• Local Hospital Networks
• Aboriginal Community Controlled Health Organisations
• Service providers (GP, allied health, mental health, aged care)
• State and Commonwealth agencies

The standard PHN consultation is online survey + 4–8 in-person workshops + key-informant interviews. Output is a narrative section in the NA, sometimes with a separate consultation report annexed.

The most useful step is feeding the data findings back to clinicians and asking "does this match what you see?". When the data and the clinical perspective disagree, that's the most interesting finding in the document.

Section 6 — Priorities and intervention areas

What it should answer

The NA must conclude with a ranked list of priority areas — typically 5–10 — that feed into the commissioning cycle.

Method

Three filters most PHNs apply:

1. Magnitude — is the gap big enough to matter at the scale of the catchment? A 5% gap in a 900,000-person catchment is more impactful than a 50% gap in a 5,000-person sub-catchment.
2. Tractability — is there a primary-care intervention that addresses it? Some gaps (e.g. unemployment-driven mental health) are upstream of primary care.
3. Alignment — does it match the Department's PHN priorities (mental health, aged care, alcohol and other drugs, population health, workforce, digital health, Aboriginal and Torres Strait Islander health)?

A short ranking table — gap × magnitude × tractability × alignment — is enough.

The cadence

A practical timeline for a needs assessment from scratch:

Phase	Weeks
Project plan + indicator inventory	2
Data assembly	4–6
Analysis + drafting sections 1–4	4
Stakeholder consultation	4
Synthesis + priorities	2
Review + Department-facing edits	2
Total	18–20 weeks

The data-assembly phase is where most teams overspend. If you have a pre-built data layer that already maps your catchment to ABS / AIHW / state datasets, this compresses to 1–2 weeks of validation rather than 4–6 weeks of stitching.

If you don't, you'll spend that month. Many PHNs do.

A few principles that hold up across NAs

1. Plan at SA1, present at LGA. Cover this in detail in Plan at SA1, Not LGA.
2. Always show the distribution, not just the mean. A catchment with a flat distribution is a different planning problem to one with a long tail.
3. Pick a comparator and stick to it. State average, all-PHN average, or 2–3 similar PHNs. Switching comparators between sections is one of the things reviewers flag most.
4. Be explicit about lag periods. Every chart should say what time period it represents.
5. Make the appendix do the heavy lifting. Tables, source citations, methodological notes go in the appendix. The body is for the narrative.
6. Build the data pipeline once. The annual update should be a re-run, not a redo.

A good NA is a short, readable document supported by a long, transparent appendix. That's the shape that gets read by board members and trusted by reviewers.

---

Domato PHN is the white-labelled population-health data layer for Australia's 31 Primary Health Networks. Pre-built mapping of every catchment to LGA, SA2, SA1 and postcode; ABS, AIHW, ACARA and state open data pre-joined. Designed for the NA + annual-update + board-pack cadence. Live in production with Sydney North Health Network. See the product page or reach out for a pilot.