NQF Wants Risk-Adjusted Outcomes, Costs, Payments

The evidence supporting differences in health outcomes between high-income and low-income populations is "indisputable," says Steven H. Lipstein, president and CEO of BJC Healthcare. Using non-risk adjusted outcomes to calculate performance measures disproportionately punishes providers who care for disadvantaged populations, he says.

A National Quality Forum panel wants the federal government to take into account socio-demographic factors for patient population health outcomes when risk adjusting pay-for-performance performance measures.

Steven H. Lipstein, president and CEO of St. Louis-based BJC Healthcare, was a member of the NQF panel that issued the report in October. Lipstein says numerous studies have shown that patients from less-affluent communities have poorer health outcomes, even though the care they receive in the hospital is the same as people from wealthier communities. As a result, he says, hospitals that serve predominantly poorer communities will be unfairly penalized.

Lipstein spoke with HealthLeaders Media about the panel's report. The following is an edited transcript.

HLM: Why does socio-economic status affect health outcomes?

Lipstein: In St Louis, mainly [in] the neighborhoods north of the city, you go into communities that are characterized by low household income, high poverty rates, high housing unit vacancy rates, high unemployment rates, and a high percentage of the population that didn't complete high school.

When you look at those five indicators you are looking at a community that is lacking social and economic infrastructure. When you live in a community without grocery stores, drugs stores, without transportation, with high crime rates, the health outcomes in that neighborhood are much worse than you would experience in neighborhoods that are characterized by a more-affluent population.

The evidence supporting the difference in health outcomes is indisputable. What is challenging right now is that the healthcare policy makers, specifically CMS (the Centers for Medicare & Medicaid Services), believe that health systems, doctors, hospitals, nurses, and other healthcare professionals, can overcome those disparities and produce the same outcomes at the same cost as compared with more affluent communities.

That is just not true. There is no evidence to support that finding.

HLM: CMS raises concerns that adjusting for socio-demographics will create tiered care standards that will adversely affect the poor. Is that a legitimate concern?

Lipstein: What people are concerned about is that if you risk-adjust for socio-demographic status, you will somehow cover up poor quality provided to patients of lower socio-economic status.

Nobody is advocating for covering up anything. In fact, after we do the risk adjustments, you will see in the recommendations that were proposed by the expert panels that the National Quality Forum convened, the performance measures would include specifications for stratification of the clinically adjusted version of the measure based on the socio-economic factors used in the risk adjustment.

That means you would actually display the results of all hospitals' performances according to different socio-demographic variables. You would see if there were better outcomes for poorer patients than for more-affluent patients. All of that would be transparent and visible so there would be no obscuring examples of poor quality.

Socio-demographic risk adjustment, if you're going to use those measures solely for quality assessment and improvement, then it is important to highlight differences among socio-demographic groups. But when you are using these measures as part of a pay-for-performance system, as CMS is doing, what you end up doing is punishing disproportionately the providers who care for disadvantaged populations.

You punish them in one of two ways: either their outcomes are not as good as hospitals that serve patients from more-affluent communities, or if their outcomes are the same and it just costs more to produce the same outcomes for a vulnerable patient as it does for an affluent patient, then CMS calls the provider who produced that same outcome a low-value provider because the quality is the same, but the cost was higher and CMS equates that to lower value.

HLM: Doesn't CMS take socioeconomic status into consideration when it designates critical access, low-volume and safety net hospitals?

Lipstein: They are not. The federal government looks at safety net hospitals as if they are all the same, and they are not. There are safety net hospitals in some jurisdictions that are able and willing to tax themselves to support safety net care.

If you were to look at the tax base that supports Denver Health, or Cook County Medical Center or Grady Memorial Hospital, or Parkland in Dallas, and then look at the taxing jurisdictions that support rural communities or the boot heel of Missouri, or East Baltimore, or North Philadelphia, you would find the local jurisdictions are very different.

There are 51 different state Medicaid programs. That means eligibility for Medicaid is variable across the United States. When CMS says we are going to look at dual-eligibility as an indicator of a person's poverty or life circumstances, think about that.

A childless adult in Missouri is not eligible for Medicaid, but in 27 other states that childless adult is eligible up to 138% of federal poverty, or an income of $32,000 a year for a family of four. So, the individual life circumstances of people who are ineligible for Medicaid are highly variable across the United States.

HLM: What do you want?

Lipstein: We want CMS and all payers, in their pay-for-performance programs, to risk-adjust health outcomes, to risk-adjust costs to produce those outcomes, and to risk-adjust premiums payments if we got to that for the socio-demographic characteristics of the community in which the patient lives; not the community where the hospital or the patient resides, but the community where the patient lives.

The risk-adjustment methodology that we use at BJC Healthcare looks at the socio-demographic characteristics of individual census tracts. Ascribing them to each individual patient, we can then calculate a discharge-weighted poverty rate for each hospital in our state. That means it's the poverty rate not of the hospital or where it resides, but the poverty rate of where the patients live who use that hospital.

In our report, hospitals with the highest discharge-weighted poverty rates by census tracts are the same hospitals with the highest all-cause readmission rates in our system. In order not to penalize doctors and hospitals under pay-for-performance of value-based purchasing or risk-based contracting, you need to adjust their outcomes, costs, and premiums to account for the degree of difficulty associated with serving vulnerable populations. By doing that we will encourage hospitals, doctors, nurses, healthcare professionals to invest future capital in serving disadvantaged patients.

Right now what I am concerned about is that CMS is presiding over a pay-for-performance environment in which they are rewarding the provider community for investing its future capital in serving more-affluent populations. The market already provides the rewards for serving more-affluent populations. CMS doesn't have to help them do that.

Right now, the fact that there are health systems in Detroit, Michigan paying the maximum penalty for readmission rates, while hospitals in Scottsdale, Arizona pay no penalty for readmission rates, is a re-direction of federal resources from a poorer community to a more affluent community, and I don't understand why our government thinks that is good public policy

HLM: Is a price tag attached to this risk adjustment?

Lipstein: It would be neutral. The penalties could still be assessed but they would begin to fall on different providers. The federal government could save the same amount of money but it would end up assessing penalties on hospitals in more-affluent communities at a higher rate than it does today.

HLM: Does your proposal create the tiered standards of care that CMS is concerned about?

Lipstein: If CMS were correct, then we may have hospitals in our health system or hospitals nationwide that are not paying penalties for readmission rates because they've served patients who are discharged into home environments or communities that can support their recovery and convalesce more easily than they could in more-disadvantaged communities.

Why would CMS disproportionately favor those hospitals even though they are doing no better at managing the experience of care when the patients are in the hospital or the transition of care to the patient in the home environment?

The key point is we are trying to be fair to providers who serve vulnerable patients and we are trying to encourage them to continue to provide services to vulnerable patients or even expand those services in the future. But if they are disproportionately affected by the federal government's pay-for-performance and value-based purchasing program, that will be a contributing factor to their financial challenges.

If you lose your ability to meet your financial challenges by shifting costs to more-affluent patient populations, you go out of business.

The number of hospitals that have failed in the St Louis-metro area—and we presented almost 40 years of data—are all hospitals that lost the ability to address their financial challenges and they disproportionately served vulnerable communities.