Application of the Insight Diagnosis Related Grouper (Part 1)

A Diagnosis Related Group (DRG) is a tool with which to classify hospital admissions into clinically intuitive and statistically homogenous categories.

Photo by Markus Winkler | Unsplashed

A Diagnosis Related Group (DRG) is a tool with which to classify hospital admissions into clinically intuitive and statistically homogenous categories. Clinical intuitiveness means that categorisations can be easily understood and interpreted by someone with (limited) clinical knowledge. Statistical homogeneity means that all the admissions within a particular category are expected to be associated with similar levels of resource utilisation and similar costs. The development and maintenance of the Insight DRG algorithm will itself be detailed in future blogs.

When correctly deployed by healthcare funders and providers, a DRG contributes to the development of a more sustainable healthcare system which delivers more cost effective and better-quality care. Amongst its many applications, DRGs allow for the scientific quantification of the mix and severity of hospital admissions. This is commonly referred to as case mix.

An understanding of case mix and the ability to compensate for variations in case mix is a prerequisite for the preparation of meaningful cost and quality benchmarking assessments. Such analyses are needed to create a more efficient healthcare system. Benchmarking assessments highlight more efficient ways to deliver care which can me mimicked by those delivering less efficient care.

 

Scenario Example

Imagine benchmarking the cost per admission at a particular hospital relative to that of its peers without being able to quantify and offset variations in case mix. This engagement between the funders and providers would look something like the interaction below and yield little if any value.

  • Healthcare funder:

    “Your average cost per admission is higher than that of your peers. Remedial actions are required to address your cost efficiencies”.

  • Healthcare provider:

    “A high cost per admission does not equate to cost inefficiencies. This is a function of our hospital admitting patients for major cardiac procedures. Such complex procedures are invariably associated with higher-average costs per admission relative to a benchmark which includes less complex procedures such as tonsillectomies. In turn, the analysis is flawed and cannot be considered further”.

The application of a DRG does not alone guarantee that the benchmarking assessment is robust. This is in the same way that the use of high-quality equipment in a manufacturing plant does not guarantee that the manufactured output is of the requisite standard. A robust manufacturing process requires appropriate raw materials, high-quality equipment and skilled machine operators. A robust benchmarking assessment requires the same.

The data which underpins the measurement is the raw material. The DRG used to facilitate the case mix adjustment necessary to make fair comparisons is the equipment. The actuary responsible for preparing the analysis is the skilled machine operator. If the raw materials are flawed or the machine operators are insufficiently skilled, the benchmarking exercise will not be sufficiently sound. This is irrespective of the application of the DRG.

 

 

For example, a sufficiently robust DRG-based cost efficiency assessment requires (at least) the following data validations and technical and operational processes.

 

  • Outliers
    Hospital admissions with unusually high or low costs even after case mix adjustment are referred to as outliers. Outliers are typically the result of inadequate clinical coding or extreme complications. Outliers have the potential to unduly distort cost efficiency assessments and should be removed. Interquartile statistical trimming is commonly used for this purpose.

 

  • Suitable DRGs
    Not all DRGs are well suited to benchmarking based on the cost per admission. Some admission types are simply too unpredictable[1]. DRGs should be delineated into three categories based on statistical homogeneity.DRGs deemed to be sufficiently homogenous based on the cost per admission should be benchmarked with reference on the cost per admission. DRGs deemed to be sufficiently homogenous based on the cost per day but not the cost per admission should be benchmarked with reference to the cost per day. The remaining (less homogenous) DRGs should be excluded from cost benchmarking exercises or ring-fenced with the appropriate caveats.Homogeneity may be assessed with reference to volumes, standard deviations and coefficients of variance. Whilst the Insight DRG allows for most admissions to be robustly benchmarked with reference to the cost per admission, the importance of benchmarking based on the cost per day and the exclusion of particularly volatile admission types cannot be overstated.

 

  • Case Mixes
    DRGs, no matter how robust, can never fully explain all variations in case mix. Where appropriate, structural supply-side factors should be accounted for. Care must be taken to ensure that supply-side factors which explain cost variations, but which do not justify these variations are not incorporated into the case-mix adjustment process.An example of a structural supply-side factor which should be accounted for is the distinction between general acute hospitals and specialised mental healthcare facilities. It is widely accepted that specialised mental healthcare facilities attract more severely ill mental healthcare patients than general acute hospitals even within the same DRG.An example of a structural supply-side factor which (typically) should not be accounted for is an undersupply of basic and inexpensive but clinically adequate equipment coupled with an oversupply of advanced but not clinically superior equipment. This could explain but not justify higher costs at a particular facility.

 

  • Like-with-like comparisons
    At certain facilities, hospitals deliver extraordinary services which are not offered by their peers. This artificially inflates hospital costs whilst deflating hospital-related costs. For example, some hospitals provide and bill for the consumables used in cataract procedures whilst at other hospital ophthalmologists provide and bill for these consumables. Hospitals which provide and bill for the consumables will be associated with inflated hospital costs and deflated hospital hospital-related costs. Carve-ins and carve-outs are needed to ensure like-with-like comparisons.

 

  • Sharing Information
    The extent of the information which is shared between healthcare funders and providers is arguably as important as the assessment itself. Inadequate information sharing inhibits remedial actions and culminates in inaction and frustration.

 

This can best be illustrated by way of an example.


Some funders highlight to hospitals that the anaesthetist costs in their facilities are higher than expected whilst failing to provide further information. Hospitals are blind to anaesthetist costs and can only have somewhat superficial engagements with practitioners without being empowered with more information. At a minimum, reporting should allow for an understanding of whether cost differentials stem from higher prices or variations in practice.

Conversely, a minority of hospitals are too quick to demand access to excessively detailed and sensitive information which cannot be reasonably shared by a funder before acting on cost variations. The need for never-ending and often immaterial technical adjustments similarly inhibits strides towards a more sustainable healthcare system.

Healthcare funders and providers should ideally strive towards a mutually agreeable reporting cost benchmarking framework. The framework should detail the scope and frequency of reporting. Rather than being a source of friction between health funders and providers, DRGs are being successfully employed by many healthcare funders and providers to foster a more collaborative and sustainable way forward. Adversarial discussions on tariff levels are being superseded by deliberations more favourable tariff adjustments can be funded by tackling the cost variations highlighted in DRG-based cost benchmarking reports.

Insight has been fortunate to participate in the structuring of such arrangements and has observed first-hand the tremendous benefits which they can bring both to healthcare funders and providers. The extent of the information which is shared between healthcare funders and providers is arguably as important as the assessment itself. Inadequate information sharing inhibits remedial actions and culminates in inaction and frustration.

Further insights into these and other DRG-related considerations will be detailed in an upcoming series of articles. For now, we will reiterate our belief that DRGs have tremendous applications for healthcare funders and healthcare practitioners and these applications can contribute to the realisation of a more efficient and sustainable healthcare system.

 


 

[1] For example, the Poisoning or Toxic effects of Drugs and other Substances with Major Complications and Co-morbidities DRG is typically associated with a coefficient of variance of over 3.0 with reference to the cost per admission and a coefficient of variance of over 1.0 with reference to the cost per day. Such a DRG is poorly suited to cost-benchmarking.

1

0 Comments

Get an email whenever we publish a new thought piece

Topic:

By signing up you consent to our terms and conditions

More from Craig Getz

Email

Subscribe

Email

Subscribe