Practice Incentives Program Quality Improvement Measures: annual data update 2023–24

Interpreting PIPQI data

Overview

Results included in this report should be interpreted with care, taking into consideration the points raised above. In addition, it should be noted that this report provides information on a specific set of items for PIPQI and does not provide information about the complete range of care that is provided to a client.

Where data are presented as a time series, the results represent national point-in-time proportions of cohorts with a recorded result at each quarter.

These data should be interpreted in conjunction with other administrative and survey data collections where the data from these client-provider interactions are captured, for example, Medicare Benefits Schedule (MBS), the Australian Immunisation Register (AIR), the National Diabetes Service Scheme (NDSS) Register, the Australasian Paediatric Endocrine Groups (APEG) state and territory registers, the National Cancer Screening Register (NCSR), the National Health Survey, and State and Territory Health surveys.

Impact of temporal reference periods on QIM results

In the past two national reports, AIHW reported on variations in the QIM results generated by different extraction tools, revealing discrepancies in the interpretation and coding of the PIPQI technical specifications. Factors contributing to these data discrepancies include the ambiguous wording in the technical specifications and differences in data coding applied by software providers. AIHW, in collaboration with clinical software providers, examined the source of these discrepancies. As QIM 8 is one of the more complex PIPQI measures which refers to the recording status of the four necessary risk factors to enable a cardiovascular disease risk assessment, the following sections examine the impact of the data discrepancies on the overall QIM 8 results by using the recording of tobacco smoking status as a specific example.

The two major CIS providers in Australia, BP and MD, record information on behavioural risk factors, including tobacco smoking status. In both systems, manual input is required to indicate that a client’s smoking status is up-to-date, even if it remains unchanged since the client’s last visit. BP and MD each have two ways of storing smoking status information that can be extracted from the CIS: the first is in a generic clinical patient record and the second is a more detailed smoking-specific record. 

The generic clinical patient record summarises the most recent information for a patient and is updated with a single timestamp that corresponds to the last time any clinical details were updated in the CIS for a patient; this includes family and social history, occupation, smoking and alcohol consumption status. The more detailed smoking-specific record includes information such as current smoking status, types, quantity per day, year started and stopped, cessation advice/support, past smoking history, and is only updated when the client’s smoking status information is updated in the CIS. When assessing how recently the information was recorded, the smoking-specific record is the more reliable data source because the date in the generic clinical record does not necessarily correspond to when smoking status was last updated. Furthermore, when the technical specifications call for information to be current, this does not mean the resulting captured data reflect the most recent episode of care.

Since the recording of a patient’s most recent smoking status in the CIS depends on manual input, and the smoking status for most clients remains unchanged between recent practice visits, applying a condition that only counts smoking status records that have been updated within a fixed temporal reference period (e.g., within the last 12 months) will exclude all records that have not been manually updated within that timeframe. This also means that the recently updated smoking status records are more likely to capture instances where smoking status has changed since the previous visit (e.g. from a smoker to non-smoker) and are less likely to capture instances where smoking status was the same as the previous visit. On the other hand, applying a fixed temporal reference period can be useful for assessing smoking status in at-risk age groups. For example, a 12-month reference period to assess the tobacco smoking status for clients below 30 years old in QIM 2 helps monitor the smoking status and effects of interventions in this age cohort. If no temporal reference periods are applied, so that clients with any smoking status ever recorded are included, then there is no bias introduced because this approach does not rely on manual updates to be made in the CIS when the current smoking status is the same as the previous visit.

Due to ambiguity in the technical specifications, each ET provider may have different approaches for using the data within the generic table or smoking-specific table (or a combination of both tables) from CIS when determining whether a client had smoking status recorded for QIM 8. In April 2024, AIHW analysed a sample of de-identified aggregate PIPQI data from an extraction tool to examine the impact of different reference periods for recording smoking status on the overall QIM 8 proportion, while keeping the reference periods for other risk factors constant. This extraction tool used both the generic and smoking-specific tables to determine a client's smoking status, prioritising the specific table if data was available in both sections and included data for all male and female regular clients aged 45 to 74 years that had no pre-existing CVD. The analysis showed that applying a 12-month reference period for updating smoking status resulted in only 6% of regular clients with all their risk factors recorded. Extending the reference period to 24 months increased this proportion to 12%. Removing the smoking status reference period entirely (i.e. effectively implementing smoking status ever recorded) increased the proportion of regular clients with all CVD risk factors recorded to 45%.

PHN boundaries and residential population

In 2015, PHNs were established with the key objectives of increasing the efficiency and effectiveness of medical services for patients, particularly those at risk of poor health outcomes, and improving coordination of care to ensure patients receive the right care in the right place at the right time (Department of Health 2023). Where possible, boundaries of the PHNs align with Local Hospital Networks (LHNs) or equivalents, or cluster of LHNs, to facilitate collaborative working relationships and reduce duplication of effort. The analysis of data at a regional level allows for the planning, commissioning and provisioning of health services based on the local needs assessments of the community.

In 2022, PHN population by usual residence varied from 63,678 to 1,887,653 people (ABS 2022). Some residents of PHNs may seek health services provided by other adjacent or non-adjacent PHNs. Across PHNs, the percentage of adults (>15 years of age) who visited a general practice in the previous 12 months in 2019–20 varied from 78.6% to 87.1% (AIHW 2021b). Please see Figure 5 for an overview of the PIPQI QIMs and Estimated Resident Population by PHN geographic boundaries.

For estimated resident population in each PHN, please refer to the data tables provided in the Practice Incentives Program Quality Improvement Measures – Data tables for download.