Patient Risk Scores Are the Key to Value-Based Payment

The U.S. healthcare system is steadily transitioning from fee-for-service (FFS) reimbursement to fee-for-value (FFV) payment. This change has already started to affect medical practice revenue, and it will have an even bigger impact in the years ahead.

Unfortunately, most physicians and practice managers understand only part of the FFV equation. Under FFV, they know the quality data they report to payers will affect their reimbursement, but many do not understand exactly how payers use these data to adjust payment.

What is the missing piece of this equation? Patient risk scoring.

Under many value-based payment models, payers adjust reimbursement to reflect the relative health or sickness of patients. These adjustments are meant to reflect expected costs, so they can have a big effect on payment. Depending on what risk factors are present, appropriate risk scoring can double or triple per-patient reimbursement.

The challenge is that patient risk scoring is complex. It is easy for medical practices to under-report risk and, therefore, to miss out on full reimbursement. There are some crucial challenges they must confront to understand and properly utilize patient risk scoring.

The Nuts and Bolts

Physicians and practice managers must first understand the nuts and bolts of patient risk scores.

The overall goal of value-based payment is to reward physicians who provide high-quality care. However, because of differences in patients and patient populations, physicians may see wide variations in outcomes and costs regardless of the quality of the care they provide. If value-based payment is to be fair, there must be a way to account for variation in patient risk.

The solution is risk adjustment—using statistical modeling to convert a patient’s individual health risk factors into an overall patient risk score.

Payer risk models are technically complex. The most commonly used system is the Hierarchical Condition Categories (HCC) risk-adjustment model. The Centers for Medicare & Medicaid Services (CMS) introduced the HCC in 2004 to adjust capitated payments for beneficiaries enrolled in Medicare Advantage plans.

The HCC model incorporates 79 diagnostic categories covering high-cost chronic diseases and some acute conditions. Specific HCC categories include:

• Diabetes without complication (HCC19)
• Diabetes with chronic complications (HCC18)
• Morbid obesity (HCC22)
• Rheumatoid arthritis and inflammatory connective tissue disease (HCC40)
• Drug/alcohol dependence (HCC55)
• Major depressive, bipolar, and paranoid disorders (HCC58)
• Congestive heart failure (HCC85)
• Acute myocardial infarction (HCC86)
• Dialysis status (HCC134)

The HCC model also includes patient demographic information (age and gender) and patient Medicaid status. Within the model, each HCC and demographic category is assigned an individual risk factor. The sum of each patient’s individual risk factors is his or her total risk score. This patient risk score is also known as the Risk Adjustment Factor (RAF).

In general, the RAF is low for young, healthy patients and high for senior patients and those who have chronic comorbidities.

Risk Scores Determine Payment

Medicare Advantage determines plan payments by multiplying the base capitated rate by the patient’s RAF. For example, if the base rate is $9,000 and a beneficiary’s RAF is 1.450, the risk-adjusted capitated payment for that patient will be $13,050. The HCC model also applies to beneficiaries enrolled in state Affordable Care Act (ACA) marketplaces. Many healthcare finance experts believe that most payers will continue to adopt patient risk scoring in some form in the years ahead.

In a risk-based contract, plans calculate and may pay physicians based on a similar methodology. However, physician and practice managers should be aware of two important points:

• First, the HCC is based primarily on claims data. Claims are typically associated with a face-to-face encounter with a medical professional or care management services. In addition, documentation within the medical record must support all diagnoses.
• Second, the entire system is prospective. Demographic and diagnostic data reported in one year are used to establish the patient’s RAF for the following year. The corollary is that diagnostic data are not carried over from year to year. On Jan. 1, each patient’s RAF is reset to zero. Patients are considered free of any disease conditions until the provider submits new claims for payment with diagnostic data.

Be a Data Creator

One of the major problems for medical groups and healthcare systems is that the entire system of value-based payment depends on accurate reporting of diagnostic and demographic data, and physicians are not trained to be “data creators.” There is often a big gap between the data a physician captures and the data required to generate an accurate patient risk score. The result is lower-than-appropriate payment for physicians who provide complex care.

FFV payment models provide higher reimbursement for care that physicians render to complex patients. It’s a great opportunity for doctors who treat large numbers of senior patients and patients with chronic conditions. Unfortunately, common documentation problems cause many physicians to receive less risk-based reimbursement than they deserve.

Most physicians want to spend as little time as possible on documentation. However, documentation is a requirement, and information technology systems and automation can be used to aid the requirement. In the coming years, the financial health of your practice will increasingly depend on accurate diagnostic documentation and coding.

The first step in addressing this issue is to understand how your practice may be under-reporting risk. The following are five common documentation problems that suppress patient risk scores:

1. Not documenting the highest specificity level. Patient age and gender are important factors in risk modeling, but patient diagnoses can have a dramatic effect on patient risk scores. For example, a patient with major depression might have nearly twice the risk score of a patient without this disorder.The problem is that physicians are used to documenting procedures using Current Procedural Terminology (CPT) codes. Most are not accustomed to documenting diagnoses and are not trained in using International Classification of Diseases (ICD)-10 codes to the highest level of specificity. As a result, they fail to capture diagnostic data—and the data they do report fail to reflect the complexity of their patients.For instance, a physician may document a patient as “diabetic without complications” when, in fact, the patient also suffers from kidney disease. This will incorrectly force the patient into a low-risk category, resulting in lowered reimbursement.
2. Poor EHR Template Design. Most providers have been trained to use an electronic health record (EHR). However, many have not received sufficient education on setting up EHR templates correctly and using them to document encounters accurately and comprehensively.In one common scenario, physicians customize their EHR templates in such a way that the templates limit accurate documentation. We worked with one physician who customized a template to document up to four chronic conditions. As a result, it was impossible to fully document any patient with five or more chronic diseases.
3. Poor EHR setup. When the government introduced the Medicare EHR Incentive Program, many practices quickly took advantage of available dollars to purchase EHR software. In many cases, however, their implementation efforts were flawed.Typically, practice leadership focused on capturing financial data but did not give adequate thought to the long-term capture of discrete clinical data. For example, many practices have not configured their EHRs to capture lab data with Logical Observation Identifiers Names and Codes (LOINC). As a result, the system lacks a powerful mechanism for identifying and documenting patient conditions.
4. Reluctance to document certain diagnoses. Physicians sometimes hesitate to document diagnoses for nonclinical reasons. For example, a physician may fail to document “morbid obesity” because he or she does not want to stigmatize the patient—or initiate an uncomfortable conversation.Alternatively, physicians may be reluctant to document conditions that will become a permanent part of the patient’s medical record. Under the Affordable Care Act, pre-existing conditions are not a concern. But, due to uncertainty surrounding healthcare reform, some worry that documenting a condition like high cholesterol today could one day lead to a loss of coverage.
5. Failure to recapture ongoing diagnoses. Many risk-based payment models are prospective. For example, under Medicare Advantage, diagnostic data reported in one year are used to establish the patient’s risk score for the following year. In addition, as noted above, diagnostic data do not carry over from year to year. On Jan. 1, every patient is considered 100 percent healthy until proved otherwise through documented diagnoses.

Unfortunately, many physicians are not aware that their risk-based patients are considered “healthy until proved sick.” As a result, they fail to recapture diagnoses by documenting ongoing chronic conditions annually and reporting them on claims. In addition, significant medical history—for example, a mastectomy performed five years ago—may be “lost” if it is not re-reported to payers.

The Next Step

All of these scenarios produce an artificially low patient risk score, leading to lowered reimbursements. The solution is to improve documentation of patient data, especially patient diagnoses.

A careful review of recent claims will uncover potential instances of under-documentation and under-coding. This will typically allow a practice to recoup additional payments.

For example, Medicare Advantage allows providers a 12-month prospective period to review coding accuracy and submit corrected claims. For a practice with $4 million in annual claims, a review will typically uncover $250,000 to $1 million in additional payments. Just as important, by improving documentation and data management, physicians can ensure accurate patient risk scores and payments going forward.

Medical practices that implement effective documentation processes and optimize their use of technology will significantly improve their performance under risk-based payment. In addition, capturing diagnoses accurately and fully will help leverage data to improve patient care.

Gene Rondenet is the president of qrcAnalytics.