Comparative Effectiveness Research: A Cornerstone of Healthcare Reform?

Abstract

Comparative Effectiveness Research (CER) has recently emerged as a major theme in health care reform. Unfortunately, there is a widespread lack of understanding about what it will do and fear that it will do more harm than good. These concerns include threats to individual physician's autonomy and professionalism, as well as fears that care will be rationed based on such findings. In this paper, we argue that the main components of the current healthcare reform (HCR) bills, which include expanding insurance while increasing efficiencies through cost containment, should embrace CER. This type of research will provide a “safeguard” against “blind” cost-containment, so that the new financial incentives being introduced can be actualized effectively and safely. Evidence for this is provided from examples from the authors' prior and current research as well as from the literature. We also argue that the requirement for data from CER will create long-term disincentives for “me-too” drugs and devices and, therefore, become a catalyst for effective innovation.

Comparative Effectiveness Research (CER) has recently emerged as a major theme in the health policy arena and as a component of the health care reform (HCR) legislation of 2009/2010. Its purpose is to help determine what individual tests and treatments work best and to assess ways of improving the delivery of healthcare.

This type of research has existed for a long time and has been known by various names, such as health services research, clinical epidemiology or outcomes research. However, there are certain aspects that have emerged as most important within the context of HCR with its emphasis on expanding insurance coverage, health system and reimbursement changes, disease management, and health information technology (1). CER has been suggested as a part of the strategy in healthcare reform for improving outcomes and helping to curb the growth in costs.

Although CER is included in one way or another in all of the current HCR proposals, there is a widespread lack of understanding about what it will do, and fear that relying on CER will do more harm than good. These concerns include threats to individual physicians' autonomy and professionalism, as well as fears that care will be rationed based on such findings.

In this paper, we first describe, in broad strokes, the components of the current HCR bills. We then review the distinctive elements of CER which differentiate it from the research currently required for approval of new drugs and devices. We will illustrate some of the methods for CER, as well as its potential value, through examples from the first author's prior and both authors' current work. Finally, we will make the argument that CER should be viewed by the medical community and by academic medicine, on net, as an asset and not principally as a threat of rationing, of cookbook medicine, or to the doctor-patient relationship.

COMPONENTS OF HEALTH CARE REFORM

Table 1 displays the key elements of HCR contained in the current proposals (1). Insurance coverage would be expanded to increase the number of people insured through mandates for coverage and broadening eligibility for group purchasing and assistance; there would be regulations on insurance companies, such as prohibiting exclusions of individuals with prior illness and guaranteeing transportability of coverage. Payment rules would reward quality while creating incentives to decrease utilization by bundled payments for certain diseases or conditions, and penalties for events such as readmissions for conditions such as congestive heart failure after discharge. In addition, HCR would fund computerization and electronic medical records to increase efficiencies in the system. Reform would also create changes in health care delivery systems (first by experiments) to incentivize disease management programs and to establish comprehensive care models such as the reorganization of practices to create “medical homes”. Finally, some of the proposals also included tort/malpractice reform which should result in reductions in defensive medicine.

TABLE 1.

Components of Health Care Reform

Change Insurance Coverage	Individual Mandate
	Employer Mandate
	Purchasing Tools
	Refundable Tax Credit
	Medicaid/SCHIP Eligibility
	Open Access to FEHBP
Change Payment Rules	Physician Pay for Performance
	Hospital Pay for Performance
	Bundled Payment
Change Health Services Delivery	Disease Management
	Health Information Technology
Change Legal Environment	Medical Malpractice Reform
Comparative Effectiveness Research	New agency or AHRQ

Thus, the driving forces behind HCR are expanding insurance in order to increase access to health care, while, at the same time, laying the groundwork to slow the rise in health services expenditures. For insurance to remain affordable by both employers and individuals, and thereby for expanded access made possible by HCR to remain viable, incentives for reductions in the utilization of expensive new and complex technologies will be needed to achieve efficiencies and reduce cost. And well they should, if our country is going to have a chance to provide medical care to all.

CER: THE “SHARPER TOOL” FOR CHANGE IN HCR

Many entities in the US are already involved in CER activities. Examples include the work of the Effective Healthcare Program currently ongoing at the Agency for Healthcare Research and Quality (AHRQ), as well as the Centers for Medicare and Medicaid Services, the Veterans Administration, the National Institutes of Health (NIH), and the Office of the National Coordinator for Health Information Technology (2). Yet, it was only when the federal government substantially increased the investment in CER through the American Recovery and Reinvestment Act (2009), that national interest in this topic emerged. Congress allocated 1.1 billion dollars to the NIH, the Department of Health and Human Services and AHRQ to be spent over a 2-year period to initiate and disseminate CER.

Virtually every western country that has reformed its health care delivery system has instituted one form or another of comparative effectiveness research. The most famous is the National Institute for Clinical Excellence (NICE) in the United Kingdom. CER is used by NICE to decide which technologies, drugs or devices the National Health Services will provide and pay for. In other countries such as Canada, although used for policy decisions, the link is less direct (3).

As mentioned, in the United States, cost containment resulting from health care reform will be driven by financial incentives, programmatic initiatives and organizational changes. However, this does not mean or guarantee that individual clinical decisions will be in the patient's best interest or that, in aggregate, economic policy changes and incentives will lead to enhancing the publics' health. Conversely, these forces should be thought of as relatively “blunt instruments” which nudge the health care system toward efficiency without precise direction. They must be accompanied by “sharper tools” which inform change and guarantee, to the extent of medicine's ability, that the health of the population is enhanced, not diminished, by the changes. CER clearly stands out as one of these “sharper tools”. In the following sections, we describe how CER fits in as an integral component of HCR.

THE FUNDAMENTALS OF CER

A number of definitions for CER have been proposed in the medical literature (4). However, the most widely accepted is that of the Institute of Medicine (IOM) committee on comparative effectiveness research, which was commissioned by Congress and which released its report in June 2009 (5). According to the report, CER is defined as: “the generation and synthesis of evidence that compares the benefits and harms of alternative methods to prevent, diagnose, treat, and monitor a clinical condition or to improve the delivery of care. The purpose of CER is to assist consumers, clinicians, purchasers, and policy makers to make informed decisions that will improve health care at both the individual and population levels.” The IOM report emphasized that CER research should focus on selected clinical areas and aspects of the health care delivery system. The report identified 100 study topics related to a range of diseases and systems that are of most importance to the health of the U.S. population (6).

In addition, the report emphasized that attention and resources would be needed, not only to stimulate CER, but also to develop relevant research methods to conduct this type of research in valid and reliable ways. First, CER research must involve direct comparison of already known to be, and thus established, efficacious interventions to show the real marginal benefit of the proposed intervention. To do so, CER studies must include a valid comparator treatment that represents the current standard of care. One challenge that these comparisons will have to overcome is the anticipated small difference between treatment effects of the interventions that are currently used and the new one. Studies that compare a treatment to placebo have little value from a CER perspective despite showing potentially large treatment effects. Second, a CER study should include patients that are seen within the context of day-to-day clinical care and that, therefore, represent the real world of healthcare. The study should not be restricted to a select subset of patients who fit standards for inclusions and exclusion in a standard clinical trial, since these standards may be too restrictive to adequately represent the target population for such a drug or device. Finally, the study should attempt to account for the heterogeneity in treatment effects across the spectrum of individuals, and thereby, to the extent possible, provide information for decisions tailored to the characteristics of individual patients.

To address these issues, the IOM report emphasized that CER research will often need to rely on observational data that are available in the form of large claims data sets, registries or electronic health records. Analytic plans should use advanced statistical models that address biases that arise when using observational data. When RCT's are required, studies should employ, to the extent possible, pragmatic or practical clinical trials designs rather than traditional ones.

EXAMPLES OF CER: REFLECTIONS ON PRIOR AND CURRENT WORK

The comparative effectiveness of therapies: implantable cardiac defibrillators

Implantable cardiac defibrillators (ICDs) are a relatively new technology that has been controversial with regard to whether it is advantageous over existing medical treatments. This device's ultimate place in clinical care and decision making has been uncertain.

The Multicenter Automatic Defibrillator Intervention Trial (MADIT I) was the first randomized trial of an ICD for primary prevention of cardiac death in individuals at risk of ventricular tachycardia or fibrillation. Patients who had had a myocardial infarction and who had documented non sustained ventricular arrhythmias not responsive to medications when tested in the electrophysiology lab were randomized to receive an ICD or to be placed on ongoing medical management as part of their usual care. Patients were followed prospectively and called or seen every month in order to determine end points of mortality and cardiac events. Also, detailed health care utilization information was collected in order to evaluate the economic consequences of using an ICD as compared to medical management (7).

The data were analyzed according to intention to treat criteria via Kaplan Meier survival curves, with the main endpoint being all-cause mortality. The left panel of Figure 1 shows that the ICD was associated with a sizable survival benefit with a hazard ratio of 0.48. This research was followed by MADIT II, in which the criteria for an ICD were expanded to a population of patients with CHF post-MI (EF <30%) but who had no evidence of ventricular arrhythmias. The results of this trial, as shown in the right panel of Figure 1, were less dramatic than MADIT I but still indicated a benefit for the ICD over medical therapy (8).

Fig. 1.

Survival of patients treated with ICDs versus medical therapy (CONV) for prevention of ventricular arrhythmias.

The economic data were analyzed in conjunction with the effectiveness information in order to calculate the incremental cost-effectiveness ratio (iCER) associated with the use of ICDs in these two clinical situations. For the higher-risk population in MADIT I, the iCER was approximately $27,000 per year of life gained, whereas the iCER for the patients in MADIT II was considerably higher, at but still below a $100,000 threshold (9, 10).

These studies are illustrative of the way CER can be used to inform clinical health policy of a new therapy. They evaluated currently accepted treatments as well as a comparator alternative. Although they included stringent inclusion and exclusion criteria and were conducted in academic health centers, there was an attempt to recruit from a wide referral base and to maintain real world conditions. Importantly, they made a special effort to catalog the health care consequences of each alternative and to enumerate the resultant costs. At the time of these studies, there was significant controversy not only about whether ICDs worked but also about how they should be used clinically. There was also considerable concern about their cost and whether insurance companies including Medicare should pay for this technology.

The information from the trials was critically useful in delineating the clinical indications for the use of ICDs when they first came out and then to guide their extension beyond the first clinical indications to additional populations of patients. Furthermore, although ICDs are expensive, the research showed that, if used appropriately, there was a reasonable return on the investment in terms of the incremental survival achieved, particularly for the higher risk group represented by MADIT I. It was, thereby, not only useful for clinical decision making, but also to argue for the inclusion of ICDs in health insurance benefit packages and to delineate the clinical characteristics that would justify the use of an expensive new technology.

The comparative effectiveness of diagnostic tests: suspected coronary artery disease

The optimal strategy for diagnosing coronary artery disease (CAD) in patients with non-specific chest pain is unclear. Most studies of test accuracy have used coronary angiography as the gold standard, which may underestimate the prevalence of the disease. The objective of the Coronary Artery Disease Diagnosis and Outcomes Study (CADDO) was to determine the optimal exercise stress test for diagnosing CAD in symptomatic women and men. It included a prospective cohort study of patients with chest pain or other symptoms suggestive of CAD. A consecutive series of patients were referred from 44 primary care practices located in the Rochester, NY area. Each patient underwent exercise stress testing with electrocardiography (ECG), echocardiography (ECHO) and sestamibi (MIBI) perfusion imaging at the same time. Patients were followed for 3 years for physician-reported cardiac events, including stable and unstable angina. The study compared the result of each test (ischemia, yes or no) to 2 different gold standards at 3 years of follow-up: cardiac events and the clinical assessment of cardiologists (informed by knowledge of cardiac events but not test results).

Included were 349 patients (185 women and 164 men). At 3 years of follow-up, 10% had experienced cardiac events and 33% were classified by cardiologists as having CAD. Using cardiac events as the gold standard, the sensitivity and specificity of the tests were compared, as shown in Figure 2 for men. The combination of sensitivity and specificity as represented by the Youden's index was highest for MIBI, but the differences were small. Using cardiologists' clinical judgment as the gold standard, the relative performance of these tests persisted and specificity was similar, but sensitivity decreased for both women and men (11).

Fig. 2.

Comparison of three diagnostic test for coronary artery disease in men.

These findings suggest that although MIBI may be a slightly more accurate first-line test for patients with suspected CAD, that the differences between all three tests are quite small, and that whereas the sensitivity may be higher for one, it may not have more specificity. The preferred test, therefore, will depend on whether one wants to maximize sensitivity or specificity. From a CER perspective then, these data could be used to argue that less expensive testing may be justified, at least in sizable proportion of patients, and thereby, to guide and support a clinically rational cost-containment strategy.

Comparative effectiveness observational studies using registries: total joint replacements

The Hospital for Special Surgery (HSS) Total Joint Replacement (TJR) registry was established as a result of an ongoing collaboration between the Department of Public Health at Weill Cornell Medical College and HSS and is funded by a Center for Education & Research on Therapeutics (CERT) grant from the Agency for Healthcare Research and Quality. The registry, which was started in May of 2007, has enrolled more than 10,000 patients to date. Enrolled patients complete a baseline battery of surveys preoperatively [demographics, SF-36 (12–14), WOMAC (15,16), Lower Extremity Activity Scale (17) (LEAS), Expectations (18–20), and EuroQ EQ-5D (21)], and complete follow-up surveys at 6 months (focusing on self-reported post-surgical complications), 2 years and 5 years (overall satisfaction (22), SF-36, WOMAC, LEAS, and Euroqol) after surgery. In addition, subjects consent to data collection from medical chart abstractions conducted at HSS (mainly for implant, procedure and co morbidity information). Current HSS TJA volume is between 5000 and 6000 cases per year. Approximately 85% of patients undergoing these operations have consented to enroll in the registry. The follow-up survey completion rate at 6 months is approximately 80%. Two-year follow-up has already started. Research Assistants enter survey information into a secure web-based database system (SCTR, Look Left Group, New York, NY) which has been developed specifically to support the registry data activities. The SCTR system is designed to interact automatically with other databases in the hospital. To assure quality of data collection and entry, validation checks occur on a regular basis.

The registry provides an excellent vehicle for comparative effectiveness research in orthopedics. More than 100 hip and knee replacement prosthesis designs are available on the market today for patients considering total joint replacement. The designs vary in shape, material, size, and cost (from $600 to $15,000). Yet there is little evidence comparing the effectiveness of these prostheses. In fact, if available at all, evidence is limited to comparing prosthesis with regard to revision rates. Much less is known about prostheses in terms of their “functional” effectiveness.

One main purpose of this registry is to compare the effectiveness of the 5 most used total hip replacement prosthesis designs in terms of their 2-year functional outcomes. With such a large volume, that is not otherwise available, the registry is enabling researchers to study the factors affecting TJA outcomes. For example, outcomes data on complications (Table 2) can be used to compare the effectiveness of different implant designs for a total hip replacement despite the infrequent occurrence of such complications.

TABLE 2.

Adverse outcomes at 6 months of registry total joint replacement procedures (May 1, 2007–July 31, 2008)

Complication	Hip (n)	Hip (%)	Knee (n)	Knee (%)
Pulmonary embolus	8	0.3%	24	1.0%
Deep vein thrombosis	28	1.1%	67	2.8%
Infection around joint	29	1.1%	91	3.8%
Dislocation	56	2.2%	11	0.5%
Fracture of or around joint	17	0.7%	24	1.0%
Pneumonia	17	0.7%	23	1.0%
Stroke	6	0.2%	7	0.3%
Myocardial infarction	6	0.2%	8	0.3%

The patients enrolled in the total hip replacement registry are identified from a database that includes all total hip replacement procedures. Baseline demographic and functional status information (mentioned in the previous section) are merged with comorbidites and in-hospital complications downloaded from medical record sources. The database is then linked to a device table that has been developed specifically to identify the type and manufacturer of the prosthesis used. This table is then used to distinguish the different prosthesis types to be compared. Patients identified at baseline are followed with mailed questionnaires to assess functional status and to inquire about complications at 6 months and then again at 2 years.

Data available from patient registries such as this include populations that are not represented in randomized clinical trials, which should make the results of registry-based studies more generalizable and applicable to clinical practice. Yet, unlike randomized clinical trials, observational research has several inherent limitations that require attention. Like other cohort studies, patients are not randomized to each of the prosthesis designs. The implanted prosthesis depends on multiple factors that include, but are not limited to, the experience of the surgeon with the implanted prosthesis and the requests of the patients. As a result, observed differences between function at 2 years may be susceptible to selection bias, which must be addressed by using statistical methods such as instrumental variables and propensity scores.

POTENTIAL ROLES FOR CER

As we have described, health care reform (HCR) will eventually pit the goals of expanding health insurance coverage against strong pressures to reduce the growth in health care costs. If left to measures in the proposed legislation, cost-containment resulting from health care reform will rely primarily on marketplace incentives, programmatic initiatives, and organizational changes to offset partially the costs of expanding coverage (1). Unless we want our health care system to be driven solely by financial and regulatory incentives, CER should become a cornerstone of HCR that should be embraced, rather than resisted, by the clinical and public health community.

The findings from CER should provide a buffer against “blind” cost containment, preserving what is of value in our current system, even if the items are expensive. As we described above, expensive implantable cardioverter defibrillators (ICDs) for certain categories of patients, have been shown to be more effective than medications for some patients post-MI with ventricular arrhythmias. Additionally, when used selectively and carefully, they are an efficient use of resources. Based on information from CER studies, ICDs should be provided for selected indications.

In our era in which high-tech care is generally favored, there is little likelihood that a less complex and cheaper therapy will emerge as preferred unless there are data to suggest such. However, CER can lead to such insights. For example, cheap diuretic medications have been demonstrated by the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), to be more effective than to Angiotensin Converting Enzyme (ACE) inhibitors, calcium channel blockers and alpha blockers for treatment of hypertension (23). We will need to rely on insights such as this to suggest ways in which the new financial incentives being introduced can be actualized effectively and safely. The study of diagnostic tests for suspected coronary artery disease we outline above provides another example, since the results support the use of more basic tests for the initial evaluations of patients with non-emergent and ongoing chest pain.

Although we have not included such evidence in this paper, CER should improve medicine in ways other than its direct effects on clinical decisions. The reliance on, and perhaps the requirement for, data from CER will create disincentives for “me-too” therapies. It should become a catalyst for effective innovation. The presence of CER should make it clear to pharmaceutical companies and device manufacturers that any new product will not be accepted in the medical market place unless it is also proven to be superior to currently available therapies or diagnostic devices. Once the medical community begins to expect and demand clear evidence of superiority, manufacturers will be forced to devote more of their resources to the discovery of true break-throughs. Furthermore, if the emphasis is also on cost-reduction, then there is an incentive to do the research and development that will bring to the marketplace not only superior therapeutics, but equivalent interventions that are cheaper and, therefore, more cost effective. Clement and colleagues (2009) recently illuminated the process of innovative drug introduction in 3 countries (UK, Australia and Canada) that already use comparative effectiveness findings to evaluate and determine coverage (24). Only drugs with strong evidence were approved upon first submission. For example, all these countries rapidly approved Ranibizumab, which is indicated for age-related macular degeneration, because of the strong supporting evidence from randomized trials of superiority to alternatives. On the other hand, Insulin Glargine, which has only slight advantages over Insulin NPH, was resubmitted 5 times in Australia before being approved at a much lower price than originally proposed, and NICE in the UK approved the drug only for a subset of patients: type I and a small niche of type II patients who might be more likely to benefit. At present incentives for such decisions are absent here.

An increasing reliance on CER for health policy in the United States may have an even greater impact. The US is by far the largest market for drugs, and new drugs are almost always approved and released here before they are released in the rest of the world. Countries that do not have additional CER requirements frequently approve what the FDA approves (i.e., “me-too” as well as breakthrough drugs). If CER is successful in shifting the focus here toward the development of drugs and devices that really improve outcomes, fewer investigational new drug applications will be reviewed by the FDA. This should shorten the approval time for drugs to enter the US market, thus additionally benefitting everyone.

CONCLUSIONS

On balance, the medical and public health community should be enthusiastic about CER and support it. It is physicians' first line of defense against blind cost containment and a way for medical professions to responsibly provide input and help with this task. On the basis of CER our drug and device industry will be encouraged to produce products that really matter. Research of this type will require that our medical research establishments fully engage in the creation of the information needed and that resources are available to do so. CER requires, almost by definition, that physician scientists play key roles, and thereby opportunities for physicians to lead such research will increase. The final result should be that important medical decisions will be guided and influenced by the scientific community, not solely the capricious nature of the marketplace.

Rather than, “will it have a role” we should be asking “how best to include” the evidence from CER into clinical health policy and coverage decisions as we move forward. We should also determine how to organize and support the medical and research establishments to create the best evidence about what works in health care and whether one approach is better than another. Finally, we should seize the opportunity that CER offers to rightfully expand the investment in research that will make a difference in patient care and improve the outcomes of our health care system.

DISCUSSION

Powe, San Francisco: Al, thanks for a comprehensive overview of where comparative effectiveness research is going. I have a question related to where you think this will best be placed in terms of who organizes this. My understanding, from the bills currently circulating in the House and in the Senate, is much like the discussion of the public and the private insurance option. In fact, the House bills place the organizational unit for overseeing comparative effectiveness within the federal government within the Department of Health and Human Services, probably either the NIH or the Agency for Healthcare Research and Quality, whereas the Senate bills place them in, or as you said this, private non-profit units outside the federal government, which would bring with it input from a different set of stakeholders. So my question is, have you thought about where the best locus is in terms of objectivity, access to all the stakeholders that need to be involved in this process and access, importantly, to academic institutions, like researchers like yourself, who are interested in doing that research?

Mushlin, New York: Neil, that's a great question. First of all, I hesitate to say exactly what I think would be best. I think that there are probably a number of different valid opinions, and I can see pluses and minuses of all of the configurations that are being thought of. I think you did a really good job of explaining a principal difference with the current options, at least that I've heard of. I guess I will venture a little speculation of my own. I mentioned it when I referred to the fact that this type of research has been known by other names before: health services research, outcomes research and so forth. This type of investigation is finally becoming mainstream in academic medicine, and I think to all of our benefits. Therefore, I would very much like to see comparative effectiveness research placed in the federal government in such a way that the links with academic medicine are made as robust as possible. I guess that would suggest that this research be placed in a more public, rather than a more private organization. If done right it could also call for and lead to a closer link between the NIH, for example, and the Agency for Healthcare Research and Quality. So I think it's a very important question. It's the second most important question next to how much new research money is going to be provided for comparative effectiveness. The second one close behind this is how the organization for this research will be structured. It's going to be very important to express our own views about this and to watch what happens in the next few months in both the Senate and the House.

Ludmerer, St. Louis: I want to thank you for that presentation and commend you on your work. It is important work. This is an important activity. I wanted to make an observation linking what you are saying now to the discussion yesterday from Holly Humphrey and education. In addition to comparative effectiveness, the problem we face is soaring costs. I would submit there is another element that is right there at the center and typically goes undiscussed, and that's the thoughtless practice of medicine, the decerebrate practice of medicine. Recently I have been attending on the medical service at Barnes Hospital. We received a patient who had been in the emergency room for close to 24 hours with a complaint of abdominal pain. During this time the patient had every conceivable blood test, which were all negative, had five imaging tests, all of which were paid for by Medicare, CT of the chest, pelvis, abdomen, a renal ultrasound and a right upper quadrant ultrasound, all negative, sent to us for evaluation. Well, it turns out that no one during this 24 hours listened to the patient. The problem was not in the abdomen. The pain was in the scrotum, and when it became very severe, it radiated into the abdomen. No one during this time had done a GU examination. He had acute epididymitis and suffered needlessly for 24 hours without treatment, not to mention the extraordinary cost that Medicare paid for because this was done in the emergency room. I think this illustrates in the microcosm experiences that all of us have and the challenges we face practicing medicine in a thoughtful fashion, which not only improves quality of care but is a powerful tool toward combating this culture of excess. We have not been doing things because they are needed, but because they are there, and this effects costs profoundly. This, I think, is the link with Holly Humphrey and her presentation yesterday. What excites me most about the University of Chicago approach and defending basic educational principles is that it deals with the development of people as physicians. It is important not only to produce pinpoint medical knowledge but to import clinical effectiveness so people are doctors, not just individuals with medical knowledge. My personal prediction is that one of two things is going to happen. Either we will get back and reaffirm basic principles and teach medicine and practice medicine in a thoughtful fashion, thereby improving quality and cost, or we are not going to do it, and it is going to be done for us, and ultimately we will have some type of rationing.

Mushlin, New York: Absolutely, and I totally agree with your last point. I think comparative effectiveness research is at least one tangible way of re-engaging the profession and avoiding the over dependence on “blunt” financial incentives to change healthcare delivery.