A new clinical complexity model for the Australian Refined Diagnosis Related Groups.

BACKGROUND: The Australian Refined Diagnosis Related Groups (AR-DRG) underwent a major review in 2014 with changes implemented in Version 8.0 of the classification. The core to the changes was the development of a new methodology to estimate the Diagnosis Complexity Level (DCL) and to aggregate the complexity level of individual diagnoses to the complexity of an entire episode, resulting in an Episode Clinical Complexity Score (ECCS). This paper provides an overview of the new methodology and its application in Version 8.0. METHOD: The AR-DRG V8.0 refinement project was overseen by a Classifications Clinical Advisory Group and a Diagnosis Related Groups (DRG) Technical Group. Admitted Patient Care National Minimum Dataset and the National Hospital Cost Data Collection were used for complexity modelling and analysis. RESULT: In total, Version 8.0 comprised 807 DRGs, including 3 error DRGs. Of the 321 Adjacent DRGs (ADRGs) that had a split, 315 ADRGs used ECCS as the only splitting variable while the remaining 6 ADRGs used splitting variables other than ECCS: 2 used age and 4 used transfer. DISCUSSION AND CONCLUSION: A new episode clinical complexity (ECC) model was developed and introduced in AR-DRG V8.0, replacing the original model introduced in the 1990s. Clear AR-DRG structure principles were established for revising the system. The new complexity model is conceptually based and statistically derived, and results in an improved relationship with actual variations in resource use due to episode complexity.

Australian diagnosis related groups: Drivers of complexity adjustment.

BACKGROUND: In undertaking a major revision to the Australian Refined Diagnosis Related Group (ARDRG) classification, we set out to contrast Australia's approach to using data on additional (not principal) diagnoses with major international approaches in splitting base or Adjacent Diagnosis Related Groups (ADRGs). METHODS: Comparative policy analysis/narrative review of peer-reviewed and grey literature on international approaches to use of additional (secondary) diagnoses in the development of Australian and international DRG systems. ANALYSIS: European and US approaches to characterise complexity of inpatient care are well-documented, providing useful points of comparison with Australia's. Australia, with good data sources, has continued to refine its national DRG classification using increasingly sophisticated approaches. Hospital funders in Australia and in other systems are often under pressure from provider groups to expand classifications to reflect clinical complexity. DRG development in most healthcare systems reviewed here reflects four critical factors: these socio-political factors, the quality and depth of the coded data available to characterise the mix of cases in a healthcare system, the size of the underlying population, and the intended scope and use of the classification. Australia's relatively small national population has constrained the size of its DRG classifications, and development has been concentrated on inpatient care in public hospitals. DISCUSSION AND CONCLUSIONS: Development of casemix classifications in health care is driven by both technical and socio-political factors. Use of additional diagnoses to adjust for patient complexity and cost needs to respond to these in each casemix application.

Keeping Our Classification UP to Date.

In Australia, the National Centre for Classification in Health (NCCH) is ultimately responsible for updating ICD-10-AM disease codes and Australian Classification of Health Interventions (ACHI) procedure codes, and the accompanying Australian Coding Standards (ACS). New editions of these publications are released every two years. This article outlines the updating procedure and lists the sources of information upon which the NCCH draws when compiling data for new coding manuals.

An evaluation of the quality of obstetric morbidity coding using an objective assessment tool, the Performance Indicators For Coding Quality (PICQ).

The Performance Indicators for Coding Quality (PICQ) is a data quality assessment tool developed by Australia's National Centre for Classification in Health (NCCH). PICQ consists of a number of indicators covering all ICD-10-AM disease chapters, some procedure chapters from the Australian Classification of Health Intervention (ACHI) and some Australian Coding Standards (ACS). The indicators can be used to assess the coding quality of hospital morbidity data by monitoring compliance of coding conventions and ACS; this enables the identification of particular records that may be incorrectly coded, thus providing a measure of data quality. There are 31 obstetric indicators available for the ICD-10-AM Fourth Edition. Twenty of these 31 indicators were classified as Fatal, nine as Warning and two Relative. These indicators were used to examine coding quality of obstetric records in the 2004-2005 financial year Australian national hospital morbidity dataset. Records with obstetric disease or procedure codes listed anywhere in the code string were extracted and exported from the SPSS source file. Data were then imported into a Microsoft Access database table as per PICQ instructions, and run against all Fatal and Warning and Relative (N=31) obstetric PICQ 2006 Fourth Edition Indicators v.5 for the ICD-10- AM Fourth Edition. There were 689,905 gynaecological and obstetric records in the 2004-2005 financial year, of which 1.14% were found to have triggered Fatal degree errors, 3.78% Warning degree errors and 8.35% Relative degree errors. The types of errors include completeness, redundancy, specificity and sequencing problems. It was found that PICQ is a useful initial screening tool for the assessment of ICD-10-AM/ACHI coding quality. The overall quality of codes assigned to obstetric records in the 2004- 2005 Australian national morbidity dataset is of fair quality.

A profile of coding staff in Sydney metropolitan public hospitals.

This survey assessed the profiles of ICD-10-AM coding staff employed in 13 major, acute care public hospitals in Sydney, Australia, during a two-week period in 1999. Approximately 90% (56/61) of respondents gave their job title as Clinical Coder or Coding Clerk; of these, 20 (36%) were qualified Health Information Managers, of whom 10 coded for >or=90% of their work-time and three for <75% of the time. One quarter of all Clinical Coders/Coding Clerks spent >25% of their work time performing duties other than coding. Five Health Information Management (HIM) Clinical Coders/Coding Clerks were paid under the Clerical, rather than the HIM, Award.

Correction: Re: Acknowledgment: A Profile of Coding Staff in Sydney Metropolitan Public Hospitals Health Information Management, Vol 32(2).

The authors would like to thank Adam Bennett, who collected the raw data used in this study for his thesis submitted for the degree of Bachelor of Applied Science (Health Information Management) (Honours) at The University of Sydney.

Coding productivity in Sydney public hospitals.

The aims of this study were to compare Sydney public hospitals regarding medical record coding times to compare observed coding times with coding times necessary to avoid backlog and to evaluate the impact on coding time of casemix complexity, coder age, experience, job satisfaction, employment status, and salary. Coding time (in minutes) for each medical record over a two-week period was documented by 61 coders employed in 13 hospitals: six principal referral (PR), six major metropolitan (MM), and one paediatric specialist (PS) hospitals. The mean coding time for each coder was estimated by averaging across coding times for all records during the two-week period. In order to compare hospital mean coding times, the hospitals were grouped into PR and MM/PS groups. The mean coding time necessary to avoid coding backlog (expected coding time) for each hospital group was based on the total number of annual separations and filled full-time equivalent coding positions. The observed mean coding time was longer in the PR group than in the MM/PS group (p = 0.019); however, the observed coding time was within the expected coding time limit in both the PR and MM/PS groups. Casemix complexity tended to influence coding time, but neither age, experience, job satisfaction, employment status nor salary had any impact. In conclusion, the expected coding times, if reliable, indicate that coders in the two hospital groups were keeping coding up-to-date. Thus, the variation between hospital groups in coding time is of little importance, given that the main objective in coding productivity is to maintain the coding workload.