ABOUT VOLUME 2

Introduction

Previous versions of the NSW Public Hospitals Comparison Data Book have included a second chapter, known as Casemix Measures, on costs by hospital by Australian National Diagnosis Related Group (AN-DRG) (Commonwealth Department of Human Services and Health & 3M Health Information Systems, 1996). However, the costs reported were derived indirectly using cost weights developed for the State and then applied to the total acute care costs reported by individual hospitals.

In order to meet the demand by various stakeholders for more detailed cost information it was decided to expand the range and complexity of the data available. To achieve this aim a second volume of the NSW Public Hospitals Comparison Data Book devoted solely to Casemix was produced.

Volume 2 uses data from the NSW 1997/98 Hospital Cost Data Collection (HCDC) (NSW Health, 1998). That is, the cost averages by AN-DRG are based on the AN-DRG-level costs reported by hospitals submitting data to the HCDC. Where this information was not available for the hospitals within the scope of the HCDC, the methodology employed and the hospital affected are outlined in these introductory notes.

The version of AN-DRGs upon which the costing presented in this publication is based is Version 3.1. Even though a later version of the classification is now available, Version 3.1 is still the official version for use during 1999/00 in NSW. This extended use allows time for a better understanding of the classification to occur system-wide.

Volume 2 includes cost data at a broad level (that is, Peer Hospital group, Area Health Service and hospital), and at a detailed level by hospital. Clinical analyses are presented by AN-DRG and Service Related Group (SRG) (NSW Health, 1996).

Classifications used

Australian National Diagnosis Related Group (AN-DRGs) Version 3.1
(Commonwealth Department of Human Services and Health & 3M Health Information Systems, 1996)

The AN-DRG classification is a classification of acute inpatient services. The version of the classification used in this publication is Version 3.1, which was released in 1995. The classification has 667 classes. For the purposes of this publication, ‘error’, rehabilitation, and mental health and drug and alcohol AN-DRGs have been excluded from the list of AN-DRGs presented. Episodes grouped to ‘error’ DRGs should be re-coded and re-grouped (they are not final DRGs). Any cost averages calculated for these classes would not reflect the true average for the DRG because of the heterogeneity of episodes grouped to the class. Rehabilitation DRGs are not considered acute for the purposes of funding and benchmarking hospitals in NSW (NSW Health, 1999b). Also, the costs reported for mental health and drug and alcohol DRGs were not considered robust. This is because of the inability of some hospitals to separate the costs of mental health patients treated in designated and non-designated wards.

Service Related Groups (SRGs) (NSW Health, 1996)

The SRG classification is a method for grouping hospital inpatient records into categories corresponding to clinical divisions of hospital activity. The classification was developed to assist with the planning of health services. The basis of the classification is the AN-DRG (Version 3.1) classification, although there are categories in the classification that rely on the diagnoses and procedure codes of the record rather than an AN-DRG to be allocated to a class. The classification is used extensively in NSW for activity analyses, cross-Area flow patterns and projections of hospital use. The advantages of presenting information using SRGs over DRGs is that SRGs summarise clinical categories and they reflect services rather than body systems. This means that where a pattern or problem is occurring with a particular service stream, it can be identified at the outset using the SRG classification rather than a DRG-by-DRG analysis. The classification has 36 classes.

Peer Hospital Groups 1997/98 version (NSW Health, 1999c)

The NSW Peer Hospital Classification is a classification of NSW hospitals into groups of hospitals based on size, role, and acuity. The grouping is used as the basis for activity and cost comparisons between hospitals. The version used in this publication is the 1999 revision, which is also used in Volume 1 of the NSW Public Hospitals Comparison Data Book. The analyses in this report are confined to acute hospitals, District level (that is, 2,000 or greater separations) or above, including specialist acute and paediatric hospitals. The seven groups are the groups in the scope of the NSW HCDC.

For more information about Peer Groups click here.

The NSW Hospital Cost Data Collection (HCDC)

The 1997/98 NSW HCDC was the third annual study of the costs of acute care services in NSW public hospitals. The HCDC closed with cost data received from 79 hospitals out of the 83 hospitals in scope. The hospitals within the scope of the HCDC are District level and larger based upon the NSW Peer Group classification.

Both clinical costing and cost modelling approaches to costing are used in NSW. Each approach contributed approximately half of the episodes in the HCDC.

Rigorous standards are applied to the process of collecting cost data in NSW. A Costing Standards Manual containing minimum standards for the collection of cost data is produced to support each year’s collection. Further guidelines and clarification of standards are also provided during the costing process through a group E-mail facility that is subcribed to by costing staff in the system.

Upon submission to NSW Health, the data are further checked for completeness and accuracy. The checks involve logical checks such as the absence of intensive care costs for a hospital with intensive care services. Validation of data also involves comparisons with peers to assess the plausibility of the costs reported.

It is expected that the release of the results of the HCDC in publications such as the NSW Public Hospitals Comparison Data Book will assist in improving data quality. Hospitals will have an opportunity to explore whether any differences in costs are genuine and not due to differences in the costing process.

Methodology

There were six steps involved in obtaining the cost reported presented in this publication.

1. Selection of hospitals in the scope of the publication

Data for this publication were collected from hospitals within the scope of the NSW HCDC. These hospitals are District-level and larger, as classified by the NSW Peer Hospital Groups classification. There were instances where hospitals did not submit data to the collection, or the data were not of sufficient quality, and therefore, could not be used in the analysis. The average cost per casemix weighted separation for these hospitals still contributed to the analysis and was derived by using cost information from the Unaudited Annual Return (UAR). Further information on hospitals for which UAR data were used, rather than HCDC data, is provided in Step 6.

2. Merge of activity data reported through the HCDC and the Inpatient Statistics Collection (ISC)

The second step, which was undertaken as part of the process of developing cost weights for NSW, was to merge the activity files submitted by hospitals through the HCDC with activity data obtained from the ISC. The purpose of this action was to reconcile the activity reported to ensure that all cases that were in scope were being included in the HCDC. The expected differences between these two collections were:

Program/service category classification. Costing officers were specifically asked to examine a combination of sources to allocate episodes to the appropriate Program using the NSW Program Structure and the appropriate service category on admission. That is, episodes classified as acute were not necessarily automatically assigned to the Acute Care Program – other sources were examined to determine the Program and/or the service category to which they should be classified. Where there were differences in Program assignment and/or service category assignment, the HCDC assignments were used.

AN-DRG classification. Episodes are grouped locally, and then re-grouped when submitted to NSW Health after editing. The reasons for differences in AN-DRG classification may be due to the presence of ungrouped cases at the time of costing or re-grouping as a result of edits applied to the episode. The ISC DRG assignment was used as the definitive assignment in instances where legitimate AN-DRG was assigned. Cases reported as ungrouped were removed from the analysis.

Qualification status of newborns. Where there was insufficient clinical support for a baby episode to be classified as qualified, the qualification field was overwritten in the ISC. This reassignment was then used as the definitive assignment.

The unexpected differences occurred where there were additional or insufficient episodes in the HCDC activity file or the ISC. That is, the number of cases reported in these collections did not match. Efforts were made to examine the reasons for the differences, particularly when they were large. Many mismatches were resolved, however, some still remained.

The ISC AN-DRG and number of cases were used when there were differences between the two collections.

3. Merge of cost results by hospital and DRG with activity data

Same day episodes are costed separately from longer stay episodes in the NSW HCDC. The costs of same day episodes in DRGs with large numbers of both same day and longer stay episodes are not currently well understood, therefore, the average DRG cost reported in this publication is the average across all cases grouped to the DRG. This average was calculated for each DRG as follows:

(Acute same day Separations*Same day Average Cost)+(Acute overnight Separations*overnight Average cost)
(Acute same day Separations+Acute overnight Separations)

The same day average cost and the overnight average cost are the average costs by AN-DRG as reported by hospitals.

The averages calculated through the above process were then merged back into the main database by episode.

4. Calculation of casemix weight by episode

A casemix weight was calculated for each episode using the Casemix Standards for NSW 1999/00 (NSW Health, 1999b) for acute care episodes. All the adjustments recommended in the Standards were applied, that is, discount for private patients, discounted cost weight for selected DRGs when occurring on a same day basis, discount for transfer episodes when occurring on the day of admission or the following day, and an additional amount for long stay outlier cases. This was done to adjust for differences between hospitals above those accounted for by the casemix (that is, the DRG profile).

5. Subset of cases to be included in the analysis

A number of observations were removed from the analysis. Their associated costs were also removed where they could be quantified. These observations were of the following types:

Cases assigned to ‘Error’ DRGs:

951 Unacceptable principle diagnosis;
952 Ungroupable;
955 Neonatal diagnosis not consistent with age/weight; and
956 Unacceptable principle diagnosis.

Cases assigned to rehabilitation DRGs:

940 Planned same day rehabilitation; and
941 Rehabilitation.

Cases not assigned a DRG at the time of costing. These were episodes grouped to ‘999’,’1999’ ,’000’ or ‘1000’ DRGs in the HCDC.

Cases occurring in designated psychiatric units.

Cases grouped to mental health and drug and alcohol DRGs. These are:

841 Schizophrenia disorders
842 Paranoia & acute psychotic disorders
843 Major affective disorders
844 Other affective & somatoform disorders
845 Anxiety disorders
846 Eating & obsessive-compulsive disorders
847 Personality disorders & acute reactions
848 Childhood mental disorders
860 Alcohol intoxication & withdrawal
861 Drug intoxication & withdrawal
862 Alcohol use disorder & dependence
863 Other drug use disorder & dependence

6. Reconciliation to the total acute care costs from the Unaudited Annual Return.

The total costs by hospital using the above methodology were compared to the total acute care costs reported in the UAR. Where the totals were substantially different, further analyses were carried out to determine the source of the difference. Where the differences could not be adequately explained, UAR data were used to calculate the cost averages.

The hospitals for which the average cost per casemix weighted separation was sourced from the UAR, due to either non-submission or due to quality issues, were:

Batemans Bay District
Canterbury District
Deniliquin
Gosford
Griffith Base
Illawarra Regional
Inverell District
Muswellbrook District
Wagga Wagga Base
Wyong

Further details on the definitions used in Volume 2 can be found in the Glossary of Terms at the end of this publication.

References

Bridges, J. F. P., and Hanson, R. M. (1999). Rethinking risk: Moving forward rather than looking backward, in Commonwealth Department of Health and Aged Care. (1999). Conference Proceedings: the 11th Casemix Conference in Australia. Canberra: Commonwealth Department of Health and Aged Care.

Commonwealth Department of Human Services and Health and 3M Health Information Systems. (1996). Australian National Diagnosis Related Groups. Definitions Manual Version 3.1.

NSW Health. (1999a). NSW Public Hospitals Comparison Data Book 1997/98. Sydney: Information Management and Clinical Systems Branch, NSW Health

NSW Health. (1999b). Casemix Standards for NSW 1999/00. Sydney: Casemix Policy Unit, Structural and Funding Policy Branch, NSW Health Department.

NSW Health. (1999c). NSW Peer Hospital Groups 1997/98. Sydney: Casemix Policy Unit, Structural and Funding Policy Branch, NSW Health Department.

NSW Health. (1998). NSW Costing and Unaudited Annual Return (UAR) Standards Version 2. Sydney: Structural and Funding Policy Branch and Information and Data Services Branch, NSW Health Department.

NSW Health. (1996). Service Related Groups: A Classification of Clinical Specialties for Services Planning. Sydney: Structural and Funding Policy Branch.




	Contact Us :: Terms & Conditions