Vermont moves toward providing plethora of benchmarking information

Initiative makes 19 data collection recommendations

Imagine a one-stop benchmarking shop for consumers, insurers, and employers who want to check out health care providers and plans in your state. It could be reality sooner than you think.

In Vermont last year, such a Herculean effort was launched by the Department of Banking, Insurance, Securities and Health Care Administration in Montpelier, a state agency charged with licensing managed care organizations and health insurers. The agency convened an 18-member panel representing payers, providers, consumers, employers, and government interests to study what health care information should be provided to the public as well as how it should be collected, says Steve Kappel, director of data analysis.

"One of our highest priorities is to inform consumers. Another priority is to improve quality and contain costs of health care," he says. "We believe both goals can be met by providing good, accurate information."

The challenge, he points out, is that each audience for health information - consumers, employers, providers, health plans, policy-makers, legislators, and regulators - has a distinct set of needs. "Everybody wants information, but with a little different slant. Collecting, reporting, and analyzing data is expensive and time-consuming, so one of our main tasks was to identify a core information set we could agree upon."

Despite the difficulties, the attempt to bring comparative health care information together is a growing trend across the country. There are plenty of good reasons your organization may want to get involved, but primarily you'll want to protect your interests when there are decisions being made about standardization of measures and data-gathering techniques.

The Vermont project is typical in this respect. Among the group's goals are several aimed at standardization:

· Make use of existing validated measures.

· Identify most useful measures for each constituency group.

· Identify measures that can be produced using existing systems.

· Clarify definitions to make measures directly comparable.

The first three times the group met, members received background information on the processes and challenges of collecting, analyzing, and presenting health care data. "One of most important tasks early on was to make sure we all had a shared vocabulary. We wanted them to have the whole picture as a common starting point," Kappel explains.

The group also reviewed current measurement systems such as Health Employer Data and Info rmation Set (HEDIS), the Joint Commission's ORYX program, and the Commission on Accredi tation of Rehabilitation Facilitates (CARF).

The Vermont group's experience is a good case study in what you can expect to encounter in your state. In the course of his team's work, Kappel says several common themes arose:

· Administrative data don't tell all.

"Data systems are best at what they are designed to do. For example, administrative databases are designed to track claims payment. Because they are focused on reimbursement, they are limited in usefulness of gauging actual quality," he explains.

· Quality data tell more - but you'll pay highly for them.

Data derived through manual abstraction of a chart or survey are much more useful than administrative data, but the collection cost may be prohibitive, he says. "It's a trade-off between the cost of the perfect data world and less expensive - and less useful - data. We knew we wouldn't be able to get to that perfect world, but we wanted to get as close as we could."

The working group also agreed that because hospitals lack standardized quality and satisfaction measures, comparisons are difficult. More over, current reporting systems often are weak in measuring the information consumers want. "They want an indication of the 'hassle factor' - how difficult was it to negotiate discharge, for example," he says.

Finally, the group agreed that because health care performance measurement is a rapidly evolving science, it would be advantageous to adhere to national standards where they're available.

In the next four sessions, the working group outlined 19 recommendations in three categories: measures that should be collected immediately, those that needed more refinement, and those that should be collected in the future.

1. Immediate data collection. The group recommended that the state should begin collecting from all managed care organizations HEDIS measures relating to access and availability of care, effectiveness of care, and use of services. Some of those measures would be available from administrative databases, and others would require chart-abstracted or survey data. Some would need to be extracted for the managed care companies' operation within the state; other measures would be planwide.

"We will work with the MCOs operating in multiple states to break out their Vermont data separately. We want to assure that sampling methodologies reflect our state's experience as well as make sure the sampling is consistent among plans," Kappel says. "If measures are not gathered through the same method, we will not be able to use them for comparisons."

2. Measures needing refinement. By the end of 1999, MCOs will need to use a common consumer survey tool, the group recommended. "We recognize that discussions about a com -mon survey instrument are going on with the National Committee on Quality Assurance and the Health Care Financing Administration. But if, by 1999, they have not reached an agreement, we will require a standardized reporting tool," Kappel says.

After a review of various tools, the group's preferred survey instrument is the Consumer Assessment of Health Plans (CAHPS) to assess the following:

· ease of using administrative services;

· courtesy and respect by health plan staff, physicians, and other health care professionals;

· adequately consistent information about plan procedures, benefits, and services;

· coordination and continuity of care;

· cost, including premium, deductible, coinsurance, copayments, and balance billing;

· overall satisfaction with care received;

· enough time with physicians and other health care professionals;

· interpersonal and communication skills of physicians and other health care professionals;

· timely access to preventive services.

To measure consumer satisfaction with hospitals, the team recommended the state should use existing tools developed by JCAHO and/or CARF.

3. Measures for the future. One of the recommendations in this section was that the state should continue to address the problems of collecting consumer satisfaction data about MCOs, particularly in those areas not addressed with CHAPS. "There just isn't a `gold standard' for satisfaction," he says. Such a standard, he adds, would include basic questions such as:

· Based on your experience with submitting claims, billing, scheduling, or customer service within the last six month, would you recommend this plan to your family members or friends?

· Based on your experience with health care services provided by doctors and other health care professionals within the last six months, would you recommend this plan to family members and friend?

Finally, the team recommended that the state should find out how satisfied providers are with MCOs and health insurers. "The information collected should concentrate on the administrative and clinical management processes that affect the quality of care," he explains. Areas to be measured include the following:

1. Professional autonomy and practice climate. This measure should include indicators showing how autonomous primary care providers are in coordinating and managing utilization as well as the autonomy of referral specialists in devising and managing specialty treatment plans, says Kappel. Continuity and communication between primary care providers and referral specialist also should be measured, he adds.

2. Financial factors. Kappel says measures should determine if financial arrangements between providers and plans influence clinical decision making by examining the following:

· plan-to-plan variations in reimbursement rates;

· volume-based incentives;

· disincentives to making specialty referrals, recommending treatment, or ordering tests.

3. Community standards. "The measures also should determine whether care management guidelines and processes vary significantly among health plans. This indicated different expectations of quality of care," Kappel says. Processes to measure include referral, utilization management, forms, paperwork, and information requirements.

4. Management tool. Do administrative controls over medical management decisions improve or adversely affect the quality of care? The answer can be found by measuring credibility and clinician agreement on practice protocols, adequacy of decision support tools, the utilization management process, patient access to therapeutic drugs of choice and medical supplies as well as required paperwork and documentation.

5. Organizational resources. "We'll look at the clinical expertise of medical directors and consultants as well as the availability of medical technology and equipment," he says.

6. Denial and appeals process. Measures will include the length and complexity of the process. The "arbitrary and capricious use" of multiple practice guidelines that result in denials also should be measured, the team recommended.

As the Vermont example clearly demonstrates, state efforts to provide consumers and employers with comparative data on health care providers and plans could have significant impact on your data collection and reporting operations. If your state has such a project under way, you owe it to yourself and your organization to get involved. At the very least, you will stay up to date on its actions and possibly have some input into what directions those actions take.

For more details, contact Steve Kappel, director of data analysis, Department of Banking, Insurance, Securities and Health Care Administration, 89 Main St., Drawer 20, Montpelier, VT 05620-3601. Telephone: (802) 828-2900.