outlook for patients with diabetes is improving, however, as better IT-enabled diabetes management is believed to create value by of diabetes …

The Value of Information Technology-Enabled Diabetes Management

Davis Bu, MD, MA Eric Pan, MD, MSc Douglas Johnston, MA Janice Walker, RN, MBA Julia Adler-Milstein David Kendrick, MD, MPH Julie M Hook, MA, MPH Caitlin M Cusack, MD, MPH David W Bates, MD, MSc Blackford Middleton, MD, MPH, MSc

Center for Information Technology Leadership Partners HealthCare, Inc One Constitution Center Information Systems Department, Second Floor West Charlestown, MA 02129

2007 by the Center for Information Technology Leadership Published and distributed by the Healthcare Information and Management System Society HIMSS All rights reserved No part of this publication may be reproduced, adapted, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher Requests for permission to reproduce any part of this work should be directed to Ellen S Rosenblatt, Manager of Operations Center for Information Technology Leadership Partners HealthCare, Inc One Constitution Center Information Systems Department, Second Floor West Charlestown, MA 02129
erosenblatt@partnersorg ISBN: 0-9777903-5-5 For more information about CITL, please visit wwwcitlorg

Contents CITL
Chapter 1 Introduction 1
Background 1 Value Proposition and Research Gaps 3

Chapter 2 Approach to the Analysis7
Introduction 7 Available Technologies: ITDM Taxonomy 8 How Technologies Affect Processes of Care: ITDM Impacts Engine 10 How Changes in Processes of Care Result in Improved Clinical and Economic Outcomes: ITDM Disease-Burden Engine 11 Costs of the Systems: ITDM Implementation-Cost Engine 13 How Patient Participation Affects Value: ITDM Population-Selection Engine 16 How Costs and Benefits Apply to Different Settings: ITDM Net-Benefit Projection Engine 17 Sensitivity Analysis and Stability Testing 19 Approach-to-Analysis Summary 19

Chapter 3 Results 23
Common Assumptions23 Technologies Used by Payers 23 Technologies Used by Providers 25 Technologies Used by Patients 27 Integrated Provider-Patient Systems29 Net Benefit to Organizations31 Sensitivity Analysis and Stability Testing 31 Additional Benefits 31

Chapter 4 Discussion 33
Key Limitations 34 Conclusion 36

Acknowledgments39 Appendix 1: Literature Summary 41 Appendix 2: Intervention-Cost-Engine
Estimates 43 Appendix 3: Expert-Panel Biographies 47 References 53 Index 57

Chapter : Chapter title

iii

Chapter 1: Introduction CITL
Background
Diabetes, a condition in which the body has lost the ability to produce, or to correctly utilize insulin, is the fifth-leading cause of death by disease in the United States1 An estimated 208 million Americans have diabetes,2 and the American Diabetes Association ADA reports that in 2002 the direct and indirect costs of the disorder totaled more than 132 billion3 Because of demographic shifts–changes in population size, age distribution, and ethnic diversity, for example–the ADA projects that by 2010 the number of patients with diabetes will have risen by 20 percent and associated costs will be 156 billion3 Even more alarming is the fact that these estimates may actually be conservative, given that more Americans than ever are struggling with diabetic risk factors such as obesity The outlook for patients with diabetes is improving, however, as better knowledge of diabetes provides them a broader array of options Landmark studies have shown that tight management of the disease can prevent many complications, including stroke, blindness,
heart disease, and death Further, treatments such as laser eye surgery can help control complications after they develop In addition, an improved understanding of risk factors for diabetes has enabled earlier diagnosis, even prevention4 Although these advances have resulted in widely cited guidelines to providers, patients with diabetes often fail to receive the recommended care; a survey conducted by McGlynn in 2004 revealed that physicians complied with diabetic guidelines less than half the time5 This noncompliance results in part from the structure of our health care system, which–despite recognition that chronic conditions such as diabetes account for the majority of health care costs–is oriented toward treating acute problems6 Chronic conditions require ongoing multidisciplinary care, as opposed to infrequent visits to a physicians office, and they require patient education on self-care–blood-sugar monitoring, adherence to dietary recommendations, exercise, and regular foot inspection, for example The chronic care model and disease management are approaches designed to meet many of these needs7 The chronic care model integrates community resources, health system
organizations, and self-management with the aid of mechanisms such as decisionsupport and clinical-information systems Figure 1-18,9 It provides multidisciplinary

Chapter : Chapter title 1: Introduction

evidence-based care and gives the patient the education and tools needed for intimate involvement in the management of his or her disease Chronic Care Model 8

1-1

Figure

Disease management is defined as a system of coordinated health care interventions and communications for populations with conditions in which patient self-care efforts are significant10 It is rooted in the assumption that care for chronic diseases can be greatly improved without the organizational changes required by the chronic care model7 Most disease management programs are built around four specific processes concerning patientsidentification, enrollment, engagement, and retentionwhich work in concert to improve quality of care and outcomes Figure1-2 In fact, payers, recognizing that disease management may help avoid or delay costly complications of chronic diseases, have been investing in disease-management programs, many of which only indirectly involve physician practices11 Such programs emphasize
the empowerment of patients to manage their own care and to use evidence-based guidelines to help them stay as healthy as possible10 A growing body of literature suggests that diabetes-management programs in particular need an information technology IT backbone in order to be effective For instance, the Health Care Delivery Work Group from the National Institutes of Healths 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 it12 The advantages of IT tools include: promoting better

The Value of Information Technology-Enabled Diabetes Management

provider-guideline compliance by presenting recommendations at the point of care; helping to identify patients overdue for care and assisting providers to proactively reach out to them; enabling patients to manage their own care through education and communication tools that allow them to receive direct feedback; and providing numerous other benefits as well Components of Disease Management
Figure

1-2

Value Proposition and Research Gaps
IT-enabled diabetes management is believed to create value by
improving processes of care, which reduces the rate of diabetic complications, which in turn produces both cost savings and enhanced quality of life But the literature on cost-benefit and cost-effectiveness measures of diabetes management has been limited Shortcomings in the literature include brief study durations, a lack of generalizability of results to external settings and populations, and failure to account for factors such as identification, enrollment, and retention of patients In a 2004 analysis of the general disease-management literature, the Congressional Budget Office reported that there is insufficient evidence to conclude that disease-management programs can generally reduce overall health spending13 Examples of questions unanswered by the current literature include: Do short-term improvements result in long-term benefit? By necessity, studies reported in the existing literature usually focus on narrow and specific outcomes over

Chapter 1: Introduction Executive Summary

short periods of time The time and money required to conduct a long-term population study are not available for most organizations implementing diabetes management When short-term improvements
are seen in these studies, they may not be indicators of long-term benefit, merely delaying diabetic complications rather than preventing them How do study results generalize to other settings? The effectiveness of diabetesmanagement programs depends in large part on the characteristics of its patients A relatively healthy patient population will suffer fewer complications than a population at higher risk and will have less opportunity for improvements in outcomes Therefore results from a study conducted on young patients with diabetes may not apply, for example, to an older and less healthy Medicare population What are the total costs of the programs? Many studies fail to provide any cost data Others report some costs but fail to include measures such as the cost of the intervention itself or of identifying eligible patientsWithout a full accounting of all the costs of diabetes management, cost-benefit analyses will paint an overly rosy picture Are all eligible patients identified and enrolled in the program? Studies conducted in controlled environments may fail to account for problems associated with the identification and enrollment of patients with diabetes in real-world
situations Failure to efficiently identify and enroll all eligible patients–typically, because of tight constraints on time and money–may result in some patients not receiving the benefit of management Such omissions reduce the economic benefits that the program might have realized What happens to benefits when the patient changes programs? Because diabetic complications are slow to develop, continuity of care is critical to the success of disease management; it may take years of tight control for a patient with diabetes to avoid complications such as a heart attack or blindness When patients choose or are forced to change health plans or providers, and are obliged to leave a diabetes-management program prematurely, they may not have had time to change the course of their disease and reap the benefits of management Does an IT-enabled diabetes management program make sense for smaller provider organizations? Because small-group practices deliver a substantial fraction of chronic care, disease management in those venues can substantially improve the health of patients with diabetes However, because small practices may not be able to realize the benefits of economies of scale, their
costs of implementing diabetes-management programs may be prohibitive A thorough understanding of the cost considerations of alternate approaches to IT-enabled diabetes management will allow practices of all sizes to make informed decisions regarding the net value of such programs

The Value of Healthcare Information Exchange and Interoperability

Does it make sense for payers? IT-enabled diabetes management may provide payers with substantial savings from avoided health-care-utilization costs However, premature health-plan switching by patients may prevent payers from realizing these savings Further, payers are limited in their management options Unlike providers, they cannot prescribe medications directly; they can only encourage providers and empower patients to make the best decisions What are the technology options? In choosing an IT-enabled diabetes-management program to implement, today stakeholders are faced with a multitude of options Confusion results from the lack of comparative-benefit studies, without which an intelligent choice is reduced to mere guesswork In this report, the Center for Information Technology Leadership CITL has addressed these and other
outstanding questions with regard to IT-enabled diabetes management

Chapter 1: Introduction

Chapter 2: Approach to the Analysis CITL
Introduction
IT-enabled diabetes management ITDM helps to improve diabetic-care processes, which in turn reduces the rate of diabetic complications, thereby generating clinical and economic benefit Figure 2-1 For instance, ITDM promotes strict dietary compliance, which improves blood-sugar levels and lessens damage to small blood vessels throughout the body This reduction in microvascular disease lowers rates of diabetes complications such as blindness, lower-extremity amputations, and end-stage renal disease Such outcomes not only improve patients quality of life but also reduce utilization of health care resources, potentially leading to cost savings We have thus focused our analysis on savings that result from improved care and reductions in complications, specifically for patients with Type 2 diabetes Improvements Lead to a Reduction in Complications adapted from CBO Report 13

2-1

Figure

CITL convened a highly qualified expert panel of nationally recognized experts who were interviewed by phone using a structured set of questions,
participated in a oneday roundtable discussion, and were consulted throughout the project The members of that panel included: MadhuAgarwal,MD,ActingDeputyChiefOfficerofPatientCareServices,Veterans Administration BrianAustin,DeputyDirector,TheImprovingChronicIllnessCareProgram,Group Health Cooperative, Seattle Stephen J Brown, President and CEO, Health Hero Network, Redwood City, Calif LawrencePCasalino,MD,PhD,AssistantProfessor,UniversityofChicago

Chapter : Chapter title the Analysis 2: Approach to

Timothy G Ferris, MD, MPH, Partners/MGH Institute for Health Policy, Massachusetts General Hospital, Boston JeremyGrimshaw,MBChB,Director,CentreforBestPractices,InstituteofPopulation Health, University of Ottawa KarenMKuntz,ScD,AssociateProfessor,HarvardSchoolofPublicHealth,Boston JohnAMerenich,MD,RegionalDirector,ChronicDiseaseManagementProgram, Colorado Permanente Medical Group, Denver David Wennberg, MD, President and COO, Health Dialog Analytic Solutions, Boston This chapter describes how CITL approached fundamental questions in order to model the value of ITDM, estimate costs associated with implementing it, and understand how efficiencies influence realization of net
benefit The chapter provides detail on the taxonomy that CITL developed for these diabetes-management strategies, the approach used to construct the value model, the methodology for determining model inputs, and the way in which we determined which process outcomes were to be evaluated To these ends, the sections in this chapter address the following questions: WhatITDMtechnologiesareavailable? HowdoITDMtechnologiesaffectprocessesofcare? HowdoITDM-influencedchangesinprocessesofcareresultinimprovedclinical andeconomicoutcomes? WhatarethecostsofITDMsystems? HowdoespatientparticipationaffectthevalueofITDM? HowdothesecostsandbenefitsapplytoITDMinaparticularsetting? HowstablearetheITDMcostandbenefitprojections?

Available Technologies: ITDM Taxonomy
To help frame subsequent analysis, CITL derived the following taxonomy of ITDM technologies, which is described in detail below: Technologiesusedbypayers Technologiesusedbyproviders o Disease registries o Clinical decision-support systems Technologiesusedbypatients o Self-management o Remote monitoring Integratedprovider-patientsystems

Technologies Used by Payers
Payer systems interface with electronic-claims systems to track
patients with diabetes and monitor diabetic-specific information Payer systems compare patient data with

The Value of Information Technology-Enabled Diabetes Management

recommended guidelines in order to identify opportunities for improved management; provide feedback to patients and providers by telephone, email, or postal mail; and can focus on behavioral change, using health coaches to convey educational and motivational information to patients The systems, which may be based at the payer organizations themselves or at separate disease-management companies, involve only those two entities and patients There is typically no point-of-care component, though many programs do follow up with providers during or after an intervention

Technologies Used by Providers
Diabetes registries track patients with diabetes and store information specific to their care At the point of care, registries may generate concise patient reports for clinicians that highlight areas for attention during the office visit Registries may also aggregate information across the population to generate report cards, which show, for example, the proportion of patients with diabetes who had foot exams during the
prior six months or a list of patients with hemoglobin A1c HbA1c levels above 7 percent, which indicates that the patients blood-sugar level is too high Registries may also use these report cards to facilitate a providers communication with patients–for instance, to generate lists for announcements about available diabetes-education sessions or to facilitate appointment scheduling Clinical decision-support systems CDSSs generate alerts and reminders for clinicians during a patient visit Such communications may caution providers about potential errors or remind them of recommended guidelines for improving quality of care In addition, CDSSs may offer information that helps providers navigate the complex array of treatment options by suggesting regimens based upon a patients condition Unlike diabetes registries, CDSSs are built on electronic medical records EMRs, which maintain comprehensive health data about each patientThough EMRs are not generally designed for population-level reporting, providers and their office staff may query them to generate such information

Technologies Used by Patients
Self-management technologies provide patients with educational resources and
datagathering mechanisms for managing their own care between provider visits These technologies include automated phone systems that generate reminders or offer educational content; electronic diary tools that collect information to be taken to a visit; interactive educational programs on computers; and online resources, such as peer support groups, sponsored by providers Remote-monitoring technologies capture and send providers information that is needed to facilitate diabetes management between office visits; patients periodically submit structured electronic data about their condition–via a telephones touchtone keypad, for example–and they receive feedback and instructions Newer remote-monitoring programs use Web sites that accept data uploaded directly from glucometers and other

Chapter 2: Approach to the Analysis

home devices Although the focus of remote-monitoring technologies is on the data sent from the patients homes to providers offices, some systems also deliver educational content to patients, content such as self-care advice in recorded phone messages, when they submit their data These systems may also connect patients to resources such as EMRs, endorsed
educational materials, interactive self-care tools, and provider e-mail

Integrated Provider-Patient Systems
CITL envisions a fully integrated chronic-care platform that would coordinate the delivery of evidence-based care across multiple care settings, including outpatient clinics and patients homes This system would include a disease registry to manage chronically ill populations, patient education tools to support self-management, and remote-monitoring devices to measure and report patient symptoms and clinical progress to providers between visits Although we could not find an example of such an integrated system, we included it in our analysis to show its potential impact

How Technologies Affect Processes of Care: ITDM Impacts Engine
CITL surveyed the literature to find the best estimates for the impacts that each form of ITDM has on Type 2 diabetes care processes We then created the ITDM Impacts Engine to transform reported evidence of physiologic and care-process improvements into expected process improvements for new care settings over time

Data Sources
CITL relied on a variety of sources to estimate the impact of each diabetes-management strategy on care processes We
searched academic publications and a wide array of nonacademic literature, including trade journals, government publications, the general press, vendor and consultant studies, proprietary research services, and studies by foundations and professional associations Preference was given to evidence published in peer-reviewed literature For each data element in the model, we chose from the evidence a single best estimate, based on study design strength and closeness of fit to the taxonomy We applied a standardized quality-scoring sheet to each study without regard to the direction or magnitude of its result We used data from these studies as inputs into the ITDM Impacts Engine In certain instances, CITL was required to convert some of these data into a format usable by the modelTable 2-1 details the final ITDM impact data used in the ITDM Impacts Engine; Appendix 1 summarizes the studies from which these inputs were derived

The Value of Information Technology-Enabled Diabetes Management

Physiologic and Care-Process Improvement Evidence from Literature Review and Evidence Synthesis
Blood Glucose
HbA1c 836 to 80214

2-1

Table

Blood Pressure
SBP mmHg 1325 to
128715

Cholesterol
LDL mg/dl 114 to 104615

Eye Exam Screening
Rate 40 to 4816

Foot Exam Screening
Rate 2 to 2517

Microalbumin Screening
Rate 273 to 37316

Payer

Payer Clinical Decision Support Systems Diabetes Registries

Provider

HbA1c 84 to 81718 HbA1c 73 to 6120 HbA1c 95 to 8621 HbA1c 745 to 74222

SBP mmHg 1381 to 13918 SBP mmHg 140 to 139120 SBP mmHg 141 to 13121

LDL mg/dl 1267 to 11218 LDL mmol/dl 32 to 2720 LDL mg/dl 100 to 9421 Total Chol:HDL 57 to 51322

Rate 122 to 19319 Rate 36 to 6920

Rate 462 to 55619 Rate 67 to 8820

Rate 233 to 43619 Rate 27 to 5520

Patient

Remote Monitoring Selfmanagement

HbA1c hemoglobin A1c, SBP systolic blood pressure, LDL low-density lipoprotein, HDL high-density lipoprotein

Model Development
Six care-impact measures were chosen for inclusion in the Impacts Engine, based on whether they 1 were consistent with diabetes-care guidelines,4 2 could be incorporated into the CITL Diabetes Disease-Burden Engine, described below, and 3 could be reported across all technologies in the taxonomy These criteria ensured that the resulting impacts would be consistent with scientific knowledge, project both clinical and economic
benefit in our model, and produce results that were comparable across all forms of ITDM in our taxonomy The resulting measures were changes in HbA1c, systolic blood pressure SBP, and cholesterol levels, as well as the rates of eye exams, foot exams, and microalbuminuria screening Some outcomes of diabetic-care guidelines, such as pneumococcal vaccinations rates, were not included because their clinical impact could not be incorporated into the model Other outcomes, such as emergency-department visits and avoided admissions, were not modeled because the impact of ITDM on them was inconsistently reported Also, although we built into our model the ability to incorporate foot-ulcer and amputation rates, we did not find evidence in the literature that ITDM affected these outcomes

How Changes in Processes of Care Result in Improved Clinical and Economic Outcomes: ITDM Disease-Burden Engine
The next step in assessing ITDM value was the projection of care-process improvements on clinical outcomes and complications from Type 2 diabetes To this end, we created the Disease-Burden Engine
Chapter 2: Approach to the Analysis

Data Sources
Because many disease-state models have been
developed to project the future disease burden of diabetes and its complications, we surveyed these models to find one that we could reuse and extend The model had to account for the social, economic, and health care environments of the United States It had to be modifiable, so that it could incorporate the effects of ITDM interventions Finally, it had to be flexible enough to handle additional model parameters to address hypotheses of interest We chose as a starting point for our Disease-Burden Engine a model created by the Centers for Disease Control and Prevention CDC in conjunction with the Research Triangle Institute RTI23,24 This model was developed for the health care environment in the United States and allowed modifications for measuring the full spectrum of ITs potential alterations of the course of disease However, the model was based only on newly diagnosed diabetics and did not account for the impact of changes in preventive screening rates and other process-of-care improvements Therefore we extended theCDC/RTIDiabetesCost-Effectivenessmodeltoincludesucheffects,andwecombined it with another published model to account for the impact of diabetic retinopathy screening25
We also expanded the model to simulate patients with pre-existing diabetes and other demographic variations in the patient population

Model Development
The resulting Disease-Burden Engine characterizes five complications of diabetes: nephropathy, peripheral neuropathy, retinopathy, coronary heart disease CHD, and stroke Figure 2-2The Engine specifies various levels of severity for each complication and how patients with diabetes may progress through those disease states For example, in the case of CHD, a patient may progress from normal to angina, suffer a heart attack, and ultimately die Combined with the ITDM Impacts Engine, the Disease-Burden Engine predicts the degree to which care-process improvements decrease the chance that a patient will progress to a more severe disease state Estimates of such changes were derived from the medical literature, consistent with the original CDC model Where credible medical evidence was lacking, no benefit was projected For example, while the landmark UnitedKingdomProspectiveDiabetesStudyreportedthatimprovedglucosecontrol might lower the risk of heart attacks,26 this result just missed statistical significance and was therefore not included
in the model

The Value of Information Technology-Enabled Diabetes Management

Simplified Schematic of Model Disease States

2-2

Figure

Costs of the Systems: ITDM Implementation-Cost Engine
CITL created the Implementation-Cost Engine to estimate the expenses involved in implementing and operating each form of ITDM

Data Sources
Because published cost estimates are not widely available, CITL relied primarily on market research for these data With the help of the Disease Management Association of America DMAA, CITL collected information via structured phone interviews with 38 organizations currently implementing ITDM or that sell diabetes-management technologies CITL interviewed at least one and up to sixteen confidential sources for each form of ITDM represented in our taxonomy

Chapter 2: Approach to the Analysis

Model Development
With the exception of Payer ITDM in which diabetes management is generally outsourced, the Implementation-Cost Model was built from the following common components: Identification Costs incurred for identifying all eligible patients with diabetes from the entire pool of patients Hardware Costs incurred for obtaining hardware eg, computers
necessary to support the IT diabetes-management program Software Software-licensing costs, when software must be purchased Programmer costs, when applications must be developed Interfaces Costs to transform data from external systems into a standard format for the diabetes-management application IT support staff Staff costs required specifically for IT components of the management program Training Costs to train users of the systems Non-IT programmatic costs Costs incurred to implement diabetes-management programs that are unrelated to their IT components Technologies Used by Payers For payer technologies, we assumed that the payers identify eligible patients in their plans and then hand over diabetes management to an external disease management vendor Thus the Implementation-Cost Engine applies costs to the payers in setting up and running claims analyses that identify patients with diabetes, and it applies the costs of staff to support the program as well The Engine then applies program-management costs including an implementation fee and a per-intervened-patient-per-month PIPM cost, wherein many of the cost components are bundled together This PIPM fee varies by size of the
intervened population and their insurance status Technologies Used by Providers For diabetes registries, the Implementation-Cost Engine assumes that practices purchase a registry application from a vendor and run it as an application service provider ASP model, thus minimizing the requirements for on-site development and support The main cost driver of this approach is the expense of building interfaces to existing practice systems We assumed that providers would require three such interfaces, based in part on reports in the literature27 Additional expenses include the costs of personal computers PCs for providers to run the application at the point of care, annual license fees for software and hosting services, costs of setting up and running claims analysis to identify patients with diabetes, and costs of quarterly alerts and reminder mailings to patients For clinical decision support, the Implementation-Cost Engine focuses on costs associated with modifying existing EMRs in order to sustain robust diabetes management The main expense, applied as a fixed cost to the organization, is in the time required to

The Value of Information Technology-Enabled Diabetes
Management

develop and implement diabetes guidelines in the EMR We did not include the costs of implementing the original EMRs themselves Other expenses include the cost of a part-time program manager to oversee the effort

Technologies Used by Patients
For remote monitoring technologies, the Implementation-Cost Engine assumes that each patient with diabetes is given a device that connects a glucometer to a telephone line and uploads data Major cost drivers in this approach are data-transmission devices and software-licensing fees for each patient, as well as personal computers, softwarelicensing fees, and staff time for registered nurses who monitor and respond to incoming data Additionally, there are the costs to set up and run claims analysis to identify patients with diabetes, interface costs for ensuring that practice systems will accept incoming data, and training costs for staff and patients For self-management technologies, the Implementation-Cost Engine assumes that patients are given access to a Website with education modules and customized selfmanagement tools The main cost drivers of this approach are PCs for clinical staff dieticians, diabetes educators, staff time
dedicated to providing tailored feedback and to moderating group forums, and the annual per-patient and per-dietician software fees charged by the vendor Additional costs are those incurred in setting up and running claims analysis to identify diabetics, building an interface to the practice system, and training staff and patients

Integrated Provider-Patient Systems
For integrated diabetes-management systems, the Implementation-Cost Engine assumes that diabetes registries are extended with remote-monitoring and self-management technologies to allow a more comprehensive diabetes-management approach–one that fully involves both providers and patients The cost of this approach is the sum of the costs for the diabetes-registry-application, remote-monitoring, and self-management platforms Any double-counted costs, such as identification and interface costs, were removed

Key Assumptions: Cost Estimates for Components
Several assumptions applied throughout the cost modeling: PhoneandInternetconnectivityexistwhereneeded,andthesecostsarenotincluded in the model Hardwareisreplacedeveryfiveyears Annualsoftwaremaintenancerepresents20percentofacquisitioncost
Annualinterfacemaintenancerepresents175percentofacquisitioncost OrganizationsbeginITDMactivitywithnodiabetes-specificITordisease-management intervention already in place Economiesofscaleandvolumediscountsaretakenintoaccount,wherepossible

Chapter 2: Approach to the Analysis

Costs were normalized to 2004 dollars, using Consumer Price Index data28 Staff costs were calculated as annual fully loaded salaries base pay plus fringe benefits and then broken down into hourly costs for a 40-hour workweek Cost estimates for the approaches were not intended to account for all of the costs associated with any particular intervention approach; they were solely meant to capture an average cost associated with using a particular type of IT Therefore CITL excluded several costs from the model These included the costs of 1 conversion of legacy data or data cleaning, 2 sales or pre-sales activities such as contracting, 3 patient-support technologies such as basic glucometers, 4 patient or provider recruitment or marketing, 5 program planning and development, 6 increases or decreases in time for providers and their staff to use the IT and information contained within it, and 7 practice
re-engineering For detailed cost assumptions, please refer to Appendix 2: Intervention-Cost-Engine Estimates

How Patient Participation Affects Value: ITDM Population-Selection Engine
The potential of diabetes management may be wasted if patients with diabetes are not actively identified, enrolled, engaged, and retained in such programs We therefore modeled the effects of churn: the entry and subsequent withdrawal of patients from IT-enabled diabetes-management programs

Data Sources
To assess rates for churn, we performed a targeted literature review and conducted interviews to estimate rates at which patients move through diabetes-management programs To factor in the rates at which patients with diabetes would be eligible for participation in the first place, we extracted diabetes incidence and prevalence rates from the 2001-2002 National Health and Nutrition Examination Survey NHANES dataset These incidence and prevalence rates were applied to all patients in the pool of patients, without regard to how frequently they are seen by a provider Table 2-2 summarizes the various patient turnover rates used by the Population-Selection Engine

The Value of Information
Technology-Enabled Diabetes Management

Annual Rates Used by the Population-Selection Engine
Remote Monitoring

2-2
Self-Management Integrated Registry CDSS Payer

Table

of Patients Retained Diabetes Incidence Diabetes Prevalence Successfully Identified for ITDM Program of Identified Patients Enrolled in ITDM Program of Enrolled Patients Retained in ITDM Program

9529

9330, 31

142

115 9332 9516, 34 9836

115 9332

115 9133 1821 9821

115 9133 6035 8837

115 9332

100

Physician participation rate assumed to be 100

Model Development
Building and maintaining participation in diabetes-management programs involve several steps Eligible patients must first be identified and invited to participate However, enrollment into the program does not guarantee continued participation Patients with diabetes may leave the program directly, or they may leave the provider panel or payer plan and no longer be eligible This exit phenomenon is offset by new patients with diabetes entering provider panels and payer plans, and by non-diabetic member patients who may develop diabetes over time We
created the Population-Selection Engine to account for such factors The Engine assumes that provider panels and payer plans remain at a constant size on average, and the number of annual new members is calculated accordingly

How Costs and Benefits Apply to Different Settings: ITDM Net-Benefit Projection Engine
CITL created the Net-Benefit Projection Engine to apply the above engines to specific settings The potential of ITDM to improve the health status of populations depends not only on the characteristics of the technologies but also on those of the population and care environments For instance, populations with higher prevalence

Chapter 2: Approach to the Analysis

rates of diabetes or with higher disease severity may have the potential to benefit more from ITDM In any case, the model was designed to project the value of ITDM for any population, as defined by a set of key demographics and average health status of patients with diabetes

Data Sources
Data from the US Census and Centers for Medicare and Medicaid Services, among others, were used to create the distribution of payer panels38-43 Data from the American Medical Association AMA and Community Tracking Study
Physician Survey were used to create the distribution of provider practices We derived national averages for diabetic-population characteristics from the 2000 Census and 2001-2002 NHANES44

Model Development
The Net-Benefit Projection Engine characterizes the distributions of providers and patients in organizations of various sizes The Engine also describes the distribution of those organizations in order to aggregate the net benefit into a national figure To be consistent with the CDC-RTI model, CITL eliminated all diabetes patients younger than age 25 CITL split the national payer scenario into four arms: Medicaid Managed Care, Medicaid Fee-for-Service, Medicare Fee-for-Service, and a Payer Mix group that included both Medicare Managed Care and commercial insurers To avoid double counting, individuals eligible both for Medicare and Medicaid were assigned to Medicare if they were age 65 or older and to Medicaid if they were younger than 65 Uninsured individuals and those covered by military health or consumer-directed health plans were excluded For provider-level analyses, CITL used AMA data to estimate the number of patients from the number of full-time specialist and
primary-care providers who would treat diabetes45 We then applied age and gender adjustments46 to account for part-time providers To determine the distribution of diabetes-care providers by practice size, CITL used estimates of all primary-care physicians by practice size extracted from the 20002001 Community Tracking Survey47 and then scaled those estimates by the percentage of providers who treat patients with diabetes45 US Census data were used to characterize the basic demographics of the general population as well as the population of each of the four insurance pools On the basis of these demographics, which included age, gender, and ethnicity, the health status of an individual was projected Health-status information included basic physiological parameters, such as systolic blood pressure, hemoglobin A1c, and cholesterol ratio, as well as prevalence of pre-existing co-morbidities such as heart attack, blindness, and stroke

The Value of Information Technology-Enabled Diabetes Management

Sensitivity Analysis and Stability Testing
In CITLs ITDM Model, we defined specific disease states for patients with diabetes, such as having angina, and determined the probability of
progressing to new disease states, such as having a cardiac arrest Because it was impractical to analyze, one by one, all possible outcomes for all possible combinations of starting demographics and health statuses, the model was run as a Monte Carlo simulation That is, a starting set of demographics and health statuses was randomly chosen, and the progressions for that simulated patient set was tracked and recorded This process was repeated hundreds of thousands of times, in each run of the model, to estimate the outcomes for a particular population of patients Because these simulation outcomes are not absolute, the model was run several times to test the stability of results Such testing showed that the financial results varied by less than one percent between runs Key model inputs were also tested through a series of one-way sensitivity analyses These variables included ITDM impact estimates used by the Intervention-Impacts Engine, patient-turnover rates used by the Population-Selection Engine, and discount rates used across the model Model runs were conducted with values for key variables changed to reasonably high and low limits, as gleaned from our evidence collection process
Where no limits were discovered in the evidence, variables were increased and decreased by 25 percent

Approach-to-Analysis Summary
The current literature does not include the long-term benefits of ITDM, and without such analysis it is difficult for organizations to make decisions on whether to invest and what kind of technology to invest in CITLs five-engine model addresses these issues The model was developed to allow for the input of various ITDM technologies, patient-population characteristics, features of health care organizations, and attributes of the US health care system as a whole The interaction of these inputs and model engines can be seen in Figure 2-3

Chapter 2: Approach to the Analysis

2-3

Figure

ITDM Model Overview

The Value of Information Technology-Enabled Diabetes Management

To fully appreciate the importance of each component of the CITL ITDM model, it is helpful to return to the questions and limitations listed earlier in the introduction to this report

Do short-term changes using ITDM project out to long-term benefit?
The ITDM Impacts Engine applies best available evidence to the Disease-Burden Engine to show how process improvements can
gradually alter the course of disease

How do study results using ITDM generalize to other settings?
By applying specific demographic, epidemiologic, and organization data to the Disease-Burden Engine, Implementation-Cost Engine, and Net-Benefits Projection Engine, the value of ITDM can be estimated for different organizations and the nation

What is the total cost of an ITDM system?
The Implementation-Cost Engine provides a comprehensive estimate of costs associated with implementation of ITDM, excluding certain costs as assumed EMR, connectivity Both cost of routine care and cost of diabetic complications are incorporated into the Disease-Burden Engine

Are all eligible patients identified and enrolled in ITDM programs?
The Population-Selection Engine dilutes the potential value of ITDM by taking into account how successful programs are in identifying and enrolling diabetics into the program

What happens to benefits when the patient changes ITDM programs?
The Population-Selection Engine accounts for patients with diabetes leaving and entering the program, and existing patients developing new-onset diabetes

Does ITDM make sense for smaller organizations?
The Net-Benefits
Projection Engine combines the estimated benefits with costs from the Implementation-Cost Engine to illustrate how economies of scale may operate with regard to ITDM

Does ITDM make sense for the payers?
The Net-Benefits Projection Engine addresses the complete cost and benefit picture to show how ITDM may benefit payers as well as providers

What are the ITDM technology options?
The ITDM taxonomy organizes the technology options into a framework that facilitates analysis, and the ITDM impact engine shows how various forms of ITDM may impact processes of care

Chapter 2: Approach to the Analysis

Chapter 3: Results CITL
Common Assumptions
The model projected the value of each form of ITDM separately–combinations of two or more forms of ITDM were excluded from our analysis–and the results that follow reflect the impact of its full national adoption We assumed that ITDM would be 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 older than age 25 were included in the analysis We assumed that the full impact of ITDM on process of care would be achieved
in the first year of implementation, given that each of the contributing studies reported impact within 12 months We also assumed that this impact remained constant over the ten years considered in this analysis, as long as a patient remained in the diabetes-management program The costs of care for patients with diabetes were derived from the original CDC model and updated to reflect changes in standards of care All financial values were calculated on a present-value basis, using a five percent discount rate

Technologies Used by Payers
Payer interventions commonly include an integrated set of IT tools that enable targeted telephone-, Web-, and mail-based management of a diabetic population along with communication with these patients providers The main IT component of such diabetes-management programs is an in-house payer- or vendor-owned diabetes-registry software application These systems interface with claims-based and select clinical data to update and integrate patient information Diabetes-management programs may monitor diabetic outcomes, process measures, and values against accepted standards and recommended guidelines in order to provide feedback to patients and providers
Patient Participation: Participation in payer programs would be relatively highWhile some patients with diabetes might choose not to participate in diabetes management by payers, enrollment and retention rates would generally be far higher than in remotemonitoring and self-management programs

Chapter 3: Results

Care Process: When fully implemented, national adoption of payer technologies would improve average: Retinopathyscreeningfrom14percentto26percent Peripheralneuropathyscreeningfrom45percentto58percent Microalbuminuriascreeningfrom45percentto53percent Physiology: When fully implemented, national adoption of payer technologies would improve average: HemoglobinA1cHbA1cby024percent SystolicbloodpressureSBPby54mmHg Cholesterolby11mg/dl Clinical Outcomes: Over the first ten years, national adoption of payer technologies would cumulatively reduce: Heartattacksby54,000 Strokesby19,000 Kidneyfailureby3,000 Amputationsby190,000 Blindnessby18,000 Diabetes-relatedmortalityby380,000 Financial Outcomes: Over the first ten years, the national rollout of payer technologies would result in: Implementationcostof216billion Cost-of-caresavingsof71billion
Overallnetcostof145billion

The Value of Information Technology-Enabled Diabetes Management

Technologies Used by Providers
Diabetes Registries
Diabetes registries improve diabetic care in a number of ways Through reminders at the point of care, registries can assist clinicians in making decisions that comply with diabetic guidelinesThrough tools such as report cards, registries provide an opportunity for clinicians to identify potential for improvement across the panel of patientsThrough mailed reminders to patients, registries can help heighten compliance and empower patients to be active participants in their own care Patient Participation: Participation in diabetes registries would be relatively high, as enrollment is typically automatic, and patients tend not to opt out The primary source of patient non-participation would be failure by the management program to identify eligible candidates for enrollment Care Process: When fully implemented, national adoption of diabetes registries would improve average: Retinopathyscreeningfrom14percentto62percent Peripheralneuropathyscreeningfrom45percentto80percent Microalbuminuriascreeningfrom45percentto66percent Physiology: When
fully implemented, national adoption of diabetes registries would improve average: HbA1cby050percent SBPby11mmHg Totalcholesterolby31mg/dl Clinical Outcomes: Over the first ten years, national adoption of diabetes registries would cumulatively reduce: Heartattacksby100,000 Strokesby5,200 Kidneyfailureby5,600 Amputationsby560,000 Blindnessby63,000 Diabetes-relatedmortalityby710,000 Financial Outcomes: Over the first ten years, the national rollout of diabetes registries would result in: Implementationcostof616billion Cost-of-caresavingsof145billion Netcostsavingsof834billion

Chapter 3: Results

Clinical Decision-Support Systems
Clinical decision-support systems are a powerful aid to clinicians at the point of care By combining patient information with the latest medical knowledge, they offer recommendations to clinicians for optimal individualized care Physicians can use these recommendations to improve control of blood sugars and other physiologic parameters as well as to improve compliance with screening and other care-process guidelines Patient Participation: Participation in CDSS would be relatively high, as patients generally do not choose to opt out The
primary source of patient non-participation would be failure by the management program to identify eligible candidates for enrollment Care Process: When fully implemented, national adoption of CDSS would improve average: Retinopathyscreeningfrom14percentto24percent Peripheralneuropathyscreeningfrom45percentto68percent Microalbuminuriascreeningfrom45percentto61percent Physiology: When fully implemented, national adoption of CDSS would improve average: HbA1cby028percent IncreaseinSBPby40mmHg Totalcholesterolby45mg/dl Clinical Outcomes: Over the first ten years, national adoption of CDSS would cumulatively: Increasestrokesby12,000Thisresultlikelyreflectstheprojectedincreaseinblood pressure See discussion section Reducekidneyfailureby2,600 Reduceamputationby340,000 Reduceblindnessby20,000 Reducediabetes-relatedmortalityby210,000 It would have no statistically significant effect on heart attacks at 005 Financial Outcomes: Over the first ten years, the national rollout of clinical decision support would result in: Implementationcostsof193billion Cost-of-caresavingsof107billion Overallnetcostof86billion

The Value of Information Technology-Enabled Diabetes
Management

Technologies Used by Patients
Remote Monitoring
Remote-monitoring technologies allow clinicians to gauge the degree of control of a patients diabetes between visits and to modify care plans accordingly These technologies offer the potential to improve blood glucose, blood pressure, and cholesterol, for example Patient Participation: Participation in remote monitoring would be low Many patients would choose not to enroll or to drop out after enrollment Care Process: When fully implemented, national adoption of remote monitoring technologies would have no statistically significant impact on provider decisions to screen for retinopathy, neuropathy, or microalbuminuria Physiology: When fully implemented, national adoption of remote-monitoring technologies would improve average: HbA1cby003percent SBPby056mmHg Totalcholesterolby28mg/dl Clinical Outcomes: Over the first ten years, national adoption of remote-monitoring technologies would cumulatively reduce: Heartattacksby12,000 Diabetes-relatedmortalityby270,000 It would have no statistically significant effect on strokes, kidney failure, amputations, or blindness at 005 Financial Outcomes: Over the first ten years, the
national rollout of remote-monitoring technologies would result in: Implementationcostsof683billion Cost-of-caresavingsof326million Overallnetcostof65billion

Self-management
Self-management technologies provide patients with the information and tools that allow them to become active participants in their own care Thus enabled, they can make healthy choices that improve control of blood sugar, blood pressure, and cholesterol; and they can have more informed discussions with their clinicians However, because self-management technologies only have an indirect effect on a clinicians behavior, their influence on processes-of-care choices may be modest

Chapter 3: Results

Patient Participation: Participation in self-management would be low Many patients would choose not to enroll or to drop out after enrollment Care Process: When fully implemented, national adoption of self-management technologies would have no statistically significant impact on provider decisions to screen for retinopathy, neuropathy, or microalbuminuria Physiology: When fully implemented, national adoption of self-management technologies would improve average: HbA1cby0020percent Totalcholesterolby79mg/dl It
would have no impact on SBP Clinical Outcomes: Over the first ten years, national adoption of self-management technologies would cumulatively reduce: Heartattacksby26,000 Diabetes-relatedmortalityby170,000 It would have no statistically significant effect on strokes, kidney failure, amputations, or blindness at 005 Financial Outcomes: Over the first ten years, the national rollout of self-management technologies would result in: Implementationcostsof162billion Cost-of-caresavingsof285million Overallnetcostof159billion

The Value of Information Technology-Enabled Diabetes Management

Integrated Provider-Patient Systems
Integrated provider-patient systems would most fully achieve the envisioned benefits of diabetes management They would help coordinate the efforts of all health care team members in delivering the best care possible, and they would provide patients with the tools to empower them in the management of their own health and to communicate effectively with their provider team Table 3-1 Patient Participation: We assumed that participation in integrated provider-patient systems would be high The primary reason for non-participation would be the failure to identify
eligible candidates Care Process: When fully implemented, national adoption of provider-patient systems would improve average: Retinopathyscreeningfrom14percentto62percent Peripheralneuropathyscreeningfrom45percentto80percent Microalbuminuriascreeningfrom45percentto66percent Physiology: When fully implemented, national adoption of provider-patient systems would improve average: HbA1cby068percent SBPby42mmHg Totalcholesterolby45mg/dl Clinical Outcomes: Over the first ten years, national adoption of provider-patient systems would cumulatively reduce: Heartattacksby160,000 Strokesby16,000 Kidneyfailureby7,900 Amputationsby560,000 Blindnessby64,000 Diabetes-relatedmortalityby920,000 Financial Outcomes: Over the first ten years, the national rollout of provider-patient technologies would result in: Implementationcostsof588billion Cost-of-caresavingsof169billion Overallnetcostof419billion

Chapter 3: Results

3-1
Payer
Care-Cost Savings millions System Cost millions Eye Exam Baseline 14

Table

Overview of 10-year Results
Provider Registries
7,100 14,500

Patient Remote Monitor
326

CDSS
10,700

Self Manage
285

Integrated
System

Financial

16,900

21,600

6,160

19,300

6,830

16,200

58,800

2560

6150

2350

1420

6150

Screening

Foot Exam Baseline 45 Microalbuminuria Baseline 45 HbA1c

5780

8000

6750

4490

8000

5260

6610

6140

4500

6610

-024

-05

-028

-003

-002

-068

Physiology

SBP mmHg

-54

-11

-056

-42

Total Cholesterol mg/dl -11

-31

n/s

-28

-79

-45

ESRD

3,000

5,600

2,600

n/s

7,900

Amputation

190,000

560,000

340,000

n/s

560,000

Morbidity

Blindness Cardiac Arrest Heart Attack

18,000

63,000

20,000

n/s

64,000

54,000

100,000

-4,500

12,000

26,000

160,000

Stroke Absolute Improvement

19,000
380,000

5,200
710,000

-12,000 210,000
100

n/s 270,000
130

n/s
170,000

16,000

Mortality

920,000

Relative Improvement 190

340

083

440

n/sdenoteslackofstatisticalsignificanceat 005 Mortality and morbidity results presented as reduction in ten-year cumulative incidence

The Value of Information Technology-Enabled Diabetes Management

Net Benefit to Organizations
Because some forms of ITDM may achieve economies of scale, the overall net benefit picture may vary across organizations of different sizes Across all
organizational sizes for payer technologies, remote monitoring and self-management, ITDM costs more than it saves Registries save more than they cost for all organizational sizes except single physician practices, and CDSS saves more than it costs for practices with greater than seven physicians Table 3-2 Net Benefit by Organizational Size for Registries and CDSS
CDSS 1 MD 2 MD 3 MD 4 MD 5-6 MD 7-9 MD 10-15 MD 16-25 MD 26-49 MD 50-75 MD 76-99 MD 100 MD -346,000
-293,000

Registries -17,000
46,000

3-2

Table

-231,000

120,000

-154,000

211,000

-69,000

312,000

79,000

494,000

347,000

816,000

938,000

1,530,000

1,650,000

2,360,000

3,040,000 4,220,000 17,000,000

4,080,000
5,580,000

21,40 0,000

Chapter 3: Results

Sensitivity Analysis and Stability Testing
One-way sensitivity analysis was performed against key variables: ITDM impact used by the ITDM Impacts Engine, patient-turnover rates used by the Population-Selection Engine, and discount rates used throughout the model Across the range of ITDM impacts found in the literature, overall cost of care varied by up to 27 percent Assuming a neutral effect on SBP by CDSS, cardiac complication rates improved by 21 percent,
stroke rates improved by 056 percent, and an additional 12 billion cost-of-care savings was generated, for a total of 12 billion Across a range of patient-turnover rates derived from the literature, cost of care varied by less than 40 percent The discount rate was varied from 30 percent to 85 percent; cost of care increased by up to 14 percent and fell by up to 16 percent across this range

Additional Benefits
The diabetes-management literature reports additional benefits that may add substantially to improved clinical and economic results projected by the model For example, many programs incorporate smoking-cessation guidelines4 as part of their array of interventions20,34,48,49 There is reason to believe not only that smoking cessation can improve health outcomes50 but also that patients with diabetes probably benefit more than those without51 Similarly, diabetic guidelines often include recommendations for other processes of care, such as vaccinations,4,52,53 exercise, nutrition therapy, and weight loss4 Diabetes management may improve compliance with these recommendations,20,29,54,55 which in turn may improve the health of patients with diabetes51,56,57 However, it was not
possible to incorporate these additional improvements into the model

The Value of Information Technology-Enabled Diabetes Management

Chapter 4: Discussion CITL
ITDM Improves Care
Our analysis demonstrates that all forms of ITDM improve processes of care, 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, and diabetes registries appear to be the most effective of all Based upon the current evidence, our analysis indicates that patient-centered technologies offer the least potential for benefit We believe that an integrated provider-patient platform, which adds patient-centered technologies to a registry and reminder system, would add benefits beyond a registry alone This integrated platform would most fully achieve the envisioned benefits of diabetes management

Cost Benefit Picture Varies by Technology
Not all forms of ITDM appear to be cost-beneficial Diabetes registries are the only forms that are cost-beneficial in virtually all situations CDSS is cost-beneficial only for larger provider organizations, and the remaining forms of ITDM
are not cost-beneficial regardless of organizational size

Cost Structure Varies Widely
The cost structures of the different technologies vary widely, and this has important implications on whether the associated programs can generate a net benefit For example, remote monitoring and self-management technologies have large variable costs driven by the number of patients with diabetes who are managed For selfmanagement technologies, the cost of interventions such as intense health coaching is based on a per-intervened-patient-per-month PIPM model Large components of remote monitoring costs include the costs of associated devices and subscription fees for each patient These variable costs prevent economies of scale from being realized This lack of economies of scale, together with the smaller benefit achieved by these technologies, prevent these programs from being cost-beneficial, even for large organizations in our model Payer technologies are also not cost beneficial for organizations of any size in our model, despite the presence of economies of scale Larger payer organizations are often able to negotiate lower PIPM costs However, cost reductions due to increased negotiating
leverage still do not allow large payer groups to achieve a positive net benefit in our model Diabetes registries and CDSS achieve economies of scale because most costs are fixed, regardless of the number of patients managed Benefits are highly dependant on the size
Chapter 4: Discussion

of the enterprise; each additional managed patient with diabetes brings the potential for added cost savings with little additional implementation cost Thus these technologies become cost-beneficial for larger organizations

Potential Benefit for Public Clinical Knowledge Repositories
The fixed costs associated with CDSS are the result of knowledge engineering tasks required to maintain clinical rules Clinical rule sets must be created and maintained in order for EMRs to appropriately trigger alerts, reminders and other forms of decision support If there were a publicly available national repository of relevant clinical knowledge, in such forms as alert and reminder logic, case finding definitions and report specifications, then a substantial cost would be lifted and the net benefit picture for CDSS might improve

Market Inefficiencies May Foster Suboptimal Solutions
Because cost savings from
improved care are largely reaped by payers, many diabetesmanagement programs are implemented by payers rather than providers Furthermore, providers have been slow to adopt health information technologies that underpin diabetes management programs Yet while implementation of payer technologies does improve the health of patients with diabetes and results in cost-of-care savings, providers are in the best position to improve care and control medical costs This misalignment of incentives may be causing the market to pursue suboptimal interim solutions

Key Limitations
These results reflect the synthesis of best available evidence, expert opinion, and a detailed simulation model predicting financial and clinical impact Further, our sensitivity analyses have shown that the model is robust with respect to key variables But although we believe these results to be the best estimate thus far of ITDMs costs and benefits, a few key limitations should be noted

Strength of Evidence
Estimates of the impact of diabetes management are limited by the strength of the underlying evidence, with two sets of assumptions deserving particular attention First, because the best available evidence regarding
the effect of CDSS-based diabetes management on SBP showed a detrimental effect, our model projects worsening bloodpressure control with CDSSs However, CDSSs in other settings show a neutral effect or improvement in blood-pressure control58,59 Sensitivity analysis showed additional benefits, assuming a neutral effect on blood pressure; thus the models results may underestimate the true value of CDSSs Second, the benefits of foot screenings may be overestimated We were able to identify only one randomized controlled trial demonstrating the salutary effect of foot screening

The Value of Information Technology-Enabled Diabetes Management

on amputation rates60 Because the care assumed in this study may not reflect standards of care throughout the country, it may overestimate the benefits of foot screening Additionally, given the proprietary nature of cost information, it was difficult to obtain sufficient data to reflect the wide range of approaches currently 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 current practice Based on interviews, extrapolation was required to project economies of
scale and volume discounts While we believe that all information shared in the interviews was accurate, the vendors could have under- or overestimated the costs of their program or product Similarly, some of the practices, hospitals, and payers could have unintentionally omitted specific costs Where possible, an unbiased third party reviewed the estimates Because of limited data, it was not possible to factor in the impact of organizational size in select cases For example, estimates of CDSS costs in the model do not vary by organizational size In reality, implementing an EMR modification in a large practice requires more time and planning to accommodate the requisite opinion leaders, while a smaller practice may implement a less customized set of guidelines to lower the cost The impact of these size-specific considerations was not reflected in the model Moreover, several other types of costs fall into this category The cost of patient and providerrecruitment/marketing,forexample,wasexcludedbecauseitvarieswidelyby organizational size and particular program approach

Population under Analysis
Generally, organizations target different subsets of patients for different interventions
For example, patients more severely affected by diabetes may receive interventions that are more intense and more expensive Patients newly diagnosed with diabetes may receive more educational support Organizations may adopt predictive modeling efforts to more precisely target interventions to those who would benefit the most An analysis of such targeted approaches may change the net analysis However, inclusion of severity stratification and predictive modeling was not possible for this report

Scope of Benefits
The complications included in the model are important causes of patient suffering and account for a substantial portion of health care dollars attributable to diabetes However, patients with diabetes consume health care resources that are not accounted for in our model, and savings from decreased utilization from other sources is not captured For example, decreased admissions from influenza, pneumonia, uncontrolled hyperglycemia, and other general medical conditions may account for a substantial amount of savings The American Diabetes Association ADA estimates that such complications account for 441 billion, or roughly a third of total diabetic costs3 Additionally, we did
not model indirect costs–days lost from work, for example–though there is some evidence that such costs may be avoided through ITDM61 The ADA estimates that indirect costs account for 408 billion, again a third of total diabetic costs
Chapter 4: Discussion

We assumed that diabetes prevalence rates would not increase and included only diagnosed patients in our analysis However, prevalence rates are likely to increase in the future as diabetic risk factors such as obesity become more prevalent62,63 Further, the ADA estimates that of the 208 million patients with diabetes in this country, more than six million remain undiagnosed2 Effective identification strategies within an overall disease-management program, with or without pre-diabetes components, might identify some of these undiagnosed patients Whereas our model captures the incremental benefit of transitioning patients from existing levels of care to care under diabetes management, undiagnosed patients with diabetes represent potentially greater opportunities for improvement because they are currently untreated Finally, the infrastructure for diabetes management might be leveraged for the management of other chronic
diseases, such as congestive heart failure or asthma Where possible, such reuse might allow for additional benefits to be achieved without the full costs of starting a new disease management program from scratch

Cross-Applicability of Studies
The evidence used in this analysis represents the best available data concerning the impacts of diabetes-management technologies on processes of care However, applying estimates from one setting to projections in another can be hazardous Diabetes-management programs show wide variation in a number of salient features, such as the programmatic elements included, the baseline quality of care, and the patient population For instance, remote-monitoring technologies are often targeted toward severe or difficult-to-control populations, but extrapolating that experience to all patients with diabetes nationwide may introduce error Our analysis projects the impact of ITDM when offered to all patients with Type 2 diabetes in an organization or across the country We include newly diagnosed or less severely affected patients with those at higher risk or who have already been affected by diabetic complications Clearly, some of the newly diagnosed or
less-severe patients may benefit from ITDM less than others As a result, some organizations have adopted severity-stratification strategies or predictive modeling to target intervention to those patients with diabetes who may receive greatest benefit This focusing of resources may yield a greater net benefit than is projected in our model

Conclusion
While diabetes afflicts millions of Americans and can place a tremendous clinical and financial burden on our society, diabetes management offers an opportunity to improve care processes that enhance the lives of patients with diabetes and help control the medical costs associated with their disease However, current diabetes-management strategies are limited by a lack of well-conducted studies for determining their specific impacts

The Value of Information Technology-Enabled Diabetes Management

on costs and benefits Our own study suggests that ITDM would improve the lives of patients with diabetes and generate cost savings if widely adopted, but it also suggests that misaligned incentives may cause the market to underutilize provider-based forms of ITDM Ironically, these may be the most cost-beneficial approaches of all

Chapter
4: Discussion

Acknowledgments CITL
This research was funded through the Robert Wood Johnson Foundation, grant 049931 The Center for Information Technology Leadership CITL, a research arm of Partners HealthCare, received unrestricted research support from the Health Information Management Systems Society over the time this research was conducted Please refer to the CITL Web site wwwcitlorg for a full listing of past and current sponsors None of the sponsors played a role in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; or the preparation, review, and approval of the manuscript The authors acknowledge the members of our expert panel for their contributions These expert panelists, excepting government employees, received nominal compensation for their time and effortsThe following individuals comprised the panel: Madhulika Agarwal, MD, MPH,VeteransAdministration; BrianAustin,The Improving Chronic Illness Care Program, MacColl Institute for Healthcare Innovations, Group Health Cooperative; Stephen J Brown, Health Hero Network; Lawrence P Casalino, MD, PhD, University of Chicago;Timothy G Ferris, MD, MPH, Partners/MGH
Institute forHealthPolicy,MassachusettsGeneralHospital;JeremyMGrimshaw,MBChB,PhD, FRCGP,UniversityofOttawa;KarenMKuntz,ScD,HarvardSchoolofPublicHealth; JohnAMerenich,MD,KaiserPermanenteColorado;andDavidWennberg,MD,MPH, Health Dialog Data Services The authors would also like to acknowledge the Disease Management Association of AmericaandKarenFitzner,formerDirectorofResearchandProgramDevelopment, for their help; and Chris Colonian at CIGNA, Jaan Sidorov at Geisinger, and David Wright at American Healthways for their assistance with the payer information Additionally the authors would like to thank and acknowledge the following contributors: David Abelson, ParcNicollet; Michael Albisser, HumaLink; Doug Bach, Colorado Access; Bruce Barter, Centene Corporation; Susan Becker, Evanston Northwestern Healthcare; Heidi Bossley, AMA; Steve Brown, Health Hero; Wayne Burton, BankOne; CandyChitty,QualityFirstHealthcareConsulting,Inc;JimChristian,PHCCLP;Nicole Cook, HCNetwork; Steve Delaronde, Connecticare; Gretchen Flanders, Colorado Access; JonahFrohlich;AdamHabigandTeriWallace,iSprit;JohnHaughom,PeaceHealth;John Haughton, DocSite; John Holland, LifeLink; Nancy Jarvis, ParcNicollet; Sharon
Katz, MillsPeninsulaMedicalGroup;VinceKuraitis;StanLapidos,RushUniversityMedical Center; John Larsen, Cerner Corporation; Diane Lee, MEDai; Pat Lierman,Anthem; Kevin Maher, McKesson; David McCullough, GHC; Gretchen McGinnis; John Merenich, Kaiser; Lisa Mohr, BCBS of SC/Companion Health; Monica Neubauer,
Acknowledgments

Medica; Derek Newell, Lifemasters; Jeremy Nobel, Harvard School of Public Health; GordonNorman,Pacificare;DrGregoryPreston,KeystoneMercyHealthPlan;Doug Reagan, iMetrikus; Ed Rutherford, Teleminder; Chris Selecky, Lifemasters; Wells Shoemaker, PMG Santa Cruz; Skip Sievert, PHPMCS; Charlotte Silvers, Sid Peterson Memorial Hospital; Bob Stone, American Healthways; Mike Summers, McKesson; DavidTeitelman, Pacificare;VictorVillagra, Health andTechnologyVector; Sandeep Wadhwa,McKesson;andRandyWilliams,PharosInnovations

The Value of Information Technology-Enabled Diabetes Management

Appendix 1: Literature Summary CITL
Summaries of the articles used in our ITDM Impacts Engine can be found below The selection of articles was based on study design strength and closeness of fit to the taxonomy The study findings presented below were either directly inputted into our
model or served as the basis for calculated inputs Preference was given to peer-reviewed literature

Technologies Used by Payers

Newell and colleagues14 examined data of a LifeMasters program in a statewide health maintenance organization The LifeMasters system compiled data from various sources including remote-monitoring devices and used algorithms to generate a customized intervention plan for staff to follow for each patient One-page summaries for physicians displayed trends in blood pressure, weight, and blood-glucose data together with LifeMasters notes from patient interactions Among patients with diabetes in the program for 12 months, average HbA1c decreased from 836 percent to 802 percent64 Data provided by the Geisinger Health Plan Systems disease management program15 was analyzed and systolic blood pressure fell from 1325 mmHg to 1287 mmHg while LDLcholesterollevelsfellfrom1140mg/dlto1046mg/dl65 Rubin and colleagues17 compared the experiences of patients with diabetes in seven HMOs at baseline and after 6-14 months in a program run by Diabetes Treatment Centers of America The companys electronic tracking system included information about patient contacts, laboratory
results, class attendance, hospital admissions, specialist visits, and ER use Company staff worked both with patients and their physicians, sending reminders about screening and visits, supporting education, and providing nurse advocates Screening rates for foot exams increased from 2 percent to 25 percent Villagra andAhmed compared baseline and one-year intervention results in patients with diabetes enrolled in a disease-management program in 10 urban areas16 A company software program included data from remote-monitoring devices and tracked patient progress Patients had access to Web-based education and received phone calls from nurses as well as reminders and educational mailings Annual dilated retinal exam rates increased from 40 percent in the baseline period to 48 percent in the follow-up period, and microalbuminuria screening rose from 273 percent to 373 percent16

Diabetes Registries

Montori, et al20 reported that a diabetic electronic management system DEMS increased the frequency of microalbuminuria testing 27 percent to 55 percent, eye 36
Appendix 1: Literature Summary

percent to 69 percent and foot 67 percent to 88 percent exams and increased control of HbA1c
levels, total cholesterol and blood pressure The DEMS was used by all persons on the care team, and the system included workflow tools specific to each role It issued care prompts based on guidelines from the ADA and it also allowed clinicians to set patient goals

Clinical Decision-Support Systems

Meigs and colleagues18 conducted a one-year randomized controlled trial to evaluate a Web-based diabetes disease-management application with interactive decision support After opening the application, clinicians saw a single-screen view of diabetes-related information about the patient, including trended and tabular laboratory results, reminders about routine exams, and specific treatment recommendations eg, LDL exceeds goal of 100 Consider starting fluvastatin The authors reported improved control of HbA1c levels, total cholesterol and blood pressure in the intervention group In a six-month randomized controlled trial, Lobach and Hammond19 compared use of a diabetes-specific encounter form to a standard encounter form An algorithm first compared the local clinicians version of ADA guidelines to information in the patients EMR, and it then produced a paper form showing the guidelines,
recommendations, and due dates for the patient Median guideline compliance for the clinicians receiving the recommendations was 320 percent, compared to 156 percent in the control group and included higher rates of eye, foot and microalbuminuria screening

Remote Monitoring
In a randomized controlled trial at theVeteransAffairs Boston Health Care System, McMahon, et al21 monitored patients with diabetes at home through a Website that accepted uploads from a blood-pressure cuff and glucometer, offered access to educational modules and other Web-based diabetes resources, and facilitated online patient communication with a care manager Based on glucose and hypertension treatment algorithms, the care manager provided recommendations to participants and primarycare providers After 12 months, average HbA1c fell from 95 percent to 86 percent, SBPfell10mmHgfromabaselineof141mmHg;andLDLcholesterolfell6mg/dl fromabaselineof100mg/dlPersistentusershadbetterresults

Self-Management

In Glasgow and colleagues 10-month randomized trial,22 all patients had access to an Internet site with educational materials and a goal-setting dietary program A subset of patients also had an online dietary coach
they could access twice a week, and another subset participated in online discussions with other patients and received electronic newsletters

The Value of Information Technology-Enabled Diabetes Management

Appendix 2: Intervention-Cost-Engine Estimates CITL
All estimates for the intervention-cost engine are broken down into two categories, acquisition and annual Acquisition costs are allocated to Year 1, whereas annual costs are incurred on an ongoing basis Payer Mediated Intervention Acquisition and Annual Costs
Acquisition Year 1 only Identification 2 weeks of IT staff time to set up and run claims analysis to identify patients with diabetes One time fee: 125,000 Program manager, IT staff, and physician time to support the implementation of intervention None Annual Years 1-10 IT staff time for monthly refreshes of claims analysis to identify patients with diabetes None Program manager, IT staff, and physician time to support intervention Per-intervened-member-permonth fee varied by insurance type and organization size

A-1

Table

Program Implementation Fee

Support Staff

Program Fee

Appendix 2: Intervention-Cost-Engine Estimates

A-2
Identification

Table

Registry
with Reminders Acquisition and Annual Costs
Acquisition Year 1 only 2 weeks of IT staff time to set up and run claims analysis to identify patients with diabetes Personal computer per physician: 500 None 3 interfaces: 10,000 each None Vendor expenses plus physician and program manager time for training None Annual Years 1-10 IT staff time for monthly refreshes of claims analysis to identify patients with diabetes 5 year hardware replacement Licensing fee for software and hosting services, includes IT support 175 of acquisition costs None Annual new employee training costs at 20 of acquisition costs Quarterly appointment reminder mailings: 4 per-intervened-member-per-year fee

Hardware Software Interfaces IT Support Staff

Training

Non-IT Programmatic Costs

A-3
Identification

Table

Modified EMR with Clinical Decision Support Acquisition and Annual Costs
Acquisition Year 1 only Nonepatients with diabetes identified during clinical encounters by physicians NoneEHR is preexisting technology 200 hours of IT staff time to develop ED forms, flow sheets, referral forms, smart text and order sets, plus cost of standard reporting tool None interfaces already exist 10 hours of
endocrinologist time to operationalize ADA guidelines Nonetraining occurs during initial EHR implementation None Annual Years 1-10 None

Hardware

None

Software

20 of acquisition costs

Interfaces IT Support Staff Training Non-IT Programmatic Costs

None 10 hours of endocrinologist time to update ADA guidelines None None

The Value of Information Technology-Enabled Diabetes Management

Remote Monitoring Acquisition and Annual Costs
Acquisition Year 1 only Identification 2 days of IT staff time to set up and run claims analysis to identify patients with diabetes Phone data transmission device for each patient: 100-160; PC for each RN: 500 One time fee: 500-1,500 1 interface: 10,000 None Cost of 1 web-ex training session 250 and 8 hrs of RN time; Fee per RN for IT support 12 None Annual Years 1-10 IT staff time for monthly refreshes of claims analysis to identify patients with diabetes 5 year hardware replacement Annual per patient fee 25-50 and per RN fee 70-90 175 of acquisition costs None Fee per patient 12; 20 of staff training acquisition costs for new staff Registered nurse time per 300 patients with diabetes

A-4

Table

Hardware

Software Interfaces IT Support
Staff

Training

Non-IT Programmatic Costs

Self Management Acquisition and Annual Costs
Acquisition Year 1 only Identification 2 days of IT staff time to set up and run claims analysis to identify patients with diabetes PC for each dietician: 500 One time fee: 500 1 interface: 10,000 None Cost of 1 web-ex training session 250-300 and 8 hrs of dietician time; Fee per dietician for IT support 12 None Annual Years 1-10 IT staff time for monthly refreshes of claims analysis 5 year hardware replacement Annual per patient fee 70-90 and per dietician fee 70-90 175 of acquisition costs None 12 fee per patient; cost of web-ex training 250-300, hosting an average 150 patients; 20 of staff training acquisition costs for new staff Dietician time per 300 patients with diabetes

A-5

Table

Hardware Software Interfaces IT Support Staff

Training

Non-IT Programmatic Costs

Appendix 2: Intervention-Cost-Engine Estimates

A-6
Identification

Table

Integrated Platform Acquisition and Annual Costs
Acquisition Year 1 only 2 weeks of IT staff time to set up and run claims analysis to identify patients with diabetes PC for each dietician, nurse, and provider: 500 each; phone data transmission
device for each patient: 100-160 each, scaled by organization Annual Years 1-10 IT staff time for monthly refreshes of claims analysis to identify patients with diabetes

Hardware

5 year hardware replacement

RM: Annual per patient fee 2550 and per RN fee 70-90 Software One time fee: 500-1,500 SM: Annual license fee per patient 70-90 and per dietician 7090 Registry: Annual license and hosting fee for provider Interfaces IT Support Staff 3 interfaces: 10,000 each None RM: Cost of 1 web-ex training session 250 and 8 hrs of RN time; 12 fee per RN for IT support SM: Cost of 1 web-ex training session 250-300 and 8 hrs of dietician time; Fee per dietician for IT support 12 Registry: vendor expenses plus physician and program manager time for training 175 of acquisition costs None

Training

12 fee per patient; cost of web-ex training 250-300, hosting an average 150 patients for combined self management and remote monitoring training; 20 of registry training acquisition costs for new staff training

SM: Dietician time per 300 patients with diabetes Non-IT Programmatic Costs None RM: Registered nurse time per 300 patients with diabetes Registry: Quarterly appointment reminder mailings: 4
per-intervened-member-per-year fee

The Value of Information Technology-Enabled Diabetes Management

Appendix 3: Expert-Panel Biographies CITL
Madhulika Agarwal, MD, MPH, Acting Deputy Chief Officer of Patient Care Services, Veterans Health Administration
Dr Agarwal is an internist who currently serves as the Deputy Chief Officer of Patient CareServicesintheVeteransHealthAdministrationsOfficeofPatientCareServices Inthiscapacity,sheistheprincipaladvisortotheVHAsUndersecretaryofHealthon policy issues that relate to patient care and clinical services VHAprovideshealthcaretomorethan51millionveteransand76millionenrollees throughout the United States With a medical-care budget of approximately 30 billion,VHAdirectlyemploysmorethan196,500healthcareprofessionalsatmorethan 1,300 sites of care, including hospitals, community- and facility-based clinics, nursing homes, domiciliaries, readjustment counseling centers, and various other facilities In addition to its medical-care mission,VHA is the nations largest provider of graduate medical education and a major contributor to medical and scientific research More than 125,000 volunteers, 80,000 health profession trainees, and 25,000
affiliated medical facultyareanintegralpartoftheVHAcommunity Dr Agarwal received her MD degree from Rajasthan University in India She completed her training in internal medicine at theVA Medical CenterGeorgetown University program in Washington, DC She is a Diplomate of the American Board of Internal Medicine She has also completed her Masters in Public Health at George Washington University in Washington, DC Dr Agarwal previously served as the Associate Chief of Staff for Ambulatory Care at theVA Medical Center,Washington, DC, where she oversaw primary care, emergency-room services, and the Community-Based Outpatient Clinics She has also been involved in training medical students and residents as well as in health-services research She holds a faculty appointment as Assistant Professor in the Department of Medicine at Georgetown University

Brian Austin, Deputy Director, Improving Chronic Illness Care, and Associate Director, MacColl Institute for Health Care Innovation, Seattle
Mr Austin is Deputy Director of Improving Chronic Illness Care ICIC, a national programoftheRobertWoodJohnsonFoundation;andAssociateDirectoroftheGroup Health Cooperatives MacColl Institute for Health Care
Innovation, which he helped found in 1992 The MacColl Institute is devoted to developing and disseminating innovations in the delivery of health care, especially within a primary-care environment
Appendix 3: Expert-Panel Biographies

Mr Austin is also a co-developer of the Chronic Care Model, a systems approach for improving the delivery of care to the chronically ill, which MacColl has been testing and refining since the mid-1990s The Model has been widely published and broadly adopted as an organizing framework, both nationally and internationally Since 1998, the MacColl Institute has also served as the home for ICIC, for which Mr Austin has been the lead administrator since its inception He is also a member of the administrative leadership team of the Group Health Cooperatives Center for Health Studies the Group Health Cooperative is the MacColl Institutes parent The Center conducts epidemiologic, health-services, behavioral, and clinical research addressing a wide and evolving range of clinical and public health questions

Stephen J Brown, President and CEO, Health Hero Network, Redwood City, Calif
Mr Brown is the founder of the Health Hero Network During his leadership as
CEO, the company has secured over 50 million in private financing to develop and commercialize the Health Buddy system, a technology platform that educates and monitors patients at home and links them to chronic-care improvement services The Health Hero Network is recognized as an industry-leading innovator and solution-provider in care management, with demonstrated and published outcomes showing quality improvement and cost-effectiveness for major health care institutions Mr Brown began his career by developing disease-management systems and software for pharmaceutical and medical-device companies while conducting research on interactive technologies for improving patient self-care and health-related behavior Mr Browns research has included some of the first interactive television and information-appliance applications for disease management, patient education, and behavioral health, resulting in over 50 patents assigned to Health Hero Network Mr Brown graduated with a BS in Physics from Stanford University

Lawrence P Casalino, MD, PhD, Assistant Professor, University of Chicago
Dr Casalino is a physician and health-services researcher at the University of Chicago In addition to
his 20 years as a family physician in private practice, he earned a PhD in health-services research, with a focus on organizational and institutional sociology and economics He is a recipient of an Investigator Award in Health Policy Research from the RobertWood Johnson Foundation Dr Casalino conducts research in two main areas: the effects of varying forms of organization on physician practice and the effects ofphysician/hospitalandphysician/health-planrelationshipsonthequalityandcostsof medical care He also studies the ways in which public and private policies shape these forms and relationships Among the journals in which his work has been published are the New England Journal of Medicine, the Journal of the American Medical Association, Health Affairs, Health Services Research, the Journal of Health and Social Behavior, and the Journal of Health Politics, Policy and Law

The Value of Information Technology-Enabled Diabetes Management

Timothy G Ferris, MD, MPH, Partners/MGH Institute for Health Policy, Massachusetts General Hospital, Boston
DrFerrisisapracticinggeneralinternistandpediatrician,ViceChairforQualityand
SafetyforPartnersPediatrics,andaseniorscientistinthePartners/MGHInstitutefor Health Policy He directs disease-management programs at Partners HealthCare, with specific responsibility for design, oversight, and evaluation of programs to improve quality and efficiency of care for high-risk patients such as those with heart failure His research has focused on quality-improvement interventions for adults and childrens health care In addition, he has studied the effects of the organization and financing on the costs and quality of care, risk adjustment of quality measures, and disparities in health care Dr Ferris has been a member of the Health Care Quality and Effectiveness Research study section of the Agency for Healthcare Research and Quality, and he has chaired two Technical Advisory Panels for the National Quality Forum

Jeremy Grimshaw, MBChB, Director, Centre for Best Practices, Institute of Population Health, University of Ottawa
Jeremy Grimshaw received an MBChB MD equivalent from the University of Edinburgh,UKHetrainedasafamilyphysicianpriortoearningaPhDinhealthservices research at the University of Aberdeen He moved to Canada in 2002 His research focuses on the evaluation of
interventions to disseminate and implement evidence-based practice Dr Grimshaw is Director of the Clinical Epidemiology Program at the Ottawa Health Research Institute; Director of the Centre for Best Practices at the University of Ottawas Institute of Population Health; Director of the Canadian Cochrane Network and Centre; andTier 1 Canada Research Chair in Health KnowledgeTransfer and Uptake He is a full professor in the Department of Medicine, University of Ottawa, with a cross-appointment to the Faculty of Medicines Department of Epidemiology and Community Medicine Earlier, Dr Grimshaw held a Personal Chair in Health Services Research at the University ofAberdeen,UKandwastheProgramDirectoroftheEffectiveProfessionalProgram within the Health Services Research–probably the largest research-implementation programintheUKDrGrimshawhasestablishedacomparableprograminOttawa He has a full registration with the General Medical Council and is member and Fellow of the Royal College of General Practitioners Dr Grimshaws research interests can be grouped according to three themes: systematic reviews of professional, organizational, financial, and regulatory interventions to improve
professional and health care system performance; the design, conduct, and analysis of rigorous evaluations of dissemination and implementation strategies; and guideline-development methods

Appendix 3: Expert-Panel Biographies

Karen M Kuntz, ScD, Associate Professor, Harvard School of Public Health, Boston
DrKuntzisAssociateProfessorofDecisionScienceintheDepartmentofHealthPolicy and Management at the Harvard School of Public Health She is an internationally recognized decision analyst with extensive experience in the methods and applications of simulation modeling for evaluating clinical and public health strategies She is currently principal investigator of one of the NCI-funded Cancer Intervention and Surveillance Modeling Network CISNET grants to evaluate national trends in colorectal cancerincidenceandmortalityDrKuntzhasbecomeoneoftheleadingauthoritieson describing errors and biases that can occur in disease modeling She received her masters and doctorate, both in biostatistics, from the Harvard School of Public Health

John A Merenich, MD, Regional Director of the Colorado Permanente Care Management Institute Chronic Disease Management Program, Colorado Permanente Medical
Group, Denver
Dr Merenich is a practicing physician and Regional Director of the Colorado PermanenteCareManagementInstituteCMIforKaiserPermanente,amajorhealth care provider in Colorado He supports CMIs stated vision to synthesize knowledge about the best clinical approaches and create, implement, and evaluate effective and efficient health care programs to improve the health of our members and community Or, stated more simply: To make the right thing easier to do PriortojoiningKaiserPermanenteMedicalGroup,DrMerenichcompleted10years of active duty with the US Army While at Fitzsimons Army Medical Center, he completed a fellowship and served as a research fellow in endocrinology and metabolism Along with his current responsibilities at CMI, he is an Associate Professor of Medicine at the University of Colorado Health Science Center in Denver Dr Merenichs undergraduate training was in biology and he received his MD from Hahnemann University and Hospital, Philadelphia He has published many articles and book chapters on topics that include evaluation of the role of drug therapy, cardiac risk, and lipid management in endocrine disorders He is the recipient of several awards in
recognition of his excellence in providing superior patient care

David Wennberg, MD, MPH, President and COO, Health Dialog Analytic Solutions, Boston
Dr Wennberg is President and Chief Operating Officer of Health Dialog Analytic Solutions and Director of the Center for Outcomes Research and Evaluation, Maine Medical Center He graduated from the McGill University Faculty of Medicine in 1987 Dr Wennbergs post-graduate education was in internal medicine at the Maine Medical Center Following his residency, he was a fellow in general internal medicine at the Harvard Combined Program and received an MPH from the Harvard School of Public

The Value of Information Technology-Enabled Diabetes Management

Health An internist with specialty training in health services and outcomes research, his major research interest is quality of care for cardiovascular services Dr Wennberg has worked with Health Dialog for six years, directing the companys segmentation and analytic services for the Collaborative CareSM product line In addition to helping found and run Health Dialog Analytic Solutions, he leads a nationally recognized research team at the Maine Medical Center, focusing on the drivers
of utilization and quality in the delivery of health care services

Appendix 3: Expert-Panel Biographies

References CITL
1 National Center for Health Statistics Health, United States, 2005 Chartbook on Trends in the Health of Americans Hyattsville, MD: 2005 2 Centers for Disease Control and Prevention National diabetes fact sheet: General information and national estimates on diabetes in the United States, 2005 Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2005 3 Hogan P, Dall T, Nikolov P, American Diabetes Association Economic costs of diabetes in the US in 2002 Diabetes Care 2003;26:917-932 4 American Diabetes Association Standards of medical care in diabetes - 2006 Diabetes Care 2006;29:S4-S42 5 McGlynn EA,Asch SM,Adams J, et alThe quality of health care delivered to adults in the United States: Appendix Rand Health; 2004;WR-174 6 Centers for Disease Control and Prevention CDC The burden of chronic diseases and their risk factors: national and state perspective Atlanta, GA: Centers for Disease Control and Prevention CDC; 2002 7 Casalino LP Disease management and the organization of physician practice JAMA
2005;293:485-488 8 WagnerEHChronicdiseasemanagement:whatwillittaketoimprovecareforchronicillness?Eff Clin Pract 1998;1:2-4 9 BodenheimerT,Wagner EH, Grumbach K Improving primary care for patients with chronic illness JAMA 2002;288:1775-1779 10 The Disease Management Association of America Available at: http://wwwdmaaorg/dm_definitionasp Accessed March 2, 2006 11 Atlantic Information Services Disease management: Outcomes, strategies, outlook Washington, DC: Atlantic Information Services; 2002 12 Glasgow RE, Hiss RG, Anderson RM, et al Report of the health care delivery work group: behavioral research related to the establishment of a chronic disease model for diabetes care Diabetes Care 2001;24:124-130 13 Congressional Budget Office An analysis of the literature on disease management programs Congressional Budget Office; 2004 Available at: http://wwwcbogov/showdoccfm?index5909sequence0, Accessed October 2004 14 Newell D, Exuzides A, Gertler P The impact of a disease management program for diabetes LifeMasters; Presented at the Disease Management Association of America Annual Meeting, October 2004 15 PersonnelCommunicationJaanSiderovGeisingerHealthPlanAdhocdatapullJune21,2005 16
VillagraVG,AhmedT Effectiveness of a disease management program for patients with diabetes Health Aff Millwood 2004;23:255-266 17 RubinRJ,DietrichKA,HawkADClinicalandeconomicimpactofimplementingacomprehensivediabetes management program in managed care J Clin Endocrinol Metab 1998;83:2635-2642 18 MeigsJB,CaglieroE,DubeyA,etalAcontrolledtrialofweb-baseddiabetesdiseasemanagement:theMGH diabetes primary care improvement project Diabetes Care 2003;26:750-757 19 Lobach DF, Hammond WE Computerized decision support based on a clinical practice guideline improves compliance with care standards Am J Med 1997;102:89-98

References

20 MontoriVM,DinneenSF,GormanCA,etalTheimpactofplannedcareandadiabeteselectronicmanagement system on community-based diabetes care: The Mayo Health System Diabetes Translation Project Diabetes Care 2002;25:1952-1957 21 McMahon GT, Gomes HE, Hickson Hohne S, Hu TM, Levine BA, Conlin PR Web-based care management in patients with poorly controlled diabetes Diabetes Care 2005;28:1624-1629 22 Glasgow RE, Boles SM, McKay HG, Feil EG, Barrera M,JrThe D-Net diabetes self-management program: long-term implementation, outcomes, and generalization results Prev Med
2003;36:410-419 23 HoergerTJ,RichterA,BethkeAD,GibbonsCBAmarkovmodelofdiseaseprogressionandcost-effectiveness for type-2 diabetestechnical report 2002;RTI Project Number 6900016 24 CDC Diabetes Cost-effectiveness Group Cost-effectiveness of intensive glycemic control, intensified hypertension control, and serum cholesterol level reduction for type 2 diabetes JAMA 2002;28719:2542-51 25 JavittJC,AielloLP,ChiangY,FerrisIIIFL,CannerJK,GreenfieldSPreventiveeyecareinpeoplewithdiabetes is cost-saving to the federal government Implications for health-care reform Diabetes Care 1994;17:909-917 26 UKprospectivediabetesstudyUKPDSgroupIntensiveblood-glucosecontrolwithsulphonylureasorinsulin comparedwithconventionaltreatmentandriskofcomplicationsinpatientswithtype2diabetesUKPDS33 UKprospectivediabetesstudyUKPDSgroupLancet 1998;352:837-853 27 HummelJBuildingacomputerizeddiseaseregistryforchronicillnessmanagementofdiabetesClinical Diabetes 2000;18:107-113 28 BureauofLaborStatisticsConsumerPriceIndexAvailableat:wwwblsgov/cpi 29 Sadur CN, Moline N, Costa M, et al Diabetes management in a health maintenance organization Efficacy of care management using cluster visits Diabetes Care 1999;22:2011-2017
30 SafranDG,MontgomeryJE,ChangH,MurphyJ,RogersWHSwitchingdoctors:predictorsofvoluntarydisenrollment from a primary physicians practice J Fam Pract 2001;50:130-136 31 SorberoME,DickAW,ZwanzigerJ,MukamelD,WeylNTheeffectofcapitationonswitchingprimarycare physicians Health Serv Res 2003;38:191-209 32 PetittiDB,ContrerasR,ZielFH,DudlJ,DomuratES,HyattJAEvaluationoftheeffectofperformancemonitoring and feedback on care process, utilization, and outcome Diabetes Care 2000;23:192-196 33 OConnorPJ,RushWA,PronkNP,CherneyLMIdentifyingdiabetesmellitusorheartdiseaseamonghealth maintenance organization members: sensitivity, specificity, predictive value, and cost of survey and database methods Am J Manag Care 1998;4:335-342 34 ClarkeJ,CrawfordA,NashDEvaluationofacomprehensivediabetesdiseasemanagementprogram:progressin the struggle for sustained behavior change Dis Manag 2002;5:77-86 35 FeilEG,GlasgowRE,BolesS,McKayHGWhoparticipatesininternet-basedself-managementprograms?A study among novice computer users in a primary care setting Diabetes Educ 2000;26:806-811 36 PersonnelCommunicationConfidentialJune25,2004 37
TurninMC,BeddokRH,ClottesJP,etalTelematicexpertsystemDiabetoNewtoolfordietself-monitoring for diabetic patients Diabetes Care 1992;15:204-212 38 US Census Bureau Current Population Survey, 2004 Annual Social and Economic Supplement Table HI05 HealthInsuranceCoverageStatusandTypeofCoveragebyStateandAgeforAllPeopleAvailableat:http:// pubdb3censusgov/macro/032004/health/h05_000htmAccessedMarch2,2006 39 Atlantic Information Services, Inc AISs Directory of Health Plans: 2004 Vol 1 Washington, DC: Atlantic Information Services Inc; 2004 40 Centers for Medicare and Medicaid Services Available at: http://wwwcmshhsgov/ Accessed January 15, 2006

The Value of Information Technology-Enabled Diabetes Management

41 USCensusBureauAvailableat:http://wwwcensusgov/AccessedJanuary15,2006 42 Centers for Medicare and Medicaid Services Public Payors Share of National Health Spending 1980-2005 CMS 2005 43 US Census Annual estimates of the population by sex and five-year age groups for the United States Available at:http://wwwcensusgov/popest/estimatesphpAccessedMarch2,2006 44 Centers for Disease Control and Prevention CDC National Center for Health Statistics NCHS National Health and Nutrition
Examination Survey data [2001-2002]Available at: http://wwwcdcgov/nchs/nhanes htmAccessedJuly12,2006 45 American Medical Association Physician Characteristics and Distribution in the US 2005 Edition United States: AMA Press; 2005 46 CooperRA,GetzenTE,McKeeHJ,LaudPEconomicanddemographictrendsandsignalanimpendingphysician shortage Health Aff Millwood 2002;211:140-54 47 Center for Studying Health System Change Community Tracking Study Physician Survey, 2000-2001: [United States] [Computer file] ICPSR version Washington, DC: Center for Studying Health System Change [producer], 2003 Ann Arbor, MI: Inter-university Consortium for Political and Social Research, 2003 48 Smith SA, Murphy ME, Huschka TR, et al Impact of a diabetes electronic management system on the care of patients seen in a subspecialty diabetes clinic Diabetes Care 1998;21:972-976 49 McCullochDK,PriceMJ,HindmarshM,WagnerEHApopulation-basedapproachtodiabetesmanagementin a primary care setting: early results and lessons learned Effective Clinical Practice 1998;1:12-22 50 US Department of Health and Human Services The Health Benefits of Smoking Cessation US Department of Health and Human Services, Public Health Service,
Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health 1990; DHHS Publication No CDC 908416:1-627 51 Haire-JoshuD,GlasgowRE,TibbsTLSmokinganddiabetesDiabetes Care 1999;22:1887-1898 52 Centers for Disease Control and Prevention Recommendations of the advisory committee on immunization practices ACIP MMWR 2005;54No RR-8:1-44 53 Centers for Disease Control and Prevention Guidelines for prevention of nosocomial pneumonia centers for disease control and prevention MMWR 1997;46No RR-1:1-79 54 EastJ,KrishnamurthyP,FreedB,NosovitskiGImpactofadiabeteselectronicmanagementsystemonpatient care in a community clinic Am J Med Qual 2003;18:150-154 55 Glasgow RE,Toobert DJ Brief, computer-assisted diabetes dietary self-management counseling: effects on behavior, physiologic outcomes, and quality of life Med Care 2000;38:1062-1073 56 American Diabetes Association Immunization and the prevention of influenza and pneumococcal disease in people with diabetes Diabetes Care 2000;23:109-111 57 SigalRJ,KennyGP,WassermanDH,Castaneda-SceppaCPhysicalactivity/exerciseandtype2diabetesDiabetes Care 2004;27:2518-2539 58
MontgomeryAA,FaheyT,PetersTJ,MacIntoshC,SharpDJEvaluationofcomputerbasedclinicaldecision support system and risk chart for management of hypertension in primary care: randomized controlled trial BMJ 2000;320:686-690 59 Montgomery AA, Fahey T A systematic review of the use of computers in the management of hypertension J Epidemiol Community Health 1998;52:520-525 60 McCabe CJ, Stevenson RC, DolanAM Evaluation of a diabetic foot screening and protection programme Diabet Med 1998;15:80-84

References

61 Testa MA, Simonson DC Health economic benefits and quality of life during improved glycemic control in patients with type 2 diabetes mellitus: a randomized, controlled, double-blind trial JAMA 1998;280:14901496 62 BoyleJP,HoneycuttAA,NarayanKM,etalProjectionofdiabetesburdenthrough2050:impactofchanging demography and disease prevalence in the US Diabetes Care 2001;24:1936-1940 63 WildS,RoglicG,GreenA,SicreeR,KingHGlobalprevalenceofdiabetes:estimatesfortheyear2000and projections for 2030 Diabetes Care 2004;27:1047-1053 64 Personal communication Alexandros Exuzides Lifemasters November 17, 2004 65 PersonalcommunicationJaanSiderovGeisingerHealthPlanJune21,2005

The Value of
Information Technology-Enabled Diabetes Management

Index CITL
This index is designed to help the user easily locate topics of interest When the letter f follows a page number, it indicates that the term is located in a figure; the t indicates that the term is located in a table

American Diabetes Association ADA, 1, 35 American Medical Association AMA, 18 Application service provider ASP model, 14 Automated phone systems, 9

Biographies Agarwal, Madhulika, 7, 47 Austin, Brian, 7, 4748 Brown,StephenJ,7,48 Casalino, Lawrence P, 7, 48 Ferris, Timothy G, 8, 49 Grimshaw,Jeremy,8,49 Kuntz,KarenM,8,50 Merenich,JohnA,8,50 Wennberg, David, 8, 5051 Blindness, 1 Blood-sugar monitoring, 1

Clinical rule, 34 Commercial insurers, 18 Community Tracking Study Physician Survey, 18 Complications, reduction in, 7f Coronary heart disease CHD, complications of, 12 Cost benefit, 33 Cost modeling, assumptions in, 1516 Cost structure, 33 Cross-applicability of studies, 36

CDC/RTIDiabetesCost-Effectiveness model, 12 Center for Information Technology Leadership CITL, 5 Centers for Disease Control and Prevention CDC, 12 Chronic care model, 12, 2f Clinical decision-support systems CDSSs, 9,
1415, 26, 42 care process, 26 clinical outcomes, 26 economies of scale for, 3334 financial outcomes, 26 fixed costs associated with, 34 net benefit by organizational size for, 31t patient participation, 26 physiology, 26
Index

Data, 10, 11t, 12 Death, 1 Demographic shifts, 1 Diabetes complications of, 1, 12 defined, 1 indirect costs of, 1 number of patients with, 1 outlook for patients with, 1 Diabetes-management programs, effectiveness of, 4 Diabetes registries, 25, 33, 4142 care process, 25 clinical outcomes, 25 economies of scale for, 3334 financial outcomes, 25 net benefit by organizational size for, 31t participation in, 25 patient participation, 25 physiology, 25 with reminders acquisition and annual costs, 44t Diabetes Treatment Centers of America, 41

Diabetic electronic management system DEMS, 4142 Diabetic guidelines, 32 compliance with, 1 Dietary recommendation, 1 Disease management components of, 3f defined, 2 Disease Management Association of America DMAA, 13 Disease registry, 10

Electronic-claims systems, 8 Electronic diary tools, 9 Electronic medical records EMRs, 9 Exercise, 1

Foot inspection, 1 Foot screenings, benefits of, 3435

Geisinger Health
Plan Systems disease management program, 41

Health Care Delivery Work Group, 2 Health-status information, 18 Heart disease, 1 Hemoglobin A1c HbA1c levels, 9

Information technology IT-enabled diabetes management ITDM taxonomy, 810 Insulin, 1 Integrated platform acquisition and annual costs, 46t Integrated provider-patient systems, 10, 15, 29, 30t care process, 29 clinical outcomes, 29 financial outcomes, 29 patient participation, 29 physiology, 29 Interactive educational programs, 9 Intervention-Cost-Engine estimates, 43t integrated platform acquisition and annual costs, 46t modified EHR with clinical decision support acquisition and annual costs, 44t payer mediated intervention acquisition and annual costs, 43t registry with reminders acquisition and annual costs, 44t remote monitoring acquisition and annual costs, 45t self management acquisition and annual costs, 45t

Implementation fee, 14 Information technology IT, advantages of, 23 Information technology IT-enabled diabetes management ITDM, 35, 7 Disease-Burden Engine, 1112, 13f Impacts Engine, 1011, 11t Implementation-Cost Engine, 1316 improvement of care and, 33 Net-Benefit Project Engine, 1718 Population-Selection
Engine, 1617, 17t for small provider organizations, 4

Landmark studies, 1 Laser eye surgery, 1 LifeMasters system, 41 Literature summary, 4142 Long-term benefit, 34

Market inefficiencies, 34 Medicaid Fee-for-Service, 18 Medicaid Managed Care, 18 Medicare Fee-for-Service, 18 Medicare Managed Care, 18 Models, 11, 12, 13f Modified EHR with clinical decision support acquisition and annual costs, 44t Monte Carlo simulations, 19 Multidisciplinary care, 12

The Value of Information Technology-Enabled Diabetes Management

National Health and Nutrition Examination Survey NHANES, 16 National Institutes of Healths Behavioral Research and Diabetes Conference, 2 Nephropathy, complications of, 12 Net-Benefit Projection Engine, 18 Net benefit to organizations, 31

clinical outcomes, 27 financial outcomes, 27 patient participation, 27 physiology, 27 Remote-monitoring devices, 10 Remote-monitoring technologies, 910 Research Triangle Institute RTI, 12 Retinopathy, complications of, 12

One-way sensitivity analyses, 19 Online resources, 9

Patient education tools, 10 Patients effect of participation on value, 16 technologies used by, 910, 15, 2728 Payer interventions, 23 care
process, 24 clinical outcomes, 24 financial outcomes, 24 patient participation, 23 physiology, 24 Payer mediated intervention acquisition and annual costs, 43t Payer Mix group, 18 Payers, technologies used by, 89, 14, 2324, 41 Payer systems, 89 Payer technologies, 33 Peer support groups, 9 Per-intervened-patient-per-month PIPM model, 14, 33 Peripheral neuropathy, complications of, 12 Processes of care changes in, 1112, 13f technologies effect on, 1011, 11t Providers, technologies used by, 9, 1415, 2526 Public clinical knowledge repositories, 34

Self-management, 9, 10, 15, 27, 42 acquisition and annual costs, 45t care process, 27 clinical outcomes, 27 financial outcomes, 27 patient participation, 27 physiology, 27 Sensitivity analysis, 19, 32, 34 Short-term improvements, 34 Smoking-cessation guidelines, 32 Stability testing, 19, 32 Strokes, 1

Technologies effect on processes of care, 1011, 11t used by patients, 910, 15, 2728 used by payers, 89, 14, 2324, 41 used by providers, 9, 1415, 2526

UnitedKingdomProspectiveDiabetes Study, 12

Value,effectofpatientparticipationon,16 VeteransAffairsBostonHealthCareSystem, 42

Web site, 910

Remote-monitoring, 2728, 42
acquisition and annual costs, 45t care process, 27

Index

The Value of Information Technology-Enabled Diabetes Management

Source:tnstate.edu