August 2008

Grant Results

SUMMARY

Researchers at the Center for Information Technology Leadership created an analytical model to assess how using information technology (IT) to support diabetes management systems would potentially affect costs, care processes, physiological measures and clinical outcomes.

Key Findings

  • All forms of IT-enabled diabetes management improve processes of care, improve physiological measures, prevent development of diabetic complications and generate cost-of-care savings.
  • Technologies used by providers seem to be the most effective in improving the lives of patients with diabetes, for example:
    • Diabetes registries — which providers use to track patients and their diabetes-specific information — appear to be the most effective of all IT-enabled diabetes management systems:
      • Diabetes registries were also the only form of IT-enabled diabetes management that showed a net cost savings after implementation. Scaled to the national level, over a period of 10 years, registries would save $14.5 billion in cost of care. With implementation costs of $6.16 billion, overall net cost savings would be $8.34 billion.
  • Integrated provider-patient systems, which add patient-centered technologies to a registry and reminder system, would add benefits beyond a registry alone:
    • Although implementation costs were highest for integrated systems, the potential for improved care processes, physiological measures and clinical outcomes far outweighed the other forms of IT-enabled diabetes management.

Funding
The Robert Wood Johnson Foundation (RWJF) supported this unsolicited project, from December 2003 through November 2005, through a grant of $350,005 to Partners HealthCare System, the parent company of the Center for Information Technology Leadership.

 See Grant Detail & Contact Information
 Back to the Table of Contents


THE PROBLEM

In 2005, diabetes was the fifth-leading cause of death by disease in the United States, according to a Centers for Disease Control and Prevention (CDC) report (available online). Direct and indirect costs in 2002 totaled more than $132 billion, and by 2010 the number of patients with diabetes was expected to rise by 20 percent, with associated costs of $156 billion, according to the American Diabetes Association (Diabetes Care, 26(3): 917–932, 2003. Available online).

Despite advances in the treatment of the disease, patients with diabetes often fail to receive the recommended care. A 2004 survey revealed that physicians complied with diabetic guidelines less than half the time (McGlynn EA et al., Rand Health, WR-174-1, 2006. Available online).

Research suggests that diabetes management programs — which provide multidisciplinary evidence-based care and give patients the education and tools needed to help manage their disease — require an information technology (IT) backbone in order to be effective.

For example, the Health Care Delivery Work Group of the National Institutes of Health's Behavioral Research and Diabetes Conference concluded in 1999 that in order for a diabetes management program to be successful, "it is necessary to have a clinical information system" to support it (Glasgow RE et al., Diabetes Care, 24: 124–130, 2001. Available online).

However, research assessing the cost benefit and cost effectiveness of IT-enabled diabetes management has been limited.

 Back to the Table of Contents


RWJF STRATEGY

RWJF's 2003 Annual Report noted that "Americans are living longer, but with chronic conditions — such as diabetes, hypertension, heart disease and asthma — that require ongoing, high-quality health care." Further, "[t]he Foundation addresses the quality dilemma on several fronts. RWJF is joining with providers, purchasers and others to establish standards for measuring health care quality, develop a framework for reporting those measures publicly, and create incentives for improving care."

When this grant was issued in 2003, it was under the auspices of RWJF's Quality team. The team sought to reduce the uncertainty surrounding the value of information technology in supporting disease management approaches.

The strategic use of IT was widely considered to be an essential ingredient in providing high-quality chronic illness care in the outpatient setting. One reason for the limited adoption of information technology in disease management was a lack of information about potential returns on investment. By synthesizing the evidence, providing a clear framework of how to understand IT functions that support disease management and showing the clinical, organizational, and financial value associated with these IT functions, it was hoped that the project would help provider groups and other health care organizations make informed decisions about IT investments. These better-informed decisions were expected to lead to wider and wiser adoption of IT to support disease management.

 Back to the Table of Contents


THE PROJECT

Researchers at the Center for Information Technology Leadership created an analytical model to assess how using information technology (IT) to support diabetes management systems would potentially affect costs, care processes, physiological measures and clinical outcomes.

The researchers chose to focus on diabetes management after an examination of the research literature made it clear that there were insufficient data to create an analytical model that could cover the use of IT in the management of multiple diseases.

Categories of Diabetes Management Technologies

With a panel of nationally recognized experts in IT-enabled disease management (see Appendix 1), the project team examined four broad categories of diabetes management technologies:

  • Technologies used by providers, including:
    • Diabetes registries, which track patients and their diabetes-specific information, and compare diabetes status with guidelines to display point-of-care reminders.
    • Clinical decision support systems, which compare patient information from electronic medical records against a set of rules to generate alerts about potential errors and present treatment options to providers. Unlike disease-specific registries, the clinical decision support systems maintain and use comprehensive health information about patients.
  • Technologies used by payers, including:
    • Systems that interface with electronic claims systems to track and monitor diabetes-specific information, compare patient data with recommended guidelines, identify opportunities for improved management and provide feedback to patients and providers by telephone, email or postal mail.
  • Technologies used by patients, including:
    • Patient self-management technologies, which use educational resources and data-gathering systems to improve patients' management of their own care between provider visits.
    • Remote monitoring technologies, which transmit clinical data from patients' homes to providers' offices so that providers can modify care plans between visits.
  • Integrated diabetes management systems, which would provide a suite of technologies to support a full range of diabetes management activities. These systems would:
    • Allow providers to use registries to make effective decisions.
    • Empower patients through self-management and remote monitoring technologies to participate actively in those decisions.

    Although the project team could find no published articles that show the effect of such a system, it included this category to demonstrate its potential.

Development of the Model

To develop the model, the project team constructed five software modules, or engines, to create estimates in four areas:

  • The Implementation Cost Engine estimated the expenses involved in implementing and operating each form of IT-enabled diabetes management. The project team relied primarily on market research for these data, conducting phone interviews with 38 organizations currently implementing or selling IT-enabled diabetes management technologies.
  • The Intervention Impacts Engine estimated the impact of IT-enabled diabetes management on physiological measures and processes of care. The project team used a systematic review of academic literature, trade publications and the general press to inform its estimates.
  • The Disease Burden Engine estimated the effect that process changes have on diabetes-related conditions, including coronary artery disease, circulatory problems associated with stroke, nerve disease, kidney disease and eye disease. The project team based the Disease Burden Engine on a disease simulation model previously published by the CDC and Research Triangle Institute and used the CDC's 2001–02 National Health and Nutrition Examination Survey dataset to provide epidemiologic data.
  • The Population Selection Engine estimated how patients' decisions to enter and leave provider panels, payer plans and associated management programs dilute the effect of IT-enabled diabetes management. The project team derived estimates for the rates at which patients move between plans and programs from focused literature searches.
  • The Net Benefit Projection Engine characterized the distribution of providers and patients in health care organizations of various sizes in order to aggregate the cost benefit and scale it to a national level. This engine relied on estimates of the variation in size of health care organizations, using data from the U.S. Census, the American Medical Association and the Community Tracking Study data of the RWJF-funded Center for Studying Health System Change.

Hypothetical Assumptions

The researchers' projections relied on several hypothetical assumptions:

  • IT-enabled diabetes management would be adopted nationwide and deployed uniformly over a five-year period, so that each year another 20 percent of all organizations would come on board.
  • All diagnosed and insured patients with Type 2 diabetes (i.e., "adult-onset diabetes") older than age 25 were included in the analysis.
  • The full impact of IT-enabled diabetes management on process of care would be achieved in the first year of implementation.
  • The impact remained constant over the 10 years considered in this analysis, as long as a patient remained in the diabetes management program.

Challenges

The project team spent more time than expected adapting the disease simulation model previously published by the CDC and Research Triangle Institute. This, along with the unexpected need to shift the focus of the analytical model from multiple diseases to diabetes, extended the duration of the project.

Communications

The project team published three articles on the project in peer-reviewed journals, including an article in Diabetes Care (available online).

A project report entitled The Value of Information Technology-Enabled Diabetes Management is also available online.

The project team made some 20 presentations on the project to professional groups. See the Bibliography for details.

 Back to the Table of Contents


FINDINGS

As noted in the project report, The Value of Information Technology-Enabled Diabetes Management (available online):

  • All forms of IT-enabled diabetes management systems improve processes of care, improve physiological measures, prevent development of diabetic complications and generate cost-of-care savings.
  • Technologies used by providers seem to be the most effective in improving the lives of patients with diabetes.
  • Diabetes registries appear to be the most effective of all and are the only technology that showed a net cost saving after implementation. Scaled to the national level, over a period of 10 years, diabetes registries would:
    • Save $14.5 billion ($1,016 per enrolled patient) in cost of care. With implementation costs of $6.16 billion, overall net savings would be $8.34 billion.
    • Improve care processes by improving average:
      • Screening for diabetic eye disease from 14 percent to 62 percent.
      • Screening for diabetes-related nerve disease from 45 percent to 80 percent.
      • Screening for microalbuminuria (i.e., the appearance of protein in the urine, an early warning of kidney damage) from 45 percent to 66 percent.
    • Improve physiological measures, including hemoglobin A1b levels (a measure of blood sugar control), systolic blood pressure and cholesterol levels.
    • Improve clinical outcomes by reducing:
      • Heart attacks by 100,000
      • Strokes by 5,200
      • Kidney failure by 5,600
      • Amputations by 560,000
      • Blindness by 63,000
      • Diabetes-related mortality by 710,000.
  • Integrated provider-patient system would add benefits beyond a registry alone. Although overall net costs of implementation ($41.9 billion) were highest for integrated systems, the potential for improved physiological measures and clinical outcomes far outweighed corresponding projections for the other forms of IT-enabled diabetes management. For example, over a period of 10 years, national adoption of integrated provider-patient systems would cumulatively reduce:
    • Heart attacks by 160,000
    • Strokes by 16,000
    • Kidney failure by 7,900
    • Amputations by 560,000
    • Blindness by 64,000
    • Diabetes-related mortality by 920,000.

Limitations

Although the project team believes these results to be the best estimate thus far of the costs and benefits of IT-enabled diabetes management, the team notes two key limitations to its findings:

  • The analytical model covered only costs directly related to diabetes. It did not cover the economic impact of diabetes complications, such as uncontrolled hyperglycemia or pneumonia, and reductions in indirect costs from factors such as lost days from work.
  • As a result of the proprietary nature of cost information, the project team had difficulties obtaining sufficient data to reflect the wide range of approaches found in diabetes management. Thus, it was not possible to model the costs for all ways in which each type of technology is utilized in practice, and extrapolation, based on interviews, was required to complete the analysis.

 Back to the Table of Contents


LESSONS LEARNED

  1. Clarify and finalize project scope early in the project cycle. It is difficult to create precise estimates of the resources and time required for new or novel research. The project team did not have a historical basis upon which to estimate resource and time requirements and so miscalculated the amount of time and funding the project would take. (Project Director/Middleton)

 Back to the Table of Contents


AFTER THE GRANT

The project team continues to promote the project report across traditional media and Web channels. It is also preparing a mailing of the project report to all federal congressional staffers, highlighting the importance of funding additional research into the effects of IT-enabled programs for managing diabetes and the need to develop means of funding adoption and maintenance of these programs across the United States.

 Back to the Table of Contents


GRANT DETAILS & CONTACT INFORMATION

Project

Studying the Value of Health Care Information Technology in Chronic Disease Management

Grantee

Partners HealthCare System, Center for Information Technology Leadership (Wellesley,  MA)

  • Amount: $ 350,005
    Dates: December 2003 to November 2005
    ID#:  049931

Contact

Blackford Middleton, M.D., M.P.H., M.Sc.
(781) 416-8530
bmiddleton@partners.org

 Back to the Table of Contents


APPENDICES


Appendix 1

Expert Advisory Panel

Madhu Agarwal, M.D.
Acting Deputy Chief Officer of Patient Care Services
Veterans Administration
Washington, D.C.

Brian Austin
Deputy Director
The Improving Chronic Illness Care Program
Group Health Cooperative
Seattle, Wash.

Stephen J. Brown
President and CEO
Health Hero Network
Redwood City, Calif.

Lawrence P. Casalino, M.D., Ph.D.
Assistant Professor
University of Chicago
Chicago, Ill.

Timothy G. Ferris, M.D., M.P.H.
Partners/MGH Institute for Health Policy
Massachusetts General Hospital
Boston, Mass.

Jeremy Grimshaw, M.B.Ch.B.
Director
Centre for Best Practices
Institute of Population Health
University of Ottawa
Ottawa, Ontario
Canada

Karen M. Kuntz, Sc.D.
Associate Professor
Harvard School of Public Health
Boston, Mass.

John A. Merenich, M.D.
Regional Director
Chronic Disease Management Program
Colorado Permanente Medical Group
Denver, Colo.

David Wennberg, M.D.
President and COO
Health Dialog Analytic Solutions
Boston, Mass.

 Back to the Table of Contents


BIBLIOGRAPHY

(Current as of date of this report; as provided by grantee organization; not verified by RWJF; items not available from RWJF.)

Articles

Adler-Milstein J, Bu D, Pan E, Walker J, Kendrick D, Hook JM, Bates DW and Middleton B. "The Cost of Information Technology-Enabled Diabetes Management." Diabetes Management, 10(3): 115–128, 2007. Available online.

Bu D, Pan E, Walker J, Adler-Milstein J, Kendrick D, Hook JM, Cusack CM, Bates DW and Middleton B. "Benefits of Information Technology-Enabled Diabetes Management." Diabetes Care, 30: 1137–1142, 2007. Available online.

Kendrick DC, Bu D, Pan E and Middleton B. "Crossing the Evidence Chasm: Building Evidence Bridges from Process Changes to Clinical Outcomes." Journal of the American Medical Informatics Association, 14: 329–339, 2007. Abstract available online.

Reports

Bu D, Pan E, Johnston D, Walker J, Adler-Milstein J, Kendrick D, Hook JM, Cusack CM, Bates DW and Middleton B. The Value of Information Technology-Enabled Diabetes Management. Chicago: Healthcare Information and Management Systems Society, 2003. Available online.

 Back to the Table of Contents


Report prepared by: Robert Crum
Reviewed by: Richard Camer
Reviewed by: Marian Bass
Program Officer: Stephen J. Downs

Most Requested