Title: Can Methodologists Fix the Dysfunctional “Market” For Health Care Services?
Author:
Section/SPIG: Medical Care
Issue Date:
Both advocates and opponents of a “Single Payer” system for health care financing in the United States often fail to appreciate that whether or not we decide to rationally and equitably collect the money to pay for health care in this country, there is still an important second question: how do we get the best possible value when spending our health care dollars?
While it is fashionable to malign health maintenance organizations (HMOs) for stinting on care, any method of paying affects how care is given. For example, fee-for-service payments reward doctors more the more services they deliver – whether needed or not – and doing too much can be just as harmful as stinting. Another problem with fee-for-service medicine is that it provides little incentive for preventive care and no mechanism for “creative” non-medical solutions to medical problems (such as, avoiding “bad-air-day hospitalizations” for people with congestive heart failure by paying for their air conditioning). In theory, the HMO concept, pioneered as a visionary social innovation, is a particularly appropriate structure for facilitating coordinated care for people with multiple health problems and maintaining the health of populations. Also in theory, competition among HMOs should create better value for both those who pay for and those who use their services.
However, when health care purchasers pay HMOs amounts that are only adjusted for the age and sex of their enrollees, plans make the most money by discouraging enrollment from (and encouraging disenrollment of) sick people. Without payments that reflect vast differences in health needs, HMOs that “cherry-pick” the healthiest populations make unearned profits, while plans that efficiently care for sick populations may face bankruptcy.
In the early 1980s, the Medicare program began contracting with HMOs. The federal government should be commended for using research grants to incubate a new health care information industry in predictive modeling. Starting with these grants, several research teams, including my own, developed methods to enable health-based payments to HMOs. These methods are now widely used by government groups in the United States and abroad, by private insurers, and by provider groups to establish fair prices for delivering “a year’s worth of health care” to a specific group of people. Prominent intellectual offspring of this initiative include Diagnostic Cost Groups (DCGs, <www.dxcg.com>), Adjusted Clinical Groups (ACGs, <www.acg.jhsph.edu>) and the Chronic Disability Payment System (CDPS, <www.medicine.ucsd.edu/fpm/cdps>). Fair payment is calculated, one person at a time, based on the expected cost next year of the medical problems for which that person has been seen during the current year. Because an HMO’s payment is the sum of the expected costs for its enrollees, the HMO loses a lot of money when a medically needy person disenrolls, and only a little when a healthy person does. The goal is to create an “efficient market” for health care delivery that protects sick people, their insurers and their providers, and leads to better care at lower cost.
Starting in the year 2000, the Center for Medicare and Medicaid Services (CMS) began, in a limited way, using formulas developed by my colleagues and myself (based on the DCG methodology cited above) to pay HMOs that enroll Medicare beneficiaries. Starting Jan. 1, 2004, Medicare is using these models to make 30 percent of its payments “health-based.” By 2007, all Medicare HMO payments will be health-based.
Health-based payments correct some important perverse incentives in previous payment methods for HMOs. Predictive modeling won’t save the world, but – irrespective of how the money used to pay for all this is collected – it is a crucial tool for identifying and rewarding better health care providers and delivery systems.