Private Hospital Data Collection Review Final Report

Appendix C

Private Hospital Data Collection Review - Final Report

Page last updated: 14 June 2012

    Comparison framework findings

    C.1   HCP

    C.1.1   HCP: Scope of the collection

    The Hospital Casemix Protocol (HCP) data collection was established as part of the 1995 Private Health Insurance Reform legislation to monitor the deregulation of the private health industry.117 The scope of the collection is for privately insured admitted patient episodes, whether they occur in a public or a private hospital. There are five data specifications within Hospital Casemix Protocol118. They are HCP, HCP1, HCP2, GT-Dental and PHDB. The information flow varies among each specification. HCP involves data flowing from the Hospital to the Insurer, while HCP1 and HCP2 is data that flows from the Insurer to the Department (Note: GT-Dental is not within consideration here, while PHDB is discussed in more detail below).

    117 Australian Health Service Alliance, Hospital HCP Data.
    118 Department of Health and Ageing, Hospital Casemix Protocol (HCP), [Viewed 05/04/2011].

    The HCP data specification is the focus here as the flow of information for this specification is from the hospital to the insurer. Private hospitals supply this data monthly and should submit an episode within 42 days of patient discharge. The legislation that applies to HCP are:

    1. the Private Health Insurance Act 2007 (the Act);
    2. the Private Health Insurance (Health Insurance Business) Rules 2010;
    3. the Private Health Insurance (Data Provision) Rules 2010.

    Insurers do not currently receive any meaningful data on private patient episodes in public hospital (with the exception of Tasmania where the public hospitals use a management system that is widely used in the private sector). Just over 15 per cent of separations which are funded by private health insurance take place in a public hospital (See Table 4). There is no incentive for public hospitals to record and provide additional data (e.g. patient admission, separation date, procedure and diagnoses) as it won’t change the financial outcome for the public hospital.

    Table 13 Separations funded by private health insurance by public and private hospital119.

    Separation Funding Source: Private health insurance Percentage
    Public Hospital 501,819 15.3%
    Private Hospital 2,767,947 84.7%
    Total 3,269,766 100.0%

    Top of page

    119 Australian Institute of Health and Welfare 2011. Australian hospital statistics 2009–10. Health services series no.40. Cat. no. HSE 107. Canberra: AIHW.

    This lack of private patient stays in public hospitals is an issue for insurers as they cannot compare public and private hospital stays or understand the nature of the episode in the public hospitals.

    The Department expects public hospitals to continue to provide the information required by private health insurers to enable the payment of hospital benefits for insured patients (i.e. claiming benefits). As the reporting requirements for HCP and PHDB document states “It is also expected that hospitals will work toward providing health insurers with data that complies with the HCP. No timeframe has been set for public hospitals to provide health insurers with a complete HCP dataset as per the specifications. This recognises the needs of both parties and allows additional time for hospitals to build a complete data provision capability, whilst continuing to provide existing base levels of data. DoHA expects public hospitals and health insurers to continue working towards providing the full HCP dataset."120

    120 DoHA 2010. Reporting requirements for Hospital Casemix Protocol (HCP) and Private Hospital Data Bureau (PHDB). Department of Health and Ageing, Canberra.

    C.1.2   HCP: Data item definitions

    The HCP has clinical, demographic, benefit and charge data for privately-insured admittedpatient episodes nationally. In total there are 76 data items in each HCP submission. Of these 65 data items are collected in the episode record and 11 items for the header record. The HCP collects a large amount of patient and episode level benefit and charge information which is not within scope of the APC NMDS. These data items are defined in the HCP and PHDB manual. Summary comparisons between the data specifications for HCP and APC NMDS can be seen below in Table 5.

    Table 14 Comparison between APC NMDS and HCP.

    Comparison APC NMDS vs. HCP
    Identical 18 items. See spreadsheet for details.
    Mappable 15 items. See spreadsheet for details.
    In APC not HCP 4 items. Country of Birth, Indigenous Status, Funding source for hospital patient, Source of referral to public psychiatric hospital.
    In HCP not APC 43 items.

    C.1.3   HCP: Counting rules

    The base statistical unit in the collection is the episode. However the HCP file submitted each month is split into a header and an episode record. The header record contains establishment level information like Provider Number and HCP version. The header record is often used to validate the submission as a whole. The episode contains information such as sex and admission time. The code sets used by HCP are ICD-10-AM version 7, ACHI version 6, AR-DRG (any version is accepted). The file format of HCP is a fixed width file, where the format is specified by column widths, pad character and left/right alignment.

    Top of page

    HCP data is currently held as a ‘snapshot’ at the point at which the data is sent. This means that any later adjustments are not factored in to the dataset. In comparison to the APC NMDS, DRG changes generally trigger a data re-submission in state reporting. To improve comparability between APC NMDS and HCP, HCP would require a similar system to ensure data is complete and accurate.

    C.1.4   HCP: Costing

    It is difficult to fully cost the time and effort required to provide data for a collection such as HCP. The steps involved are generally extracting data each month from a patient administration system (PAS), checking the extract file, submitting the file for validation to insurers or organisations like AHSA and rectifying issues with the extract. Wholesale changes to a source system such as a Patient Administration System (PAS) or even PAS version upgrades can add a considerable cost to extracting and creating monthly data extracts be they HCP or a jurisdiction based collection.

    C.1.5   HCP: Practicability of changing the collection

    Changes to the HCP data collection do impose a burden on insurers (i.e. data recipients) and typically take six months to fully implement. For insurers or intermediary organisations like (AHSA) change may require negotiation and meetings with IT providers. In the past changes have been signed off by the Department in December for the beginning of the next financial year and this would need to continue.

    The HCP is an established collection. The process of private hospitals extracting data from their systems on a monthly basis and submitting it will continue without a well laid out data convergence plan. The practicability of changing HCP, or any of the other collections, is discussed in more detail in Section 4.

    C.1.6   PHDB: Scope

    The Private Hospital Data Bureau (PHDB) data collection was established in 1997-98. The scope of the collection includes all private hospitals and day facilities. PHDB contains deidentified information on all private hospital separations, including patient demographics, hospital episode, clinical information (ICD-10-AM) and hospital charges for all patients in private hospitals.

    The flow of information is from private hospitals to the Department. Private hospitals supply this data monthly and should submit an episode within 42 days (i.e. six weeks) of patient discharge. The legislation that applies to PHDB is the same as that for HCP, they are:

    1. the Private Health Insurance Act 2007 (the Act);
    2. the Private Health Insurance (Health Insurance Business) Rules 2010;
    3. the Private Health Insurance (Data Provision) Rules 2010.

    C.2   PHDB

    C.2.1   PHDB: Data Item Definitions

    The PHDB format is very similar to the HCP format. Both collections have 11 items in the header record and 65 items in the episode record. The header records collect 10 identical items, with the PHDB not collecting Insurer/Group Identifier. For the episode record the difference between PHDB and HCP is that HCP collects Insurer Membership Identifier, Insurer Identifier, Family Name and Given Name. For the field Insurer Identifier, which is a code that identifies the private health insurer PHDB collects Payer Identifier. Payer Identifier is an item that categorises the organisation that payed for the episode of care. Summary comparisons between the data specifications for PHDB and APC NMDS can be seen below in Table 6.

    Top of page

    Table 15 Comparison between APC NMDS and PHDB

    Comparison APC NMDS vs. PHDB
    Identical 18. See spreadsheet for details
    Mappable 16. See spreadsheet for details
    In APC not HCP 3 items. Country of Birth, Indigenous Status, Source of referral to public psychiatric hospital.
    In HCP not APC 42. See spreadsheet for details

    C.2.2   PHDB: Counting rules

    Like HCP the base statistical unit in the collection is the episode. As mentioned though the PHDB file submitted each month is split into a header and an episode record. The header record contains establishment level information like Provider Number and ICD version. The header record is often used to validate the submission as a whole. The episode contains information such as sex and admission time. The code sets used by PHDB are ICD-10-AM version 7, ACHI version 6, AR-DRG (any version is accepted). The file format of PHDB is a fixed width file, where the format is specified by column widths, pad character and left/right alignment.

    C.2.3   PHDB: Costing

    It is difficult to fully cost the time and effort required for a private hospital to be a participant for a collection such as PHDB. The steps involved are generally extracting data each month from a patient administration system (PAS), checking the extract file, submitting the file for validation to the Department and liaising with the Department to rectify issues with the extract that the Department might discover. Wholesale changes to a PAS (or even PAS version upgrades) can add a lot of cost to extracting and creating monthly data extracts be they PHDB or a jurisdiction based collection.

    C.2.4   PHDB: Practicability of changing the collection

    The PHDB, like HCP, is an established collection. The process of private hospitals extracting data from their systems and submitting it on a monthly basis will continue without a well laid out data convergence plan. The practicability of changing PHDB, or any of the other collections, is discussed in more detail in Section 4.

    C.3   Perinatal

    C.3.1   Perinatal: Scope

    The scope of this dataset includes all births at public and private hospitals but also those at birth centres and those within the community. National data on births are based on notifications to the perinatal data collection in each State and Territory. Midwives and other staff, using information obtained from mothers and from hospital or other records, complete notification forms for each birth in each jurisdiction. Information is included in the NPDC for all live births and stillbirths of at least 400 grams birth weight or at least 20 weeks gestation.

    Top of page

    C.3.2   Perinatal: Data Item Definitions

    Like the APC NMDS data flows from private hospitals to the State or Territory health department and then on to the AIHW. The specification for each State based perinatal collection is unique. The Perinatal NMDS for 2010-11 has 23 metadata items many of which are specific to the birth itself, for example, birth plurality (i.e. singleton, twins, triplets etc.).

    The legislation that enables the collection of data for the Perinatal NMDS is the relevant state legislation. Agreement on the data items that should be included in the NMDS.

    Table 16 Comparison between APC NMDS and Perinatal NMDS.

    Comparison Count of items between APC NMDS and Perinatal NMDS
    Identical 10 – See spreadsheet for details
    Mappable Nil
    In APC not perinatal 27 – See spreadsheet for details
    In Perinatal not APC 13 – See spreadsheet for details

    C.3.3   Perinatal: Counting Rules

    The base statistical unit in the collection is the birth event. However information is also collected on the mother and the pregnancy.

    C.3.4   Perinatal: Costing

    It is difficult to fully cost the time and effort required for a private hospital to be a participant in the different jurisdiction based perinatal data collections. This is, because the collection methodology and content varies from jurisdiction to jurisdiction. For example, Victoria has only recently moved to electronic submissions (it was a paper based reporting method previously).

    C.3.5   Perinatal: Practicability of changing the collection

    The focus of each jurisdiction’s perinatal collection is local. That is, the jurisdiction collects and analyses information on and in relation to the health of mothers and babies in order to contribute to improvements in their health within that jurisdiction. For example, there are 23 metadata items in the NMDS but the Victorian Perinatal Data Collection consists of over 100 data items.121

    Top of page

    121 Victorian Perinatal Data Collection, [Viewed 30 May 2011].

    To change the ongoing jurisdiction focussed collections would be difficult. In addition both public and private hospitals have to provide the same birth report i.e. data from private and public hospitals is comparable.

    C.4   Cancer Registry

    C.4.1   Cancer Registry: Scope

    The AACR consists of the eight Australian State and Territory cancer registries, the New Zealand Cancer registry and the NCSCH at the AIHW122. Cancer registration is an important and fundamental tool in cancer monitoring. Cancer is a notifiable disease in Australia and each State and Territory has a registry that assembles local information about new cases of cancers and about cancer deaths.

    122 Australian Association of Cancer Registries. [Viewed 30 May 2011].

    C.4.2   Cancer Registry: Data item definitions

    Australian States and Territories are required by legislation to maintain a cancer registry of new cases of malignant cancer. Data are supplied to the registries from a range of sources. This includes private and public hospitals, pathology laboratories, radiotherapy centres and the relevant registry of births, deaths and marriages. The state and territory registries undertake a lot of data collation and cleansing prior to providing data to AIHW.

    The data that both private and public hospitals are required to supply to the cancer registry within each jurisdiction is the same. However the data that hospitals must supply to the cancer registries is unique to that particular jurisdiction e.g. what a Victorian private hospital supplies to the Victorian Cancer Registry is not exactly the same (although it is similar) as what a Queensland private hospital supplies to the Queensland Cancer Registry. In addition the method of submission differs between jurisdictions, for example, electronic notifications can be submitted in some states and territories. Generally hospitals must supply the cancer notification within one month of the separation date.

    Table 17 Comparison between APC NMDS and Cancer Registry NMDS.

    Comparison Count of items between APC NMDS and Cancer Registry NMDS
    Identical 6 – See spreadsheet for details
    Mappable Nil
    In APC not Cancer Registry 31 – See spreadsheet for details
    In Cancer Registry not APC 18 – See spreadsheet for details.

    Top of page

    C.4.3   Cancer Registry: Counting Rules

    The base statistical unit in the collection is the tumour. Information is also collected on the patient though.

    C.4.4   Cancer Registry: Costing

    A considerable portion of the cost of maintaining a cancer registry is the work done within the jurisdiction itself. This is because the registry receives many individual notifications from different organisations (e.g. pathology and a private hospital notification). These individual notifications are then cleaned and linked together to create one tumour based record.

    C.4.5   Cancer Registry: Practicability of changing the collection

    The focus of each jurisdiction’s cancer registry is local. In addition data collected by the cancer registry is from a number of organisations, not just private hospitals. To change this collection would be difficult.

    C.5 PHEC

    C.5.1   PHEC: Scope

    PHEC was first run by ABS in 1991-92. It is an annual survey which collects information about the activities, staffing and finances of all private hospitals in Australia. The survey is run annually although there was a break in the series in 2007-08 when the ABS did not collect data due to budgetary constraints.

    The information flow for PHEC is more complex than the other collections. The information flow generally involves data being sent from private hospitals to the ABS and from state/territory health departments to the ABS on behalf of the private hospitals. This complex flow occurs when the private hospital has been sending data to the state or territory health department and they give permission for the state or territory health department to release the data. This data is supplied annually and the legislation that applies to PHEC is the Census and Statistics Act 1905, under which reporting is mandatory. ABS is also bound by that Act in that it cannot release the data supplied by individual private hospitals.

    The comparable public hospital collection is the NPHED, which is compiled from data supplied by the State and Territory health authorities. The scope of this dataset is establishment-level data for public acute and psychiatric hospitals, including hospitals operated for or by the Department of Veterans' Affairs, and alcohol and drug treatment centres.

    C.5.2   PHEC: Data Item Definitions

    The PHEC is split into two forms. One form is for day surgeries while the other form is for acute and psychiatric hospitals. The surveys are almost identical with sections of the survey focussing on general information, hospital characteristics, activities, theatres and special facilities, staffing, finance and admitted patient data (health department can release this data if permission is given).

    Top of page

    Part 6 and 7 of the two PHEC surveys collect admitted patient data. This section collects the total number of separations for various data items. For example, the funding source for patients is collected by number of separations. The code sets used by PHEC align with the National Health Data Dictionary. PHEC also collects number of separations for principal diagnosis (ICD-10-AM version 7), procedure (ACHI version 6) and major diagnostic category (AR-DRG).

    C.5.3   PHEC: Counting Rules

    Unlike HCP and PHDB, which has a statistical unit of the episode the base statistical unit is the private hospital for PHEC. Information is collected at this level but generally reported at the state/territory level.

    C.5.4   PHEC: Costing

    Costing the PHEC is difficult, however the Statistical Clearing House (SCH) provides some information as to the effort required to participate in PHEC. The SCH provides a central clearance point for business surveys that are run, funded, or conducted on behalf of the Australian Government.123 PHEC went through a full review in 1997-98 and the average time to complete the survey was 672 minutes.124 The average time to complete the survey for the private hospitals is probably less than this now though, with the jurisdiction health departments supplying the admitted patient data.

    123 Statistical Clearing House, National Statistical Service. Australian Bureau of Statistics. [Viewed 11/04/2011].
    124 Private Health Establishments Collections, financial Year 1998/1999. National Statistical Service. Australian Bureau of Statistics. [Viewed 11/04/2011].

    The fact that PHEC collects information on the hospital, the facilities, staffing, finance and patients is a major contributor to the cost of the collection. To respond to each survey section would require accessing and analysing information from a number of databases/systems such as payroll systems and patient administration systems. Responsibility for obtaining and analysing this information to respond to the survey may be spread across a number of staff members within the hospital. This would also be a large part of the cost for private hospitals responding to ABS.

    C.5.5   PHEC: Practicability of changing the collection

    Recently the PHEC underwent a review.125.The review sought the views of stakeholders in relation to how they used the collection and how it might be improved. A key outcome for the review was to ensure the collection enables more valid comparisons between public and private hospitals. To achieve this, the ABS aimed to more closely align PHEC with the Public Hospitals NMDS.

    125 Statistical Clearing House, National Statistical Service. Australian Bureau of Statistics.

    ABS translated priority suggestions for change into statistical questions and included them in a revised PHEC form. During October - November 2010, the ABS tested the revised PHEC form with a range of private hospitals in Qld, Victoria and the ACT. This involved testing the new questions to determine if the information was able to be reported by hospitals and the relevance of the questions in a hospital context 126.

    Top of page

    126 Correspondence with ABS.

    C.6   NHCDC

    C.6.1   NHCDC Scope

    The NHCDC was established in 1996-97. The purpose of the collection is to produce the national public and private sector cost weights for Australian Refined Diagnosis Related Group (AR-DRG) and associated analytical tables contained in the Cost Report. The data is used for outcomes measurement, performance information and policy development. It provides the health care industry with a nationally consistent method of classifying all types of patients, their treatment and associated costs.127

    127 National Hospital Cost Data Collection. Australian Department of Health and Ageing.

    For the private sector the NHCDC is the main vehicle by which relativities are developed to inform negotiations between private hospitals and health insurers. These standards will impact on the results of the NHCDC in the private sector, recognising that to date the majority of private hospitals are costed using nationally derived service weights.128

    128 Department of Health and Ageing 2010. Australian Hospital Patient Costing Standards, Version 1.1. DoHA, Canberra.

    The scope of this data set is all public and private hospitals. Private day only facilities with less than 200 acute separations in the financial year are not within scope. The NHCDC is a voluntary collection so unlike HCP, PHDB and PHEC there is no legislation that underpins this collection. Data is supplied on an annual basis.

    Top of page

    C.6.2   NHCDC: Data Item Definitions

    Differences in reported costs between the two sectors result from factors such as different reporting practices and obligations, and admission practices. The key difference in the reporting of costs between public and private sector hospitals in the NHCDC:

    1. predominance of ‘cost modelling’ to produce cost estimates in the private sector, in comparison to ‘patient costing’ for the majority of public hospital.
    2. treatment of teaching, training and research costs.
    3. differing admission practices within and between sectors.
    4. pharmacy costs for private hospitals.
    5. prostheses costs.
    6. Public and not-for-profit private hospitals are partially exempt from paying fringe-benefits tax (FBT) and are not required to pay payroll tax.129

    129 Productivity Commission 2009, Public and Private Hospitals, Research Report, Canberra.

    Not only are there the above mentioned differences but the process of supplying data to the NHCDC process differs for public and private hospitals. The primary difference is that public hospitals submit their data to a State or Territory Health department officer to assess the data before it is sent on to the Department. However, private hospitals are required to undertake this quality assessment themselves.

    C.6.3   NHCDC: Counting Rules

    NHCDC is an annual collection of Australian hospital patient costing data. The purpose is to similar hospitals. The NHCDC also produces national cost weights for AR-DRGs and other statistics relevant for hospital service costing and planning.

    C.6.4   NHCDC: Costing

    Hospitals voluntarily participate in the NHCDC. Private hospitals supply data at their own cost (and generally each private hospital submits its own data). The state or territory health department supplies public hospital data within their jurisdiction.

    C.6.5   NHCDC: Practicability of changing the collection

    NHCDC has been producing national hospital costing results since 1995-96 as a voluntary collection. The collection has been evolving since then, particularly in terms of hospital participation. The NHCDC will need to continue to change though. In the National Partnership Agreement130 all jurisdictions agreed to the costing model being built on the NHCDC.

    130 National Partnership Agreement on Hospital and Health Workforce Reform, 2008.

    Top of page

    C.7   ACSQHC Safety and Quality Indicators

    C.7.1   ACSQHC: Scope

    The Commission recently commenced its official role as an independent, statutory authority,under the National Health and Hospitals Network Act 2011 131. In that capacity, its key functions comprise:

    1. promoting, supporting and encouraging the implementation of initiatives relating to healthcare safety and quality;
    2. collection, analysis, interpretation and dissemination of information relating to health care safety and quality;
    3. publishing reports and papers relating to health care safety and quality;
    4. formulating, promoting and supporting the implementation of standards, guidelines and indicators relating to health care safety and quality, as well as monitoring their implementation and impact;
    5. advising on national clinical standards;
    6. formulating model national schemes for the accreditation of organisations that provide health care services and relate to health care safety and quality;
    7. consulting and co-operating with other entities – including people, organisations and governments – on health care safety and quality; and
    8. promoting, supporting, encouraging, conducting and evaluating training programs and research for purposes connected with the performance of its functions.

    131 Australian Commission on Safety and Quality in Health Care.

    The focus of national indicators of safety and quality program is to develop a process of routine review and action by health care providers. There is a subset of the safety and quality indicators, addressing hospital-level outcome indicators for supportive feedback and review at jurisdiction and facility level. These hospital-based outcome indicators can be generated by jurisdictions or private hospital ownership groups, which hold the source data, and reported back to provider facilities. The safety and quality value lies in developing the report-review-act cycle based on the routine supply of timely and targeted data back to hospitals. A number of jurisdictions are already undertaking a similar process with many of the same indicators.

    There is no national legislation that forces private hospitals to participate in generating the hospital-based outcome indicators. In spite of this fact, it is important to note that representatives of the private sector have actively participated in the development and agreement of these indicators. Moreover, a number of private hospitals and operators of private hospitals have committed to the use of these core indicators.

    Top of page

    C.7.2   ACSQHC: Data Item Definitions

    The core, hospital-based outcome indicators recommended for local generation and review are:

    1. Hospital standardised mortality ratio (HSMR).
    2. Death in low-mortality Diagnosis Related Groups (DRGs).
    3. In-hospital mortality rates for:
        ‒ acute myocardial infarction (AMI).
        ‒ heart failure.
        ‒ Stroke.
        ‒ fractured neck of femur.
        ‒ pneumonia.
    4. Unplanned hospital re-admissions of patients discharged following management of:
        ‒ AMI.
        ‒ heart failure.
        ‒ knee and hip replacements.
        ‒ depression.
        ‒ schizophrenia.
        ‒ paediatric tonsillectomy and adenoidectomy.
    5. Healthcare associated Staphylococcus aureus (including MRSA) bacteraemia.
    6. Clostridium difficile infections.

    There is a data dictionary available for Hospital Associated Infections. However the Commission does not collect information from private hospitals, it is intended to be used at a local level.

    C.7.3   ACSQHC: Counting Rules

    The National indicators of safety and quality are generated at whole of system and hospital level. As it is not a collection as such there is no base statistical unit.

    C.7.4   ACSQHC: Costing

    Costing the whole of system and hospital based outcome indicators, like all the other collections, is difficult. For private hospitals they must participate in the legislated jurisdiction based admitted patient collections (e.g. VAED). The indicators are then calculated by the Commission using the data that flows from private hospitals, to the jurisdictions and then onto the AIHW.

    Creating the hospital based outcome indicators is not mandatory and so the cost of participation is a decision for the private hospital or private hospital group.

    Top of page

    C.7.5   ACSQHC: Practicability of changing the collection

    The commission’s indicators are not a collection as such. The indicators are either generated from existing collections (whole of system) or suggested that they be generated at the local level to improve safety and quality of patient care. In terms of improving comparability between public and private hospital data collections there is no need to change the commission’s indicators.