These programs essentially use practice guidelines to guide the care of targeted populations – chronically ill patients who are diagnosed with diseases that are currently most amendable to evidence-based, collaborative care management by primary care physicians, support-service nurses and other providers, and the patients themselves. Most programs target diseases such as diabetes, congestive heart failure and chronic obstructive pulmonary disease.

A core assumption of clinical guideline-driven programs is that variation in practice patterns for given diagnoses introduces extraneous cost and reduces quality of care – either by underusing effective interventions or overusing ineffective ones. The typical infrastructure of disease management programs is designed to identify eligible patients and stratify them by severity of risk, engage them in ongoing educational dialogue regarding their self-care regimens, and communicate to physicians best practice guidelines tracked to individual patient charts. Programs are often equipped to profile individual physician practice performance, offering national and peer benchmarks, while some attach financial incentives to performance outcomes.

Use of these programs is accelerating nationwide. A recent Wall Street Journal article, citing Mercer Human Resources Consulting, noted that the percentage of employer-sponsored health plans offering disease management programs grew to 58 percent last year from 41 percent the year before, while a number of health plans outsource the programs to more than 100 disease management organizations.

While disease management programs in Pa. appear to be expanding, a firm assessment of the impact of these programs – on patients, on physicians and on health care cost – lags considerably behind that growth. Several programs tout material clinical improvement, while critics point to methodological difficulties in measuring and attributing clinical improvement to disease management. The level of physician enthusiasm for disease management programs remains largely unmeasured, and "pay-for-quality" incentives are gradually being added to the programs to reward successful physician participation. Some physician groups have actually initiated their own disease management programs.

It is also largely unknown how disease management impacts health care cost. Evidence of cost reduction attributable to disease management is scarce and methodologically difficult to produce. While some data suggests savings – primarily in the form of "return on investment" in a program’s infrastructure, there is credible counterevidence suggesting that disease management programs may not reduce health care costs and may even increase costs. Some supporters of disease management maintain that "best" care is warranted even if it does not reduce costs.

Widespread Health Plan Impetus

Several major health insurers in Pa. are rapidly expanding the disease management programs they have had up and running for several years.

Independence Blue Cross (IBC) has had disease management programs since the early to mid 1990s, and has recently expanded them considerably, according to Esther Nash, M.D., IBC’s senior medical director, health management program. In its Connections program, IBC has contracted with Health Dialog, a vendor offering full-time telephone access by IBC’s commercial and Medicare managed care members to "health coaches" – over 90 nurses, most based in a call center in Arizona, and all licensed in Pa. – who use evidence-based information to help IBC members learn self-care skills and adhere to their physician’s treatment plan. The program also offers Web-based informational materials on evidence-based chronic care management, says Nash.

In May of last year, IBC added a component to the program called SMART Registry – a bi-annual reporting system, also produced in collaboration with Health Dialog, to identify for primary care physicians those patients in their practice with chronic diseases and comorbidities, and – as IBC’s 2004 annual update of the program puts it – to identify patients for whom there may be significant opportunities for clinical improvement, and practice-wide opportunities to enhance evidence-based practice and decrease inappropriate variation in effective care.

The system analyzes claims, laboratory, prescription and utilization data, and runs health risk assessment and predictive modeling algorithms to identify all of IBC’s Pa. members who have been diagnosed with one or more of five chronic conditions: diabetes, chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), coronary artery disease (CAD) and asthma. Twice a year, physicians receive reports on these patients that note individual members’ hospitalizations, specialist and emergency room visits, and whether the member may be in need of a recommended test or treatment for these conditions, based on practice guidelines drawn from nationally recognized sources, updated annually, says Nash. IBC has recently contracted with a different company to add end-stage renal disease patients to the reporting system, she notes.

The program also offers provider community outreach in the form of mailings, informational meetings and four full-time provider support specialists – local clinicians who have thus far made 1,500 visits to physician offices to assist them in using the Connections and SMART Registry systems, says Nash.

Between 13,000 and 18,000 health coach calls take place each month, 40 percent of which are initiated by patients and 60 percent by the coaches, according to Nash. Of the nearly 188,000 IBC Medicare managed care members who were covered under the Connections program at the end of October 2004, says Nash, 80,000 were identified as having chronic conditions, of whom 14,000 were considered to be high-risk. Eighty-six percent of the high-risk Medicare patients have spoken to a health coach.

IBC will monitor the total medical cost of its managed care subscribers, based on claims, and expects the system to reduce the rate of increase of medical cost inflation, says Nash.

IBC distributed its first registry list last May to more than 2,500 IBC network primary care practices caring for more than 150,000 patients with chronic illnesses. IBC’s second registry list, sent in November, included over 185,000 chronic condition patients, notes Nash.

IBC’s registry format was designed with physician input to be user-friendly and easy to interpret, according to I. Steven Udvarhelyi, M.D., IBC’s senior vice president and chief medical officer. Using the registry, physicians have so far referred over 600 patients for Connections health coaching and IBC has received registry feedback from physicians on more than 3,900 IBC patients.

Physician practices with at least 25 IBC patients with chronic diseases also receive performance reports on key indicators and comparisons to other in-network peers to allow physicians to view trends in their care management, while future registries will show how the practice performs over time, says Udvarhelyi.

Although IBC’s disease management program does not have a financial incentive component, the insurer’s Quality Incentive Payment System (QIPS) for primary care physicians is based on care guidelines – such as revised HEDIS measures – that overlap with those used for chronic care in the disease management program, notes Udvarhelyi. IBC continues to explore reimbursement models to reward better clinical performance and is interested in developing such models for primary care physicians and specialists, Udvarhelyi adds.

To what degree physicians use the registry – and what incentives may be needed to increase such use – are still open questions. Charles Cutler, M.D., who practices internal medicine with another physician in a practice owned by Montgomery Hospital, says that IBC’s disease management program may be "running in the background" for some physicians, and the SMART Registry may be "another list that isn’t noticed." While nearly half of his patients are covered by IBC, some of whom have diabetes and CHF, Cutler says he has not been contacted by a disease management nurse regarding any of his patients, and does not remember seeing any registry list of his patients.

The industry effort to change physician behavior toward best practices has had a poor history of success, and there is no clear data about what works to overcome obstacles such as resistance to clinical guidelines, resistance to clinical input from outside nurses, and discrepancies among guidelines from different sources, says David Shulkin, M.D., Temple University Hospital’s Chief Medical Officer. Data on physician incentives suggests that, to get physicians to respond to clinical guidelines offered by health insurers, the health plan needs to have a significant portion of the market and financial incentives need to be 20 percent or more of a physician’s income, Shulkin adds.

Highmark Blue Cross Blue Shield has used the SMART Registry disease management system for about three years, and every six months disseminates to some 4,000 primary care physicians in western Pa. a list of patients diagnosed with five chronic conditions – diabetes, CHF, CAD, COPD and asthma – as well as those with a comorbidity of hypertension, according to Donald Fischer, M.D., Highmark’s chief medical officer. Over 120,000 patients are currently included in the registry, he says.

Highmark has selected four or five clinical quality indicators for each disease state, and uses evidence-based guidelines by national organizations, to develop patient-specific flow sheets and communicate to physicians possible care gaps – such as ensuring that asthmatics are receiving the right type and quantity of controller and rescue medications in a given year, and tracking the number of times an asthma patient has visited the emergency room or been hospitalized, Fischer explains.

Since the registry system was implemented, he says, Highmark has seen statistically significant improvement in the majority of the clinical quality indicators, which he believes can translate into cost savings. Highmark has seen a return on its investment in the program, he notes.

About ten percent of patients in the chronic condition registry are designated by Highmark as high-risk patients and receive calls by the system’s health coach nurses, and well over 1,000 patients have been referred by physicians to a health coach, says Fischer. Like IBC, Highmark has had an 85 percent success rate in outreach by health coaches to its high-risk elderly population.

Through its claims data, Highmark is able to track rates of compliance with each of the indicators from one registry cycle to the next, and the measures that are used in the registry are many of the same clinical indicators that Highmark uses for its QIPS quality incentive program, says Fischer. The eventual ability to import laboratory data into the system, he says, will allow better tracking of outcomes.

Fischer notes that Highmark’s pay-for-performance activities support the concept of disease management by giving physicians incentive to improve their QIPS scores – which translates into higher reimbursement payments – by using the SMART Registry to help achieve higher quality patient care and reduced practice pattern variation.

This year, Highmark is extending its QIPS incentive program – which had previously only been offered to primary care physicians caring for Highmark’s HMO and point of service health plans – to primary care physicians with patients covered by Highmark’s PPO plans. Highmark had recently expanded the SMART registry system to include its PPO subscribers, matching a primary care physician to a PPO patient through algorithms based on care frequency, says Fischer.

Highmark has also implemented a more hands-on quality incentive system for a subset of its physician population: the Performance Based Incentive Program (PBIP), through which Highmark medical management consultants work closely with some 840 physicians in western Pa. to analyze their practice data, including SMART Registry data, and develop process improvement plans to alter patterns and better monitor adherence to evidence-based guidelines. Physicians participating in PBIP can achieve higher payments, above and beyond the QIPS program, based on achieving the goals that physician groups and Highmark jointly agreed upon to improve the delivery of care. Currently, the PBIP groups provide care for more than half of Highmark’s western Pa. membership, comprising over one million members.

Blue Cross of Northeastern Pennsylvania (BCNP) began formalized disease management programs in 1997, which has since grown to six programs covering prenatal care, asthma, diabetes, CAD, depression and smoking cessation, according to Ellen Preece, R.N., coordinator of the programs. Using retrospective claims data analysis, and to a lesser extent, case management referral and patient self referral, BCNP identifies and sends education materials to candidates for the programs, who can then decide if they want to enroll – except patients with coronary artery disease, who are automatically enrolled and have the choice to opt out. Program interventions include self-management educational materials and telephone calls from trained nurses for moderate- to high-risk patients, says Preece.

Between 5,000 to 7,000 BCNP subscribers are enrolled in the programs, the CAD program being the largest, with 2,900 enrollees, or about 85 percent of those identified with CAD, says Preece. All programs are run in-house with the use of 17 BCNP disease management nurses, except the CAD program, which is outsourced to a vendor, she notes.

Physicians of patients enrolled in the programs receive the educational literature sent to patients, as well as practice guidelines for the relevant disease state, which BCNP adapts from nationally recognized sources, with the input of BCNP’s physician advisory committees. The programs are designed to "leverage information," improve outcomes and make efficient use of physician office visit time by reminding patients and physicians about the care regimen, e.g., screening, control and medication activities, that practice guidelines indicate is appropriate for patients with given diagnoses, says Preece.

BCNP continues its effort to improve and finalize a "value trend assessment" of the programs by looking at enrollment, satisfaction, HEDIS clinical outcome measures and cost utilization, although assessment is confounded by variables such as health care cost inflation, provider contracting changes, and changes in the practice of medical care, says Nina Taggart, M.D., BCNP’s medical director for quality. BCNP has seen statistically significant improvements in four out of six HEDIS diabetes measures, and "meaningful improvement" – as defined by NCQA – in diabetes care, asthma care and prenatal care, with BCNP now being in the 90^th percentile of health plans for blood sugar control (HbA1c), appropriate use of controller medications, and number of postpartum visits 21 to 56 days after delivery, Taggart adds.

Taggart says that return rates on physician surveys are not yet sufficient to draw conclusions about physician feedback regarding the programs, but adds that anecdotal feedback indicates that physicians are pleased at how the programs make patients better informed during office visits and less demanding of the practices’ staff and time.

Not all physicians have seen the programs work effectively. Kenneth Rudolph, M.D., who has a solo internal medicine practice in Scranton, believes that it is helpful to remind chronic disease patients of proper self-care, and met with a BCNP nurse to formulate an outreach effort for diabetic patients. Rudolph personally sent a letter to 80 of his diabetic patients a year ago, asking them to participate in a disease management program and telling them that they would be called by a BCNP nurse to help coordinate their care. He has not heard whether any of his patients enrolled in the program, and he presumes the effort has fizzled.

BCNP has seen its disease management programs lead to higher utilization of health care services when they were implemented, while Taggart attributes the increase to the cost of meeting the practice guidelines, and she says BCNP is confident that the programs will reap cost saving benefits down the road – not by lowering costs, but by dampening cost increases.

Geisinger Health Plan also started its disease management programs in 1997, and currently covers diabetes, CHF, CAD, COPD, hypertension, osteoporosis and asthma, using best practice protocols conveyed by 52 in-house nurses in the form of patient education and case management in Geisinger Health System’s primary care practices in 28 counties, as well as in about 100 private physician practices that accept Geisinger’s HMO, according to Jaan Sidorov, M.D., medical director of care coordination.

In selecting which programs to initiate, Geisinger chose diseases that are common, for which usual physician care could be improved by the addition of disease management services, and for which interventions make a difference within a relatively short period of time, e.g., it is possible to control blood pressure, to improve the level of diabetes control, to get the right medications to patients with CHF and lower their admission rate – all within one year, says Sidorov.

Geisinger’s case-management approach to disease management varies in intensity from single encounters to weekly contact between patient and nurse, while the highest risk patients are consulted in face-to-face encounters, some entailing home visits, he adds.

Geisinger Health Plan provides reimbursement incentives to physicians for reviewing and approving the care plan created by the disease management nurse – in the form of a separate CPT code for that approval, worth $25 to $50 – and plans to expand pay-for-performance incentives further this summer, in the hope that it will bring more patients into its disease management programs, says Sidorov. Geisinger also has quality care incentives based on patient satisfaction data and HEDIS measures, and it tells physicians that having patients participate in the disease management programs will drive HEDIS, says Sidorov. About one-third of physicians embrace Geisinger’s disease management programs, one-third are skeptical but willing to participate to see how it works, and another third resist having a nurse engage with their patients, Sidorov notes, adding that the proportion is about the same for Geisinger Health System-employed physician and private practice panel physicians.

The programs are positively impacting clinical quality in nearly every disease state, such as HbA1c of patients with diabetes, use of steroid inhalers among patients with asthma, and level of blood pressure control among patients with hypertension, says Sidorov. Only the CHF disease management patients have not shown improvement in clinical indicators, and Sidorov is not sure why, although he notes that physicians are already achieving benchmark levels of ACE inhibitors and beta blocker use.

Geisinger has tracked claims patterns for patients prior to entering the disease management program and after entering the program, and those patterns have consistently shown increases in the use of primary care services, which Sidorov believes is a sign of more effective use of primary care, while inpatient charges for patients in disease management programs have decreased. Disease management has not produced a change in outpatient utilization, he notes.

For its diabetes program, Geisinger found that the per-member-per-month charges were $100 lower for patients in the program, compared to patients receiving the usual care, savings primarily accruing from less frequent hospitalization with co-morbidities such as heart attacks and cardiovascular effects, says Sidorov. CHF patients in the disease management program are showing a $700 member-per-month savings, he adds.

Geisinger’s disease management programs have generated savings that exceed their cost, sometimes producing a return on investment approaching seven-to-one, Sidorov notes. However, he says, the savings are not revenue and, in a health care environment dominated by exploding expense, the best that can be expected is to help slow down the rate of increase in health care costs, although no one has yet developed a general accounting principle of disease management cost savings that factors in variables such as medical cost inflation, increases in physician and hospital reimbursement, and increases in insurance premiums.

From an economic standpoint, disease management is valuable even if it merely delays costs, believes Sidorov, because the commercial health care marketplace is measured in year-to-year financial performance. The numbers that Geisinger Health Plan is obtaining with its disease management programs, he says, are sufficient to warrant continuing the programs – even those like the asthma program, which Sidorov says is particularly difficult to assess because patients are typically younger than those in other programs and do not stay enrolled as long.

Geisinger is looking at other chronic medical conditions for which its employer customers believe that disease management may offer cost-saving opportunities, such as lower back pain, carpal tunnel syndrome and migraines, says Sidorov.

HealthAmerica’s disease management programs – currently CAD, COPD/CHF, diabetes, asthma, high-risk pregnancy, oncology, and organ transplant follow-up – use a combination of modalities to remind patients about compliance with care guidelines, including face-to-face case management, telephonic nurses, and mailed and Web-based educational literature, according to the company’s President and CEO Robert L. Dawson. The guidelines are based on nationally recognized best practice protocols, he adds.

All care decisions are made by a patient’s physician, who can talk with one of HealthAmerica’s medical directors if he or she disagrees with a care protocol, while HealthAmerica can design a course of treatment if a physician’s treatment plan conflicts with the protocol, says Dawson. For example, if a physician prescribes a drug that HealthAmerica does not cover or refers a patient to a facility with which HealthAmerica does not contract, the insurer would suggest another facility or drug alternative, notes Dawson. HealthAmerica and HealthAssurance have surveyed their physicians on their satisfaction with the insurers’ disease management and case management programs, and reported that 100 percent of the respondents said they were very satisfied or satisfied.

Last October, HealthAmerica started paying rewards in the amount of $25 per quarter for each diabetic patient, available to practices that provide care to at least 150 diabetic members of HealthAmerica, HealthAssurance, and Advantra, and that satisfy standards established by the National Committee for Quality Assurance’s Diabetes Physician Recognition Program. The program is based on the nationwide Bridges to Excellence initiative, which was established by a group of large employers – including Verizon Communications, General Electric Corp., Ford Motor Co., and the United Parcel Service – and HealthAmerica is evaluating the program for possible expansion in the future, says Dawson.

In 2000, HealthAmerica launched a three-year longitudinal study evaluating the impact of its case-managed program for select diabetic members compared to general disease management educational activities for all identified diabetic members. The analysis showed that: (1) the cost for the population in the case-managed program decreased 13 percent and the costs for those not in the case management program increased by 17 percent; (2) the inpatient days decreased by 31 percent for members in the case management program and increased by 22 percent for those not in the case management program; and (3) the average number of hospital days decreased from 7.6 to 3.6 per member per year for the case management program, while the days for the non-case-managed group increased from 1.2 to 1.8 per member per year.

A disease management vendor, McKesson Health Solutions, is contracting with the Pennsylvania Department of Public Welfare (DPW) to provide disease management services to 250,000 individuals – half of all the state’s fee-for-service Medicaid recipients – while the other half are dually eligible for Medicare and will have their medical benefits covered under the new Medicare benefit structure in 2006, according to David Feinberg, deputy secretary of DPW’s office of Medical Assistance Programs.

As of March 1, those 250,000 Medicaid recipients will be covered under the new AccessPlus program, whereby they will be linked with a primary care physician and will be screened by McKesson to determine eligibility for its disease management programs, including asthma, diabetes, high risk pregnancy, CHF and COPD, says Feinberg. The DPW will assess McKesson’s impact on patient outcomes, such as emergency room visits for child asthma, he says. Reimbursement to physicians will increase by 30 percent for office visits by AccessPlus patients, while McKesson will offer additional monetary incentives for physicians to participate in its disease management programs, Feinberg notes.

Health Provider Involvement

While health plan-initiated disease management programs are widespread in Pa., some health systems have their own programs in place.

In the mid-1990s the University of Pennsylvania Health System had made a significant investment in disease management infrastructure, covering 31 programs in its employed primary care physician offices, and fueled by the expected continuation of widespread full-risk contracting to provide a steady reimbursement stream to make the investment economically feasible, according to David Horowitz, M.D., medical director of disease management at Penn, and Penn’s medical director of clinical effectiveness and quality improvement. The infrastructure had three components: (1) physician-based tutorials in the form of literature-supported, online protocols and suggested outcome goals specific to given disease states; (2) re-engineering of physician offices, with administrative and clinical support – e.g., links to physicians’ online patient schedule – to target the disease management population; and (3) direct-to-patient education programs, including nurse call centers to contact chronic disease patients at frequencies commensurate with their severity.

When the market jettisoned full-risk contracting four or five years ago, health systems "failed to be reimbursed to promote wellness," and Penn significantly downsized its disease management programs – which once enrolled over 20,000 patients – to focus on a subset of patients – numbering in the hundreds – who are enrolled in the Medicaid HMO HealthPartners under a risk contract with Penn’s home care entity, says Horowitz. The programs currently function as the nurse call center for asthma, pediatric asthma, heart failure and diabetes patients, although Penn still maintains protocols for all of the 31 disease states of its initial program, and may bring it back into active service if there is a reawakened interest by health insurers in pay-for-performance arrangements, says Horowitz.

As a subset of its quality and process improvement efforts, UPMC Health System has focused on outpatient diabetes disease management and CHF inpatient care management programs, whereby the health system’s 250 community-based and 1,200 practice plan physicians are mandated to use process-of-care protocols adapted from national sources to implement care regimens and to educate patients on self-maintenance techniques with the assistance of trained nurses, according to Helen Chang, MBA, MA, vice president of quality management and executive director for UPMC’s Institute for Performance Improvement. Physician education on diabetes care was delivered throughout UPMC’s physician network via telecommunicated programs for outlying rural practices, and regional meetings and lectures from the academic faculty and community physician leaders in this field.

With the use of data extracted and integrated from multiple legacy systems including EPIC and Quest Lab systems, laboratory tests and physician intervention against national benchmarks, UPMC is able to create a quarterly report card for every physician with a diabetic patient, looking at what percent of a practice’s patients are under and over the benchmarks, how a practice performs over time, and how it compares to other practices within the health system, while a medical quality officer consults with physicians about their performance, says Chang.

While UPMC’s diabetes disease management program has been heavily physician-driven in the past two years, the health system is moving toward more patient-directed care using the HealthTrak patient portal, a Web-based communication tool that allows asynchronous communication between patient and physician, for example, encouraging diabetes patients to keep a journal of their HbA1c and care schedule, notes Chang.

Chang says the diabetes management programs have brought statistically significant improvement to patients’ HbA1c and LDL Cholesterol levels – lower than American Diabetes Association benchmarks, although hospital admission rates for those diabetes patients have not changed, probably because of confounding comorbidities, making it difficult to attribute hospitalization to diabetes alone, says Chang, although she notes that there is strong clinical evidence that lowering HbA1c and LDL decreases risk of other longer-term, higher-cost complications among diabetics.

UPMC’s CHF program identifies heart failure patients in the hospital and tracks them to primary care physicians after discharge. Identifying a population of heart failure patients such as those in DRG 127, over a two year period, Chang says there were dramatic improvements in providing key clinical best practices and decreases in both unadjusted in-hospital mortality and length-of-stay, says Chang. While UPMC believes that active quality-of-care-process management programs of this sort will lower health care costs, Chang says she does not yet have a concrete data set to support that assumption. A greater accumulation of clinical results will make those calculations more meaningful, she adds.

Some disease management programs are physician-driven, as a proactive attempt to demonstrate to health insurance companies that pay-for-performance incentive programs are a mutually beneficial proposition. Under a full-risk arrangement with IBC’s Keystone Health Plan East since 2000, the Renaissance Medical Management Company – a network of about 195 primary care physicians in southeastern Pa. – receives a percentage of premium in return for caring for some 55,000 patients in the region, according to Barry Green, M.D., regional medical director for Renaissance and founding partner of Green & Seidner Family Practice Associates, an eight-physician group in Lansdale.

For two years now, Renaissance has had in place diabetes, CHF and colorectal cancer disease management programs it developed in-house, and has set its level of performance expectations higher than HEDIS standards. By mutually agreeing to follow practice guidelines developed by Renaissance and with the assistance of about 20 nurse practitioners who consult with patients by telephone and about eight outpatient intensivists who meet patients face-to-face, the network is successfully meeting those expectations and participating physicians are receiving a "modest but meaningful" incentive-enhanced payment from Renaissance on top of IBC’s QIPS bonus, Green notes.

About ten percent of the patient population of Green’s practice comes from Keystone, and he says Renaissance hopes to attract a higher market penetration of care based on the pay-for performance model. Since his practice has used Renaissance’s disease management protocols, Green has seen an increase in the percent of his practice’s diabetes patients who are at goal for their health status, and says that his patients are receiving colonoscopy screening at a younger age than previously.

Green notes that, as a Renaissance medical director, he meets with physicians in the network to discuss ways of reducing practice variation according to established guidelines, and provides tools and clinical advice to help practices that might be struggling to meet expectations. Two practices were asked to leave the Renaissance network because they were insufficiently engaged in utilizing the program, he notes.

Whether or not disease management programs will save money in the long run, Green maintains that the programs are right to emphasize preventive care rather than medical treatment of illness. "Doing the right thing supercedes costs," says Green, who notes that costs can be expected to go up initially because of higher levels of preventive care, while lower treatment costs can be expected down the line. "There is very little in this marketplace to feel good about," notes Green, while he says an effective focus on quality initiatives like disease management programs – with a modest compensation incentive for doing so – is something physicians can and should feel good about.

Long-Term Impact Uncertain

Despite the growth of disease management programs in Pa., and some evidence by their supporters that programs improve health status and save money in the short term, the capacity for such programs to sustain such effects over the long term is unknown. The level of evidence that gives the commercial managed care market confidence in the value of disease management is not enough for large government insurance programs, for which a long-term cost horizon is a crucial calculation. In that vein, a recent Congressional Budget Office report concluded that there is currently insufficient evidence that disease management programs lower health care costs enough to warrant putting such programs into the core benefit structures of Medicare or Medicaid.

In an October 13, 2004 letter to the U.S. Senate, the CBO wrote: "The proposition that decreased use of acute care services might offset the costs of the screening, monitoring, and educational services in disease management programs is clearly appealing, but, unfortunately, much of the literature on those programs does not directly address health care costs. Instead, the focus is often on the processes of care or on intermediate measures of health, from which an overall impact on spending cannot reasonably be inferred85. [W]hile a few studies indicate that disease management programs could be designed to reduce overall health costs for select groups of patients (at least in the short term), little research directly addresses the issues that would arise in applying disease management to the older and sicker Medicare population."

The CBO said it will monitor a set of large-scale disease management pilot projects, soon to be launched by Medicare to assess more systematically their impact on cost and quality, and expected to include 10 to 12 sites around the country, covering 20,000 or more patients each.

Randomized control group studies of disease management intervention are extremely scarce, while the quasi-scientific studies that characterize most disease management outcomes literature do not establish with scientific rigor the prospect of long-term health and cost improvement, according to Autumn Dawn Galbreath, M.D., MBA, vice chair for clinical programs in the department of medicine of the South Texas Veterans Health Care System, and former project director of the University of Texas Health Science Center’s Disease Management Center.

Galbreath recently co-authored a study of about 1,100 CHF patients, which she says is one of the first that had no industry funding and had a true control group design, and found that telephonic, nurse-driven disease management improved mortality – patients lived an average of two-and-a-half months longer during the study follow-up period than those who did not receive disease management – but did not reduce the costs of hospitalization, procedures and overall care. Galbreath says that another disease management study that used a randomized control group design actually showed that disease management increased costs.

She also notes that true control group studies are difficult to conduct because of equity issues that arise if a health plan is seen as "denying care" to a control group of insureds who do not receive disease management intervention. Galbreath’s study was conducted under a contract with the U.S. Department of Defense and focused on Medicare, VA, military, and unfunded beneficiaries.

Conclusions drawn by previous disease management studies are further weakened by crude clinical and cost indexes, typically hospital readmission rate and number of hospital bed days, Galbreath argues. Claims data are typically used to track outcomes, which Galbreath says are a less accurate source of outcome data than actual medical record review, which is much more expensive and time-consuming. Galbreath’s study used clinical outcome measures such as mortality, functional capacity with severity designation, and exercise tolerance level; while cost measures included health care utilization and expenditure with severity level of hospitalization, outpatient visits and prescription drug costs.

"The best-designed studies either show no change in cost, increased cost, or they only address clinical outcomes and not economic outcomes," says Galbreath, who speculates that managed care may have already trimmed most excess utilization from the health care system, and that more frequent contact by the sick elderly with a health care provider – fostered by disease management interventions – can lead to the discovery that more care is needed for them. "You can’t squeeze the care out of people who need it," Galbreath notes. The central assumption that disease management can reduce costs by reducing practice variation needs to be better assessed, and is difficult to tease out through outcomes studies, Galbreath adds.

The status of disease management outcomes assessment is improving, according to Karen Fitzner, Ph.D., director of research and program development for the Disease Management Association of America, which is currently evaluating guidelines for robust research of disease management programs, and hopes to encourage more physician participation in such research. A literature review of the most recent studies shows that disease management definitely improves outcomes, although financial outcomes vary, and that patient satisfaction in such programs tends to be very high, while physician satisfaction varies widely, she says.

"One should not just rattle off outcomes results, whether they are clinical or financial or humanistic, because they almost always relate to a specific program, with specific interventions, on a specific population, at a specific time, for a particular disease state, within a specific benefit structure. Once there has been sufficient generalizability of the findings to other populations, diseases and times, then it will be possible to talk about outcomes in a broad sense," says Fiztner.