Risk Adjustment

You are currently browsing articles tagged Risk Adjustment.

Measuring Patient Case Mix in Medicare

March 5, 2012 in Medicaid/Medicare, Medicare Advantage | No comments

How does Medicare measure patient case mix?  For the most part, Medicare uses the Hierarchical Condition Category (HCC) model.  A recent CMS presentation describes the HCC model in more detail.  Today I review where CMS applies the HCC model, provide an overview of the HCC methodology, briefly describe its performance, and give some background on how the HCC model was developed.

 

Applications

Medicare uses the HCC model to risk adjust spending in the following applications:

  • Medicare Advantage Capitation Payment (Implemented in 2004, fully phased-in 2007)
  • Shared Savings Program Accountable Care Organizations (To be implemented in 2012)
  • Medicare Physician Quality and Resource Use Reports (Implemented in 2009)
  • Hospital Quality Measurement for the Medicare Spending per Beneficiary (MSPB) measure. –(Implemented in 2012).

HCC Methodology

CMS-HCC model classifies all conditions but not all conditions used in payment/other applications. Most disease groups are high cost medical condition (cancer, heart disease, hip fracture). Conditions can be excluded because they do not predict future cost (e.g., appendicitis) or there is a High degree of discretion or variability in diagnosis, diagnostic coding, or treatment (e.g., symptoms, osteoarthritis). These conditions are generated from diagnosis codes on claims. Diagnosis codes from lab, radiology and home health claims are not used because they are not reliable and may indicate rule-out diagnoses. The number of times a diagnosis is recorded does not affect the model’s assignment of beneficiaries to health states.

The HCC algorithm starts with over 14,000 ICD-9-CM codes which are grouped into 805 diagnostic groups and then aggregated to 189 condition categories (CCs). From the CC’s, CMS creates 70 hierarchical condition categories where hierarchies imposed. For instance, Angina pectoris/ old myocardial infarction is not included in the acute myocardial infarction HCC (#81) but the CC for AMI is included.

The HCC model also includes demographic factors:

  • 24 age-sex cells (e.g., male age 80-84);
  • Medicaid dual eligible status;
  • current disability status,
  • original Medicare entitlement status

There are three separate HCC models used for the Medicare Advantage program: community, institutional, and new enrollee.

The HCC model is also used to adjust payments for beneficiaries with end-stage renal disease (ESRD), all of whom are enrolled in Medicare FFS. There are three HCC models for the ESRD population: dialysis, transplant, and functioning graft.

Physician QRUR uses age-disabled, community, new enrollee and ESRD models.

HVBP uses a single model with indicators for whether the beneficiary has ESRD or is in long term care

Performance

The model can only moderately predict cost. The R-squared is about 12%. This should not be surprising as variation in health care cost over time can be highly variable.

Development and Maintenance

The model originally developed under contract to CMS by researchers at Boston University and Research Triangle Institute (RTI) with clinical input from Harvard Medical School physicians and is currently maintained by RTI.  The model is updated every year to incorporate new diagnosis codes and is recalibrated regularly on more recent diagnosis and expenditure data.

 

Tags: , , , , ,

Risk Scores and provider size

June 16, 2011 in P4P | 1 comment

At a recent AcademyHealth presentation, Cheryl Damberg discussed her research to design a P4P program for implementation for Integrated Healthcare Association (IHA).  One thing I noticed about the presentation was that smaller provider groups had patients with lower risk scores (i.e., healthier patients).  Is it really the case that small providers treat much healthier patients?

My guess is the answer is not.  An alternative explanation would be that small providers do not have as much time or administrative staff to help them code the patient’s comorbidities in their claims (or even EMR).  If this is the case, it would make it appear that small provider’s patients are healthier when in fact the true differences may be due to differences in the quality of the data the providers report.

Any VBP system would need to take into account these differences when evaluating providers.  Setting a lower standard for small providers, however, would provide a disincentive for small providers to expand into the large provider category, even if this expansion could (potentially) create economies of scale and improve patient care.

Tags: , ,

Problems with Risk Adjustment

November 30, 2010 in Econometrics | 1 comment

To evaluate providers based on the health outcomes or the cost of care, one must attempt to evaluate dimensions of care which are strictly within the providers control. For instance, if a physicians treats two patients with breast cancer, but one patient has a more advanced form of breast cancer, one should take this difference into account. Patient comorbidities also affect the prognosis for a successful recovery from illness, as well.

One method to take into account the patient’s health conditions upon presentation at a provider’s facility is to use risk adjustment methods. Risk adjustment methods take into account factors such as patient demographics (e.g., age, gender), health status (e.g., prior diagnosis, current illness severity), prior utilizations (e.g., previous hospitalizations) and other factors to predict the expected outcome for a typical patient. Risk adjustment, however, is never perfect. A paper by Garber, MaCurdy and McClellen (1998) review some of the problems with using risk adjustment in the health care setting.

Read the rest of this entry »

Tags:

Risk Adjustment: Predicting Future Expenditures

November 2, 2010 in Econometrics, Medicaid | No comments

Many states rely on managed care organizations (MCOs) to provide medical services for their Medicaid beneficiaries.  Contracting out medical services to private providers relies on the government’s capacity to accurately predict expected cost of care for each beneficiary.  This is typically done through risk-adjusted capitation rates.

Which risk adjustment strategy works best?  The answer of course depends on the context.  A paper by Yu and Dick (2010) examines 5 predictors specifications to predict future expenditures for Medicaid eligible children.  I list each of the five specifications and their performance (measured as the R2) below:

  • Age/Gender only: 0.2%
  • Age/Gender + subjective health status measure: 3.9%
  • Age/Gender + CSHCN: 7.3%
  • Age/Gender + HCC: 12.1%
  • Age/Gender + prior year expenditure: 43.5%

One can clearly see that the best predictor of a child’s current year expenditures is the child’s prior year’s expenditures.

Read the rest of this entry »

Tags:

Risk Adjustment Models

September 9, 2010 in Econometrics | 1 comment

Risk adjustment is important for many aspects of health care.  Medicare uses risk adjustment to modify payments to Medicare Advantage (Part C) plans based on the health of the beneficiaries they cover.  Private insurance companies can use risk adjustment to fine-tune capitation payments to physicians or determine a potential enrollee’s premium.  Providers can use risk adjustment to identify likely high cost patients and how to adjust their likely treatment pattern accordingly.

There is little doubt that risk adjustment is important, but determining which risk adjustment model is ideal is difficult.  A paper by the Society of Actuaries (2007) examines this topic.  The ideal risk adjustment method will depend on a number of factors.   These include:

  • Ease of use of the software;
  • Specificity of the model to the population to which it is being applied;
  • Cost of the software;
  • Transparency of the mechanics and results of the model;
  • Access to data of sufficient quality;
  • Underlying logic or perspective of a model that makes it best for a specific application;
  • Whether the model provides both useful clinical as well as financial information;
  • Whether the model will be used mostly for payment to providers and plans or for underwriting, rating and/or case management;
  • Reliability of the model across settings, over time or with imperfect data (models that are calibrated and tested on a single data set and population may or may not perform well on different data sets/populations);
  • Whether the model is currently in use in the market or organization; and
  • Susceptibility of the model to gaming or upcoding.

Read the rest of this entry »

Tags:

Risk Adjustment Tools for Measuring Physician Quality

November 12, 2009 in Information, Physician Compensation, Quality | No comments

A paper by Thomas, Grazier, and Ward (2004) analyzes a variety of risk adjustment software products. Using these six risk adjustment products to calculate physician efficiency scores, they found “moderate to high levels of agreement were observed among the six risk-adjusted measures of practice efficiency.” However:

“And even though our analyses suggest that 50 percent to 60 percent of adult PCPs identified by their system as being high outliers are likely to be identified by other profiling systems as well, the client has no way to know which of the identified outliers are the ones that multiple systems would agree on. Thus the profiling client must deal with practice efficiency rankings knowing that, in all likelihood, 40 percent to 50 percent of PCPs identified as high outliers are actually not among the least efficient 10 percent of primary care physicians.”

The authors also compare two quality score metrics. The first is the ratio of the physician’s observed cost with the expected cost based on the physician’s patient’s risk scores. The O/E score is equal to:

  • (O/E)k=yk/Yk

Above, yk is physician k‘s observed score and Yk is their estimated score. The authors believe, however that the O/E score is not ideal. It is biased against providers who have a small sample size of patients. Thus, physician’s with smaller patient panels in the data set are more likely to be considered outliers. On the other hand, the authors advocate using a standardized cost difference (SCD). The SCD is calculated as follows:

  • SCDk=(yk-Yk)/[σ/(Nk)1/2]

The SCD measure explicitly takes into account the physician’s sample size. A large sample size will move the SCD more towards the difference in observed and expected costs; a small sample size will move the SCD score closer to the mean of 0.

Below is a list of the six risk adjustment tools used in the paper:

  • Adjusted Clinical Groups from Johns Hopkins University. Adjusted clinical groups cluster health plan members having similar comorbidities into groups that have similar resource requirements and clinical characteristics. The ACG Case-Mix System then uses a branching algorithm to place each patient into one of 82 discrete, mutually exclusive categories based on the mix of clinical groups experienced during the time period under study.
  • Burden of Illness Score from MEDecision, Inc. This system is based on MEDecision’s Practice Review System (PRS), which partitions care into episodes of illness and assigns services, severity levels, and medications to these episodes. The BOI Score is a linear-scaled measure that indicates relative health care cost risks associated with the particular mix of episodes experienced by a patient during a defined time period.
  • Clinical Complexity Index from Solucient, Inc. The CCI methodology considers age, severity, comorbidity, hospital admissions, and categories of diagnoses (acute, chronic, mental health, and pregnancy) to assign patients into mutually exclusive CCI risk categories. Although the system provides for 1,418 different categories, 95 percent of patients fall into just 45 of these.
  • Diagnostic Cost Groups from DxCG, Inc. The DCGsystem includes a whole family of multiple linear regression models.
  • Episode Risk Groups from Symmetry Health Systems,Inc. Like BOI Score, ERGs are episode-based. The episodes underlying ERGs are created using Symmetry’s Episode Treatment Groups (ETGt) methodology, a basic illness classification system that uses a series of clinical and statistical algorithms to combine related services into more than 600 mutually exclusive and exhaustive categories. For a given patient, episodes experienced during a time period are mapped into 119 Episode Risk Groups, and then a risk score is determined based on age, gender, and mix of ERGs. For our analyses, we used the ERG retrospective risk score.
  • General Diagnostic Groups from Allegiance LLC. General Diagnostic Groups were developed using the Agency for Health Care Policy and Research’s Clinical Classification Software (CCS). CCS aggregates individual ICD-9-CM codes identified on health care claims into 260 broad diagnosis categories for statistical analysis and reporting. The GDG system then lumps together CCS categories considered to be clinically similar and to have similar associated per-patient charges into 57 diagnostic categories. These 57 diagnostic categories are used as dummy variables in a multiple regression model for predicting health care costs.

Source:

Tags: , ,