Research Collaboration
HMO Research Network

Proceedings of the 3rd Annual Meeting
May 29-31, 1997
Boston, MA

Table of Contents

• Overview and summary
• Agenda
• Introduction: "Setting the Context"
• Plenary Session on Cost-Effectiveness Analysis Guidelines
• Breakout session reports
  • Health Outcomes in Diverse Populations
  • Geriatrics
  • Pharmacoepidemiology/Drug Policy Outcomes
  • Outcomes and Effectiveness Research
  • Chronic Care
  • STDs
  • Cancer
  • Asthma
  • Changes in Health Care Organizations: Implications for Provider Profiling
  • Provider Profiling
  • Implementing Research Results
  • Mental Health
  • Women's Health and Hormone Replacement Therapy
• Poster session presentations

Conference overview and summary

One hundred thirteen investigators and public health officials gathered in Boston on May 29-31, 1997 for the third annual conference of the HMO Research Network. Participants included investigators from the 12 member HMOs*, representatives from AHCPR, CDC, HCFA, and NIH, and some of their collaborators at other institutions. The conference was co-sponsored by the Agency for Health Care Policy, the Centers for Disease Control and Prevention, The Robert Wood Johnson Foundation, and the Harvard Pilgrim Health Care Foundation. Harvard Pilgrim Health Care was the host.

The major goal of the conference was to enhance the conduct of public domain research in HMOs by:

(a) assessing the methodologic and procedural requirements of such research, with emphasis on collaborations between HMOs,

(b) creating opportunities for HMO based investigators and their colleagues in public health agencies to engage in such collaborations,

(c) contributing to national research agendas and funding priorities,

(d) advancing the HMO Research Network's ability to respond to the national need for more efficient, effective, and high quality health care delivery systems.

The meeting included plenary sessions devoted to the perspectives of HMO delivery systems, of HMO research programs, and of public health agencies. There were also 13 breakout sessions devoted to development of research agendas in specific content areas.

Joseph Dorsey, Corporate Medical Director of Harvard Pilgrim Health Care, noted the role of HMO based research in the formulation of strategy and direction of HMOs, and the additional role of HMOs as research environments in which to address questions that are relevant throughout the entire medical care system. He noted several priorities for delivery systems, including improved risk adjustment methods, methods to detect underutilization of medically necessary services, studies of cost-effectiveness, and studies of the consequences of policies adopted by delivery systems.

Edward Wagner, Chair of the Network's executive committee and Director of the Center for Health Studies - Group Health Cooperative of Puget Sound, reviewed the activities of the Network and of several initiatives to enhance research partnerships. These include ongoing discussions between Network members and AHCPR, CDC, NIH, the American Association of Health Plans, and the HMO Group.

Andrew Nelson, Executive Director of HealthPartners Group Health Foundation (Minneapolis) summarized current research capabilities of HMOs. He noted that there are approximately two-dozen research programs based in, or closely affiliated with, managed care organizations. These programs involve over 150 career investigators, have extensive research infrastructures, and a collective annual research budget close to $100 million. Approximately 50% of their research is funded by federal sources, and 25% by the HMOs. (Nelson AF, et al. Research within managed care organizations: 1997 survey. Health Affairs 1998; 17:128-138.)

Mary Durham, Director of the Center for Health Research, Kaiser Permanente North West (Portland), discussed research partnerships among HMOs. She noted the specific circumstances and substantial sustained effort required to move from competition to cooperation, and finally, to collaboration. (Durham M.  Health Affairs 1998; 17:111-122.)

Representatives from AHCPR, CDC, NIH, and HCFA offered perspectives on the state of current joint research activities between their agencies and HMOs, and on the priorities for additional initiatives.

A plenary session addressed the implications for HMO based cost-effectiveness analyses of recently published US Public Health Service guidelines. The presenters made special note of the need to accommodate both societal and system perspectives.

Breakout sessions focused on agendas for HMO based research in specific content areas. Participants were asked to identify issues for which studies involving more than one HMO would be of value. HMO investigators and public health agency personnel led each session. The topics of the breakout sessions were: outcomes and effectiveness research, implementation of research results, geriatrics, provider profiling, sexually transmitted diseases, chronic care, pharmacoepidemiology and drug policy, mental health, immunization, cancer, health outcomes in diverse populations, asthma, and women's health.

Original research was presented in poster sessions. Thirty-four posters were presented.

Speakers and breakout session leaders provided the attached summaries.

Richard Platt
Director of Research
Harvard Pilgrim Health Care
Department of Ambulatory Care and Prevention
Harvard Medical School and Harvard Pilgrim Health Care
Conference chair

*The members of the HMO Research Network are:

Center for Health Studies - Group Health Cooperative of Puget Sound, Fallon Healthcare System, Harvard Pilgrim Health Care, HealthPartners - Group Health Foundation, Henry Ford Health System, Kaiser Permanente - Colorado, Kaiser Permanente - Georgia, Kaiser Permanente -Hawaii, Kaiser Permanente - Northern California, Kaiser Permanente - Northwest, Kaiser Permanente - Southern California, Prudential Center for Health Care Research

Back to top

Agenda

Friday, May 30

Morning

8:30-8:45 Welcome

Richard Platt, MD, MS
Director of Research, Harvard Pilgrim Health Care

Joseph L. Dorsey, MD
Medical Director, Harvard Pilgrim Health Care

8:45-9:00 Introduction

Edward Wagner, MD
Director, Center for Health Studies - Group Health Cooperative,
Puget Sound

9:00-9:30 The Network's Research and Research Capacity

Andrew Nelson, MPH
Executive Director, HealthPartners

9:30-10:00 Building Partnerships for Research With-in Managed Care

Mary Durham, PhD
Director, Center for Health Research, Kaiser
Permanente NW

10:30-11:30 PHS Guidelines for Cost-Effectiveness Analysis

Milton Weinstein, PhD
Joanna Siegel, SM, ScD
Harvard School of Public Health

Mark Hornbrook, PhD
Program Director, Center for Health Research, Kaiser
Permanente

Alex White, DDS, DrPH
Senior Investigator, Center for Health Research,
Kaiser Permanente

Tracy Lieu, MD, MPH
Physician Investigator, Kaiser Division of Research,
Oakland

State of the Network
Edward Wagner, MD, M.P.H.

Friday, May 30

Afternoon

12:45-2:30 Poster Sessions

2:30-4:30 Break-out Session #1

• Outcomes and Effectiveness Research
  Diana Petitti, MD
  Director, Research and Evaluation Clinical Services
  Southern California Kaiser Permanente
  
  Carolyn Clancy, MD
  Director, Center for Outcomes and Effectiveness
  Research, AHCPR

• Implementing Research Results
  Robert Fletcher, MD, MSc
  Director of Primary Care Clerkship, Harvard Medical
  School and Harvard Pilgrim Health Care
  
  David Atkins, MD
  Senior Health Policy Analyst, AHCPR

• Geriatrics
  Nancy Whitelaw, PhD
  Associate Director
  Henry Ford Health System, Center for Health Systems
  
  Terrie Wetle, PhD
  Deputy Director
  National Institute on Aging

• Provider Profiling and Models of Care, Staffing and
  Compensation
  Cindy Christiansen, PhD
  Senior Statistician, Harvard Medical School and
  Harvard Pilgrim Health Care
  
  Kelly Devers, PhD
  Expert, AHCPR

• STD
  Delia Scholes, PhD
  Scientific Investigator, Center for Health Studies - Group Health Cooperative, Puget Sound
  
  William Kassler, MD, MPH
  Chief, Health Services Research and Evaluation Branch;
  Division of Sexually Transmitted Diseases and Prevention,
  Centers for Disease Control

Saturday, May 31

Morning

7:45-8:15 Posters

8:15-10:30 Panels: Overview of Federal Agencies' Research Agendas Relevant to Managed Care

8:15-9:00 AHCPR

Irene Fraser, PhD
Director, Center for Organization and Delivery Studies

Carolyn Clancy, MD
David Atkins, MD
Kelly Devers, PhD
Barry Friedman, PhD
Expert

9:00-9:45 CDC

Suzanne Smith, MD, MPH
Chief, Branch Health and Aging Studies

William Kassler, MD, MPH

Roger Bernier, PhD
Associate Director for Science, National Immunization
Program

9:45-10:45 NIH

Edward Wagner, MPH (moderator)

Steven Hyman, MD
Director, National Institute of Mental Health

Michael Hilton, PhD
Director of Clinical and Prevention Research, NIAAA

Lawrence Deyton, MD
Acting Director, Division of Extramural Activities, NIAID

Robert Wittes, MD
Director, Division of Cancer Treatment and Diagnosis, NCI

10:30-10:45 Posters

10:45-12:45 Break-out Session #2

Chronic Care
Edward Wagner, MD

Barry Friedman, PhD

Pharmacoepidemiology and Drug Policy Outcomes
Jerry Gurwitz, MD
Executive Director, Meyers Primary Care Institute

Steve Soumerai, ScD
Director, Drug Policy Research Group, Harvard Medical School and Harvard Pilgrim Health Care

Dennis Ross-Degnan, ScD
Assistant Professor, Harvard Medical School and Harvard Pilgrim Health Care

Mental Health
Mary Durham, PhD

Kathryn Magruder, MD
*Title, National Institute of Mental Health

Immunization
Robert Davis, MD, MPH
Scientific Investigator, Center for Health Studies - Group Health Cooperative, Puget Sound

Roger Bernier, PhD

Cancer
Stephen Taplin, MD, MPH
Associate Director of Preventive Care Research, Center for Health Studies - Group Health Cooperative, Puget Sound

Brenda Edwards, PhD
Associate Director, Division of Cancer Prevention and Control, NCI

Mary McCabe, RN
Assistant Director, Division of Cancer Treatment and
Diagnosis, NCI

Saturday, May 31

Afternoon

Jeffrey Kang, MD
Chief Medical Officer, Office of Managed Care, HCFA

1:45-3:45 Break-out Session #3

• Health Outcomes in Diverse Populations
  Robert Hiatt, MD, PhD
  Associate Director, Kaiser Permanente, Oakland
  
  Irene Fraser, PhD


• Asthma
  Jonathan Finkelstein, MD, MPH
  Instructor, Harvard Medical School and Harvard Pilgrim Health Care

• Women's Health and HRT
  Maureen Connelly, MD
  Instructor, Harvard Medical School and Harvard Pilgrim Health Care
  
  Adele Franks, MD
  Associate Director for Science, Centers for Disease Control and Prevention
  
  Lynda Anderson, PhD
  Health Scientist, Centers for Disease Control

• Outcomes and Effectiveness Research (second session)
  Diana Petitti, MD
  
  Carolyn Clancy, MD

4:004:15 Committee Report

4:15-4:30 Board of Directors Report

4:30-5:00 Evaluation and Thank you

Back to top

HMO Research Conference

Setting the Context

May 30,1997
Boston, MA
Joseph L. Dorsey, MD
Corporate Medical Director
Harvard Pilgrim Health Care

WELCOME

Welcome to Boston. It is a real treat for me as Corporate Medical of Harvard Pilgrim Health Care to welcome you to this important conference. I have been with Harvard Pilgrim, and its predecessor Harvard Community Health Plan, since 1968. In my early learning years, a considerable part of my mentoring involved early leaders of your peer group. Sam Shapiro at HIP conducted the first major studies of the use of mammography and demonstrated reductions in mortality for screened women over the age of 50. Mitch Greenlick analyzed the experience of low income beneficiaries of Kaiser in Portland and Ray Fink studied the inter-relationships of mental health utilization with the overall utilization of health services in HIP.

These researchers were also closely integrated in the leadership of their early organizations. Their results were taken seriously by the leaders of those organizations as they formulated strategy and direction for the organization.

The Network's principal goals include "Advance research that is relevant to HMOs special features..."

We need to forge a coherent, tight link between the research priorities of the Network and "needs-to-know" of HMO's and their principal customers. Based on the distinctive characteristics, driving forces and goals of HMO's. I suggest we can discern a reasonable set of priorities:

1. Re: Distinctive Characteristics:
   1. As individual physicians, we serve our individual patients in a fiduciary relationship. Health plans must support and enhance those relationships while engaging physicians in the work of serving enrolled populations within a fixed budget, the amount of which is declining.
      
      For the researcher, these characteristics facilitate the calculation of rates of disease, rates of service use, rates of resource expenditures, etc. in our defined, enrolled populations, all of which are much harder to accomplish in an indemnity system in which no provider group can truly define the populations for which it is fiscally accountable.
   
   
   
   2. Because we provide, or arrange for, all the services covered in our benefit packages, the breadth and scope of the work of an HMO is far greater than in most practice settings.
      
      For the researcher, this breadth and scope of work provides access to data describing inter-relationships across inpatient and outpatient settings, acute and chronic diseases, tradeoffs of expenditures in one area for larger savings in another.
   
   
   
   3. Our emphasis on preventive services is long standing. Development of HEDIS measures, the key performance indicators by which we are judged,closely aligns us with the interests of the public health sector. We deliver in prepaid systems services which otherwise become the responsibility and the burden of public health agencies. We need to look together at how best to conduct outreach to special needs populations.
   
   
   
   4. Our need to be accountable requires the establishment of additional performance measures, especially related to outcomes and patient satisfaction, that go beyond current HEDIS measures, are not unduly costly to implement and are likely to produce actionable results.

    

2. Re: The driving forces and goals I see as Medical Director, and what I want from the Network Research enterprise.

   The single largest driving force unique to our systems is the use of a budget or a capitation to support the delivery of health care for a defined population. I need:




1. Techniques to characterize the relative risk of enrolled populations. HCFA'sbiggest problem in addressing regional differences in adjusted average per capita cost payments is the absence of a satisfactory risk adjustment method to define the at-time-of-enrollment morbidity of different populations.
   
   Health plans could reward providers in a more aligned way with better risk adjusted capitation formulas.
   
   Academic medical centers do tend to attract more high risk patients. We have some, but not enough, tools to tell how much of the higher expenditures levels associated with practice sites based at teaching hospitals are due to case mix, and how much to traditional teaching hospital practice inefficiencies, such as the use of the hospital emergency room for off hour coverage, the use of clinically very part time senior staff to supervise inpatient house staff and the resultant longer ALOS, and the disproportionate influence of consultant specialists over PCP's in the care of "bread and butter" type cases.



2. Better systems to detect under utilization of medically necessary services. We do pretty well for preventive services through our HEDIS measurements. We can use patient satisfaction surveys, monitoring of complaints, and alignment with highly ethical, high quality providers to help, but we need better ways to assure that as we drive out the "not at all beneficial" and "marginally beneficial", we do not encroach on the "medically necessary" in our quest to reduce costs.



3. We need the research community to move beyond "best quality" to "best value." We need you to study the clinical interventions that are of special importance to HMO programs, to calculate the costs associated with the interventions you study, and to make recommendations that are actionable by us. We need to understand better how to design and implement disease management programs, and how to evaluate new therapies.



4. We need you to anticipate where we are going and study what we do. The recent frenzy of legislation on so called "drive through" deliveries was not evidence based. In fact, an in house study by Dr. Troy Brennan of over 2,000 deliveries at Brigham and Women's Hospital showed that the 30 hour or less stay (mostly ours) women and infants had no outcome differences from the over 30 hour stays (mostly everybody's else's). Yet, the national and the Mass. legislature passed bills, basically forcing a minimum 2 day LOS for most women, creating an atmosphere in which it would be" politically incorrect" to push a reduced OB LOS program, even with a well tested program, with a secure safety net, and with satisfaction study results from mothers so cared for that were comparable to the longer stay mothers. The Mass. bill cost the 300,000 Health Centers Division of HPHC $7m in 1996 (on a budget of $550m).

Imagine how much better informed the public policy debate about valid reasons for authorizing emergency room visits would have been if we had data, instead of anecdotes, informing the public debate.

As we introduce further innovations in the paradigms of clinical practice to drive costs down and better serve our Medicare, Medicaid and employer partners, help us and the public be sure we step wisely. Sometimes, doing the right thing may have to await data that is more convincing for our members and for the general public than what we had when we launched the reduced ob length-of-stay program. Data might have helped define the set of enhanced services necessary to produce a satisfying experience for the mother, her obstetrician and pediatrician.

As we look forward to Jeff Goldsmith's 4th stage of evolution of the HMO marketplace, where costs have plateaued, the winners "will no longer be those driving costs lower, or improving processes, but those able to demonstrate the greatest ability to improve the health status indices of their enrolled populations." While we need your help in learning how to do that, we need to survive in the meantime. Building the nation's trust that we in the not-for-profit sector are committed to delivering best value for the dollar requires collaboration: among yourselves, with those of us who must lead the HMO delivery systems, with federal and state agencies responsible for the nation's health and its health care (HCFA, NIH, CDC, AHCPR and State Medicaid agencies) and with the public who are looking to us in managed care to establish systems that respect each individual patient, guard carefully the resources we receive to care for populations, and produce outcome and satisfaction results that justify the public's trust.

Back to top

Plenary Session on Cost-Effectiveness Analysis Guidelines

Mark Hornbrook and Alex White organized a session highlighting the recently published Guidelines for Cost-effectiveness Analysis in Health and Medicine. Sponsored by the Public Health Service, an international panel of experts in economics, medicine, and public policy derived "good practice" guidelines for conducting CEAs. The core concept underlying these guidelines is the "Social Reference Case," which counts all opportunity costs, no matter to whom they accrue, and measures effectiveness in terms of Quality Adjusted Life-Years Saved (QALYS). The goal is to achieve comparability among CEA studies produced by different investigators using different databases.

Milton Weinstein presented the keynote address, outlining the essential features of the CEA guidelines. Mark Hornbrook introduced the concept of the "HMO Reference Case" as a standard for guiding HMO decisionmakers. Tracy Lieu discussed the challenges and obstacles to conducting CEAs in the HMO context. Milt commented that HMOs should include societal costs in their CEAs, and that the HMO perspective could be presented as part of a sensitivity analysis.

Back to top

Health Outcomes in Diverse Populations

Drs. Robert Hiatt, Kaiser Permanente and Irene Fraser, AHCPR

Participants: 15

Measurement and Methodology

1. Statistical methods for noting risk adjusted differences in utilization
2. Race as a validation variable
3. Health status measurement
• How to do it
• How to use it
• Is it culturally/functionally biased
4. Validation of self-reports of utilization and self-categorization
5. Criteria for "minority" funding
6. How to get HMOs to participate
7. Cultural competence in study design

Diagnosis

1. Differences in cancer screening
2. ER use differences - understanding differences beyond "Blaming Victim"
3. Reported alcohol abuse
4. Link- to mental health outcomes
5. Birth outcomes

Interventions

1. Technologies- e.g. cultural tailoring, smoking cessation
2. Patient communities/patient needs
• Mismatch between providers and patient cultural competence/training
• Patient provider interactions
• Patient belief systems and health models

Back to top

Geriatrics

Drs. Nancy Whitelaw, Henry Ford Health System and Terrie Wetle, NIA

Participants: 15

Criteria for judging importance/feasibility of areas for research

1. Relevance to HMO financing
2. Relevance to definable population
3. Ability to have a large impact on quality of life and quality of care
4. Ability to have a large impact on cost
5. Amount of money required to do the study
6. Amount of time to required to do the study
7. Likelihood of stimulating coordination across multiple parts of the health system
8. Likelihood of funding/joint funding
9. Stimulates cross-site collaboration

Suggested areas for research

1. Which outcomes are important to measure and how do we measure them?
2. Preventive services
• which preventive services are effective
• factors associated with use and non use of preventive services
• how to disseminate benefits of prevention (e.g. exercise) to the elderly?
3. Does a common set of risk factors predict a variety of geriatric syndromes, or are there unique risk factors for each syndrome?
4. What social (non-health/non-medical) factors affect health outcomes ? What is the mechanism? What are the implications for health plans and delivery systems?
5. Adverse reactions to treatments and interventions
• who is susceptible/what is impact
• how to intervene
• impact of self help and education programs
6. Differences in outcomes of enhanced primary care for older adults versus case management
7. what are the variations of enhanced and care management approaches?
8. Geriatric continuum of care
• what would should be included in the geriatric continuum of care?
• how should services be linked?
• how to measure use/substitution of services?
• how can outcomes across the continuum be evaluated?
• how does home care fit into the continuum and its impact on outcomes?
9. Organizational and financing issues
• how do they pose barriers to innovation?
• is there variation and why in quality and outcomes under various financing and organizing systems?
• is there variation and why in physician effectiveness and satisfaction in caring for geriatric patients under various financing and organizing systems?
10. Patient/member perceptions of what constitutes high quality geriatric care?
11. What is medical necessity for care in managed care versus fee for service?
12. What are relevant (in terms of processes and outcomes of care) subpopulations within the geriatric population?
13. can we predict transitions across these population types?

Back to top

Pharmacoepidemiology/Drug Policy Outcomes

Drs. Jerry Gurwitz, Meyers Primary Care Institute, Stephen Soumerai, HPHC, Dennis Ross-Degnan, HPHC

Participants: 16

Outcomes

1. Post Marketing Surveillance
   • Epidemiology to drug use
   • Geriatric polypharmacy
   • Effectiveness backup studies
2. Underutilization In High-Risk Populations
• Depression
• Cardiovascular illness
3. Outcomes To "Noncompliance"
• Defining concept within existing data structures
• Clinical impact to "noncompliance" (nonadherence)
4. Expanding Pharmacy - Based Severity and Comorbidity Measures
5. Non-Delivery System Costs and Benefits to Pharmaceuticals
6. Effectiveness Studies
• Randomized studies
• Observational studies
• Natural Experiments
7. RCT to Interventions to Alter Prescribing

Quality

1. Intervention'.s To Coordinate Pharmaceutical Care
2. Structure To Pharmacy Practice On Use/Adherence
3. Interventions To Improve Prescribing
4. Impact to structure on responsiveness

Policy

1. Impacts To Formulary Changes (Longitudinal)
2. Impacts To Differences In Generic Coverage
• Different delivery systems
• Drugs with narrow therapeutic index (e.g. warfarin)
3. Shift To OTC Status
4. e.g. vaginal antifungal, H2-blockers, nicotine patch
5. Financial Incentives To MD's
6. Under vs. overuse
7. Changing Benefits For Patients
• OTC coverage
• Different benefit packages within organizations.
8. Unintended Impact Of Policy Changes
9. Influence To Industry On Formulary Decision-Making
10. Impact To Direct-To-Consumer Advertising
11. Understanding to markets, interactions with formularies

META - Points

1. Characterize Disease-Specific Studies
2. Need For National Scope
• Risk adjusters
• Formulary study
3. Prioritize Studies Which Do Not Require Linking All Pharmacy Data

Back to top

Outcomes and Effectiveness Research

Drs. Diana Petitti, Kaiser Permanente and Carolyn Clancy, AHCPR

Participants: 60

1. Systems interventions to address and implement multiple guidelines and to assess dissemination methods
2. Periodic preventive visits:
• resolve controversies
• how to be more selective
• addressing patient expectations
3. Variation between clinics independent of financing
4. Assess population health using DALY, QALY
5. Gaps from recommended outcomes. How costly is it to close these?
6. How do systems improve?
7. How can patient compliance be improved?
8. Describe role models/best practices to test hypotheses about why these are the best.
9. Management to encourage physicians to improve.
10. Areas where gaps should readily be closed include diabetes and asthma.
11. Mild and moderate mental health dysfunction: What are best methods for treatment and what outcomes can be expected?
12. Arthritis - treatment differences
• effects on outcomes not clear
• outcome assessment not sufficiently used

Back to top

Chronic Care

Drs. Edward Wagner, Group Health Cooperative and Barry Friedman, AHCPR

Participants: 33

• Much interest in implementing strategies across several sites to improve quality for chronic conditions.
• Much concern about appropriate use of specialty services and problems of coordination.
• Much concern that risk adjustment (and level thereof) does not reward good quality care. (This type of research left for another group).
• Concerns about the decision processes in plans target specific conditions for intensive effort.

Projects

1. Effectiveness of disease management strategies
• Improve communications
• Self-management
• Outcomes feedback
• Info systems
• A number of strategies whose effectiveness may vary with population and organization setting.
• All have some promise to improve quality in chronic care.
2. Enhanced primary care vs. carve in-out contracts for specialty care.
• Better quality from specialty care may cost less.
• However, loss of some coordination and comprehensiveness of care may be adverse to quality of care, particularly for people with multiple conditions.
3. How to implement changes to improve care?
• Incentives in compensation
• Office/system changes (e.g. computer-assisted ordering)
• Process improvement at practice level (education about guidance with feedback)
• Best methods of implementation may vary with many situational factors as well as market conditions.

Back to top

STDs

Drs. Delia Scholes, Group Health Cooperative and William Kassler, CDC

Participants: 9

Potential projects suggested:

1. Examination of recurrent pharmacy dispensing to assess Rx failure vs. re-infection vs. a chronic sub-acute process.
2. Use of managed care data to develop STD screening criteria for use in HMOs.
3. Organization vs. provider level approaches to STD screening.

• Opportunities for collaborative funding from foundations and non-federal sources as a way to maximize scarce resources.
• Opportunities to collaborate with health departments and other major players in STD prevention.
• CDC funding and opportunities as well as the CDC RFA just released. Animated discussion of possible responses to announcement.
• Consumer rights ( maternity ) any willing provider
• Carve-out/ proprietary disease management program
• National medical groups
• Provider teams
• Organization change
• Splitting inpatient/outpatient primary care hospital based internists
• Direct contracting w/purchasers
• Conversion of not for-profit ® profit

Back to top

Cancer

Drs. Stephen Taplin, Group Health Cooperative and Mary McCabe, NCI

Participants: 12

Brenda Edwards and Mary McCabe discussed three RFAs that had either been announced or were pending.

The most protracted discussions addressed multiple/integrative interventions. There was a general feeling that disease specific screening interventions that had been proven to be efficacious need to be combined into single interventions.

Other preventive functions also need integration and testing.

There was some frustration that these integrated interventions do not do well in a study section.

1. Differences in Care
• Rx as a function of delivery system type
• Colorectal cancer chemotherapy and surgery versus surgery alone
• Non-Hodgkins lymphoma staging practices based on tissue
• Survival as a function of:
• Delivery system
• Race
• Treatment plus age
• Follow-Up
2. Intervention Studies
• Integrative intervention trials
• Mammography plus cervical cancer plus colon Test effectiveness of combined intervention approaches
• MD oriented versus system oriented primary care based interventions (in-reach) versus systematic (i.e., outreach, data systems, primary care independent)
3. Patient priorities in care
• Risk perceptions and care priorities
• "Counter-detailing"
4. Survivor issues
• Quality of life
• Cost efficient follow-up protocols
• Organization - who follows When, how is communication achieved - primary care vs specialty
5. Prevention Studies
• Diet and cancer incidence
• Exercise and cancer incidence

Back to top

Asthma

Dr. Jonathan Finkelstein, HPHC

Participants: 7

The Asthma Research Task Group assembled at the 1997 HMO Research Network Meeting to: 

Create a prioritized research agenda that focuses on topics that are uniquely suited to conduct in HMOs.

• First, we engaged in a discussion of the attributes of research activities that would be uniquely suited for a network of HMOs. Certainly, almost any research that could be done in a single HMO could be done cooperatively in more than one. However, we focused on the attributes that tended to benefit from a multi-center, collaborative approach. These attributes included:
• Requirement for very large numbers of asthma patients either to detect rare events or to answer questions regarding small subgroups.
• Diversity of membership, including racial, ethnic and socioeconomic diversity, as well as geographic diversity itself
• Studies on practice or HMO variation in care.
• Studies which aim to create or sustain a systems level intervention.
• Studies which benefit from complementary talents of investigators at cooperating sites.
• The group identified seven domains of research activity, and discussed a number of enabling methodologic and infrastructure issues. These seven areas of focus were:

1. Epidemiologic and natural history studies. These studies would presumably use the population-based medical information that HMOs have to answer questions about the diagnosis and natural history of asthma. There seemed to be significant interest in utilization of care in the period preceding a diagnosis of asthma. Other types of studies include identification of risk factors for events of interest including hospitalization, ICU admission and death.
2. Creation of population models of the patterns of utilization, costs, and outcomes for asthma patients. These population models could be focused around specific demographic groups or by mode of treatment.
3. Observational systems research. This type of research is conceived to utilize the "natural experiments" existing in the HMO network. As organizations implement a wide variety of disease management approaches, there may be the opportunity to compare outcomes among them.
4. System level intervention trials. This type of study would design and implement system level interventions at participating HMOs. Since it is not possible for any HMO to "stay still" in asthma management, the types of studies that would be feasible might include implementation of specific aspects (e.g. intensive environmental control practices) in one system. The group observed that this would be easiest to implement for aspects of care that would not be part of "usual care" in that system but which the system might see as a worthwhile addition.
5. Medication effectiveness research (observational). This would utilize the data from participating HMOs to examine the outcomes of patients depending on medications and treatments prescribed.
6. Randomized trials of new therapies. This would take advantage of the large number of asthmatics in participating HMOs for enrollment in classic trials of the variety of new asthma medications that will be available.
7. Biological questions. This group of studies might include those on biological markers of asthma or impact of asthma on lung growth or other biological processes. CAMP is one example of such a trial. The group observed that as genetic markers of asthma are uncovered, there may be more interest within population-based organizations regarding these questions.

The four areas that the group fisted as highest priority were the epi/natural history research, observational intervention research, observational medication research, and population models research. All of these have in common that they are observational modes that utilize the extant data systems of the participating organizations.

Two issues of infrastructure or methodology were raised. The first was the issue of severity adjustment in observational work on asthma. While no single investigator or group wanted to take this on independently, the group felt as part of a multi-center initiative, some basic methodologic work on severity would be critical.

The other topic of discussion was the utility of a large joint registry for asthma patients across organizations in the HMO network. Kaiser Permanente is already developing a cross divisional registry of their asthmatics. They intend to supplement the automated clinical data with some survey data of their asthma population. The data coordinating center for this effort will be at the Center for Health Research in Portland. We discussed, at length, the potential benefits of having additional HMOs be a part of such a network. Benefits included more geographic diversity (especially on the Eastern seaboard), more plan diversity, and the existence of complementary investigator talent in various HMOs.

While all agreed that the idea of a registry for asthma was very powerful, the group observed that they are relatively expensive to start and require a clear-cut "present use" for them to be of interest for funding agencies. Given the significant shifts in asthma treatment, as well as the increasing prevalence, some members of the group thought that it might be a key time to begin a registry of large cohorts of patients in which future questions could be asked.

Since the assembled group was small, the suggestion was made that these ideas be sent to the larger group of individuals who expressed interest in the asthma session. With their input, we might further prioritize these various domains of research on asthma and its treatment. This prioritized list then might inform discussions with potential funders such as federal agencies.

Back to top

Changes in Health Care Organization: Implications for Provider Profiling

Kelly Devers, Ph.D., AHCPR and Cindy Christiansen, Ph.D., HPHC

Participants: 15

The purpose of this session was three-fold:

1. Identify major changes occurring in the organization and delivery of services. This may include things such as new models of care (e.g., multi-disciplinary teams, disease management, specialty carve-outs, etc.), related changes in staff mix and staffing patterns (e.g., use of nurse practitioners or physician’s assistants, educators, therapists, etc.), and other major organizational changes that affect the delivery of services.
2. Discuss the implications of these changes for provider profiling. These questions can be thought of as questions related to the specific measures and validity of profiling techniques, use of profiling (provider selection, feedback, compensation, etc.), and use of profiles by others (consumers, press, etc.). Some examples are:
• Unit analysis. Are we profiling the right entity (individual providers versus teams)?
• How do we "profile" specialty organizations whom we contract with and the interface between them and core providers (i.e., not only utilization but continuity of care, patient satisfaction, etc.)?
• Managed care plans sharing profiles on similar providers to examine both profiling techniques and behavior of same physicians in two different plans.
• How is profiling information used within the plan?
• Use of profiles by plans and outside entities? (e.g., consumers, purchasers, legislators, media)
3. What research projects are feasible on this subject given current activity in plans? Are there any areas where a collaborative effort is particularly useful and possible? What are the topics or research questions of most importance for the field generally?

The first hour was spent allowing each researcher to discuss development in their own plan and their individual research interests and the second hour was spent in discussion attempting to develop some consensus on research priorities and possible projects.

Back to top

Provider Profiling

Drs. Christiansen (HPHC and Devers (AHCPR)

Provider Profiling Models of Care, Staffing and Compensation

Priority Areas

• Does profiling have any effect on utilization
• Risk - adjusting
• Who should be profiled?
• team
• plan
• provider within team within plan
• On what should we profile?
• If profiling works, why?
• Qualitative
• As much variance as you can get ?
• Choose good measures for profiling across site comparisons
• Does M.D. profiling lead to improved patient outcomes and/or organizational outcomes.
• e.g. plan level (inpatient care)
• Profiling as part of a larger package
• Potential for collaboration
• Multiple
• Quasi-experiments

Implications for Profiling

• Risk adjusting profiles- static or á in some measures
• incoming health states
• *outcomes
• Who is accountable?
• e.g. nursing homes
• Statistical stability of profiles
• Distribution of underlying condition
• Can you measure the phenomenon?
• Use of profiling information?
• individual compensation
• public
• media?
• Should we profile to measure?
• à quartiles vs. measure
• à shifting distribution up,
• â reducing variance
• Communicating profiles to providers and members
• members' utility

Back to top

Implementing Research Results

Drs. Robert Fletcher, HPHC and David Atkins, AHCPR

Participants: 23

Most attendees were interested in efforts to change clinicians’ behaviors and only a minority were interested in conducting research comparing alternative methods of changing behavior. We tried to steer the discussion toward research itself but failed and in the end let most of the discussion concern interventions to change clinicians’ behaviors.

Potential Interventions
(in the order they were mentioned by members of the group)

1. Provide information
• Traditional CME
• Clinical practice guidelines
2. Feedback on performance
• Compliance with guidelines
• Comparison with peers (clinicians need to feel safe for this to be effective)
3. Change the work environment: create an environment in which desired practices are encouraged and facilitated (for example, by computerized decision support) and undesirable behaviors are discouraged (such as by standing orders, pharmacy restrictions)
4. Economic incentives/disincentives
• Applied directly on individual provider
• Acting on groups of providers
5. Persuasion
   • By opinion leaders (e.g., inherently respected, often senior, physicians)
   • By local clinicians who are trained to be experts for a specific condition ("peer leaders")
   • Applied face to face or to groups
   • Academic detailing
6. Changing patient demand
• Patients’ expectations drive clinicians’ behaviors
• Tying clinician payment to satisfaction reinforces the role of patient expectations
• Counter-detailing is an approach to inappropriate patient demands

Different interventions may be needed to get clinicians to:

• Do what they already accept they should do
• To get them to adopt a new practices based on new information
• To get clinicians to give up doing what they are already doing (the most difficult)

A distinction was made between initiating and maintaining change.

It was generally agreed that combinations of interventions (e.g., opinion leaders and feed-back) should work better than individual ones (but there is only limited research information on the effectiveness of multi-faceted approaches).

Guidelines are usually based on evidence of efficacy from clinical research whereas implementation in the "real world" where net effects could change - for example, because of less skilled clinicians, less cooperative patients, or a different balance of good and harm.

The most promising interventions were, in the opinion of the group, economic incentives/disincentives and changing the work environment whereas education alone was considered relatively weak. External guidelines are likely to require a process to increase local buy-in.

The concept of necessary and sufficient interventions was developed. For example, it was considered necessary for physicians to accept the value of changing behavior, and for this change to fit into current paradigms for care (e.g., PSA testing)

Diseases/Conditions that Might be Studied
General criteria: mature science base and gap between evidence and practice

Counseling on cigarette smoking cessation

Detection and management of alcohol abuse

Management of pediatric asthma

Management of frail elderly

Follow-up of acute myocardial infarction (drugs, exercise, return to work, etc.)

Identifying and screening in the subset of patients who are not being screened (outreach and inreach)

Counseling for patients with chronic disease

In addition, the group thought there was a need for descriptive studies across organizations of how decisions regarding clinical policies are made in managed care, including surveillance on new information, official sanction, adoption in practice, and barriers to changing practice.

Other issues
Efforts to change clinicians’ behaviors should be prioritized so that clinicians are not overwhelmed.

How will priorities for implementation be set?

Cost saving information

Information that can improve outcomes but at an increased cost

HEDIS measures

Do managed care organizations have the information base for deciding which practices need attention? For knowing if an intervention has changed behavior?

A major challenge in managed care outside the staff model HMO: how to reach providers who are geographically dispersed and for whom only part of their practice is with the HMO.

Back to top

Mental Health

Drs. Mary Durham, Kaiser Permanente and Kathryn Magruder, NIMH

Participants: 21

How can we prevent mental illness and what are effectiveness prevention strategies?

Treatment Matching

1. The issue of treatment matching:
   • What patient characteristics influence the course of depression and the effectiveness of treatment?
   • What kinds of patients are appropriate for what kinds of treatment?
   • What kinds of treatment providers?
   • What department the treatment should be delivered in?
2. Treatment Guidelines/Diverse population
• Do guidelines have to vary according to culture of patients and provider groups?
• How do we implement guidelines?
• How do we maintain improvements in provider practice?
3. What do we do for patients who don't get better when given guideline-based treatment?
4. What are the most effective treatment models?
5. How can we bring more consistency to the treatment offered and measures of outcome?
• I.D. best measures
• Calibrate measures to each other
• I.D. best implementation strategies, training materials, etc.
6. What are the next important outcomes and how should we measure them?
7. How do we determine where to put resources
• Value of screening
• Value of finding outcomes, baseline and information directly to clinicians.
8. How do we increase patient adherence to treatment, regimens be it medication compliance, staying in psychotherapy, etc?
9. What kind of social support or relationship interventions improve treatment outcomes? Family violence/Health Plan School partnership in dealing with kids.
10. If we offer mental health care to high utilizers, can we decrease inappropriate medical utilization, and service maintain patient health and satisfaction?
11. How can we deliver population based cared? What are the elements of successful population based care, i.e. registries.

Depression

1. How can we improve care for depression in primary care
• setting
• physical training
• enhanced assessments
• use ancillary personal
• appropriate training
• bringing mental health providers into department of primary care:
1. who should they be?
2. what training is required?
3. what specifically should they do?
2. How can we evaluate treatment outcomes in primary care settings?
3. What treatments are most effective for dual diagnosis patients?
4. Specific treatment matching- who are the best candidates for psychopharm alone? psychotherapy alone, with and without additional psychotherapeutic type interventions

Back to top

Women's Health and Hormone Replacement Therapy

Adele Franks, M.D., CDC, Lynda Anderson, Ph.D.,CDC and Maureen Connelly, M.D.,HPHC

Participants: 11

Research Topics of Interest: (in order of interest and priority)

Top Priority

1. Provider survey: Instrument already developed and utilized at Group Health Collaborative in collaboration with CDC. Designed to measure provider attitudes about menopause care, practice variation and HRT prescribing patterns. General interest in replicating study in other managed care organizations.  Identified as top priority because the subject matter is important and not well studied, the instrument already exists, implementation would be feasible across a number of health plans, cost of a mailed survey would be limited, analysis plans and programs already available, and CDC funding potentially available. Action: CDC and GHC staff to further review possibilities for funding and contact attendees.

Secondary Priority

1. Provider education/support: GHC staff described efforts to develop provider support and education. Participants expressed interest in attempting to replicate the approach used by GHC by studying their instruments.
2. Decision Support for Patients: GHC and HPHC staff described trials in progress to assist women with menopausal and HRT decision-making. Interventions include interactive videos, audiotape booklets, workbooks, counseling. Attendees expressed interest in also studying the effectiveness of these approaches.
3. Documentation of HRT Prevalence and Utilization patterns: Utilize existing automated databases to document the prevalence and patterns of HRT use. HPHC staff reviewed comparative study ongoing with the Lovelace Health Systems. Kaiser Permanente, GA staff specifically identified this as a primary collaborative research interest. Others identified this area as a secondary area of potential collaboration.
4. Alternative Therapies for Menopause: Attendees expressed interest in trials of alternative therapies, member surveys of the prevalence of such therapies, and identification of alternative therapy use by non-utilizers of hormone replacement therapy.
5. Symptom Prevalence: The group acknowledges the lack of data documenting symptom prevalence and the antecedent factors that might contribute the degree of symptoms. Attendees also noted the methodological limitations of measuring symptoms. The group agreed that collaborative studies in this area might provide adequate power to identify relevant associations.
6. Patient Attitudes and Barriers to Therapy: Attendees expressed interest in including in any patient survey project, questions about attitudes toward menopause and HRT as well as an assessment of barriers to therapy.
7. Special Populations: Attendees specifically identified the study of non-majority racial/ethnic groups, those with relevant co-morbid illnesses (i.e., DM, osteoporosis) and patients with low socio-economic status as a priority. Collaborative efforts, again, could generate larger sample sizes from which to generate and test hypotheses.

Additional Projects Suggested but Not Included in Final Priority List:

1. Development of on-line guidelines for menopause care, osteoporosis, and other women's health topics.
2. Studies of the efficacy of specific preparations and regimens of HRT.
3. Strategies for managing postmenopausal endometrial bleeding.
4. Assessment of endometrial cancer risk.
5. Prevalence of hysterectomy rates.

Back to top

Poster Presentations

Alcoholism treatment and utilization of health services
Mary Anne Armstrong, MA; Lorraine T. Midanik, PhD; Alan W. Lazere, MSW; Arthur L. Klatsky, MD

Barrier specific telephone counseling of women overdue for a mammogram in two HMO populations: Predictors of intended and actual mammography utilization after counseling
R. Luckmann,; L. Clemow; J. Savageau; M. Mitchell; B. Haddad; M. Costanza.

Benefits and costs of screening dyslipidemic hypertensive HMO members for diabetes mellitus
Patrick J. O'Connor, MD, MPH; William A. Rush, PhD; Linda M. Cherney, MPH; Nicolaas P.Pronk, PhD

Breast cancer: Differentiation of risk
Christine Cole Johnson; Maria J. Worsham; Usha B. Raju; David S. Nathanson; Sandra Wolman

Decisions about hormone replacement therapy (HRT) among HMO women in the 1990's
Katherine M. Newton; Andrea Z. LaCroix; Nancy F. Woods; Nora L. Keenan; Lynda A. Anderson

Depo-medroxyprogesterone acetate exposure and bone mineral density in young women
D. Scholes; A. Z. LaCroix; L. Ichikawa; S. R. Ott

Developing and validating a dietary screening tool for use in primary care, or The PrimeScreen Project
Matthew Gillman; Sheryl Rifas; Jamie Kotch; Rebecca Hannigan; Cindy Christiansen; Xiaoming Gao; Walter Willett

Differences in the receipt of secondary prophylaxis for myocardial infarction between patients with HMO vs fee-for service insurance
D. McCormick; J. Savageau; J. Yarzebski; J. M. Gore; J. H. Gurwitz; R. J. Goldberg

Drinking and related problems in a managed care population
Enid M. Hunkeler, MA

Ethics and the doctor-patient relationship in managed care: A survey study of patients in an evolving HMO
S. D. Pearson; J. E. Sabin; D. Rusinak

Evaluating hypertension control in a large HMO
Mark Alexander; Irene Tekawa; Bruce Fireman; Enid Hunkeler; Joe Selby; Barry Massie;Warren Cooper

Gender Differences in product use for cardiovascular conditions
Sharon J. Rolnick, PhD, MPH; Patrick J. O'Connor, MD, MPH; Agnes Tan, PhD; William Rush,PhD

Impact and implementation of an active management of labor guideline in an HMO
Sharon J. Rolnick, PhD, MPH; Barbara Hyer, MD; Jody Jackson, BA; Linda Loes, MD

Impact of a hypertension treatment guideline on blood pressure control in primary care offices
Patrick J. O'Connor, MD, NTH; Elaine S. Quitter, RD; William A. Rush, PhD; Seongryeol Ryu,PhD

Implementing a new model of care for managing depression in the primary care setting of an HMO: Physician training, nurse telephone follow-up, peer support
Enid M. Hunkeler, MA; Joel Meresman, PhD; Jennifer M. Groebe, LCSW

Incidence of fractures in men and women fifty-plus years of age
Catherine Wisner, PhD; Tom Abbott, PhD; Lucy Fischer, PhD; Marc Berger, MD

Is continuity of diabetes care related to intensity of care and glycemic control?
Patrick J. O'Connor, MD, MPH; Jay Desai, MPH, William A. Rush, PhD; Linda M. Cherney,MPH; Don Bishop, PhD

Maternal substance use and its relation to intensive care nursery utilization in a managed care organization
Mary Anne Armstrong, MA; Gabriel J. Escobar, MD; Marc Usatin, MD; Leslie Lieberman,MSW; Bruce F. Folck

New frontiers in community care of children with chronic illness
Andrew Nelson, MPH; Barbara Staub, MD; Brooks Donald, MD; Ann Kelly, MD

Older women and the risk of fractures
Lucy Rose Fischer, PhD; Catherine L. Wisner, PhD; Thomas Abbott, PhD; Gerald Amundson,BA; Tom VonSternberg, MD

Parenting the preterm infant in the first two years of life
Maryjoan D. Ladden, PhD, RN, CS

Patient noncompliance in the managed care setting: The case of medical therapy for glaucoma
J. H. Gurwitz; S. M. Yeomans; R. J. Glynn; B. E. Lewis; R. Levin; J. Avorn

Performance failure of an upper respiratory infection (URI) clinical guidelines
Patrick O'Connor, MD, MPH; Gerald Amundson, BS; Jon Christianson, PhD

Potential savings from improved diabetes care in Minnesota and the U.S.
Patrick J. O'Connor, MD, MPH; Todd P. Gilmer; William A. Rush, PhD; Jay Desai; Donald Bishop, PhD

Predicting hospital length of stay among pre-term and low birth weight infants admitted to the NICU: Experience in a managed care organization
Gabriel J. Escobar, MD; Mary Anne Armstrong, MA; Bruce F. Folck; Marla N. Gardner; Allen Fischer, MD

Rates of hospitalization during antihyperlipidemic drug use and after discontinuation of therapy at two health maintenance organizations
Susan E. Andrade, ScD; Richard Platt, MD, MS; Gordon Saperia, MD; Marc Berger, MD

Reasons why patients started on lipid therapy fail to reach lipid goals
Patrick J. O'Connor, MD, MPH; William A. Rush, PhD; Richard J. Gray, MD

Shifting physician prescribing to a preferred histamine-2 receptor antagonist: effects of multi-facotrial intervention in a mixed-model HMO
Jill Weiskopf Brufsky, Dennis Ross-Degnan, David Calabrese, Xiaoming Gao, Stephen Soumerai

Term and preterm infant and family characteristics and their impact on family health care organizations
Maryjoan D. Ladden, PhD, RN, CS

The association between copayment and delay in seeking emergency care for patients with myocardial infarction
David Magid, MD, MPH; Thomas D. Koepsell, MD, MPH; Nathan Every, MD, MPH; Jenny S.Martin, RN; David S. Siscovick, MD, MPH; Edward H. Wagner, MD, MPH; W. DouglasWeaver, MD

The potential impact of improved glycemic control on the cost of care for patients with diabetes
Todd P. Gilmer, BA; Patrick J. O'Connor, MD, MPH; Willard G. Manning, PhD; William A.Rush, PhD

To be or not to be on HRT in the 1990's: The roles of perceived risk, health history, and attitudes about menopause and HRT
Andrea Z. LaCroix; Katherine M. Newton; Nancy F. Woods; Nora L. Keenan; Lynda A. Anderson

Using information on HMO patients to understand asthma-related health care utilization by African Americans
Barbara Tilley; Edward Zoratti; Christine Joseph; Marvella Ford

Why do physicians prescribe antidepressants for geriatric patients?
Lucy Rose Fischer, PhD; Judith Garrard, PhD; Sharon J. Rolnick, PhD, MPH; Jody Jackson, BA;Lori Luepke, BS

Back to top