Kim Farina, PhD

The widespread adoption of electronic health records is facilitating implementation of quality improvement measures across healthcare. As the US healthcare system becomes more quality focused, current methods of defining and measuring quality have come under scrutiny. Diabetes has been a major focus of early quality improvement initiatives. The speakers of a symposium titled “What Are the Ideal Quality Measures for Diabetes Care?” at the 72nd Scientific Sessions of the American Diabetes Association discussed the current state of quality measurement in diabetes.

“It is impossible to know how well we are doing without measuring something,” said Sandeep Vijan, MD, MS, associate professor of internal medicine, University of Michigan.“ There is little doubt that quality measurement is important and reasonably effective.” He explained that systems do better in those areas that are measured and reported, with financial incentives clearly driving behavior. “What you measure becomes a driver of what you do,” said Dr Vijan.

Quality measures may take the form of process measures of whether an action was performed (eg, glycated hemoglobin [A1C] measurement or eye examination), intermediate outcome measures (eg, A1C <7.0), or outcome measures (eg, hospital admissions for myocardial infarction). Defining quality is a significant challenge and prone to a number of problems.

Thus far, quality measures have generally been based on interpretations of trial evidence by clinical experts. This method poses a problem because the targets for continuous measures (eg, A1C and blood pressure) are based on mean clinical trial results. “If the mean A1C of a trial is 7.0%, then 50% of patients in the trial received less than optimal clinical care,” stated Dr Vijan. Furthermore, intermediate outcomes are not always tightly linked to patient-important outcomes.

Often, quality measures are falsely dichotomized. For example, if A1C was set as a therapeutic target, a measurement of 7.1% would be equally unacceptable as a measurement of 13%. Dr Vijan raised concern that this type of quality measure will drive treatment of patients with milder elevations—who are more likely to reach target&mdash;rather than those with severe elevations. Without considering each patient’s underlying risk profi le, he said, “These types of measures provide a potential incentive to overtreat and possibly harm patients.”

After reviewing some of the problems with existing quality measures, Dr Vijan spoke about new efforts to improve the science of quality measurement. According to him, quality measures should be based on rigorous interpretation of the evidence, incorporate reasonable exceptions, reward appropriate actions, consider how much benefit is gained, incorporate patient preference, and consider benefit and value. To date, patient involvement in defining and measuring quality has been very limited.

One potential improvement on the dichotomous quality measure is the linked action measure, also called a clinical action measure.

 Linked action measures give credit for appropriate actions. For example, for a patient who is above goal (A1C >8 or systolic blood pressure >140), the provider would get credit if new medication is added, existing dose is increased, or the therapeutic target is achieved within a certain time frame. This measure allows inclusion of reasonable exceptions. Linked action measures have already been developed, specified, and tested. Dr Vijan shared published and unpublished data demonstrating his institution’s successful use of electronic medical records and linked action measures: 85% of providers from the system of more than 1 million patients were taking appropriate action or maintained patients at therapeutic target.

The next step, explained Dr Vijan, is to develop and test patient-selected, personalized measures of quality. To exemplify the importance of patient preference in measuring quality, he presented data from a cost-effectiveness analysis of intensive glucose control among patients 65 years or older that considered patient preference.

 Intensive glycemic control has been shown to be cost-effective among patients younger than 50 years. The results of this study suggest that among older patients, factoring in their patient preferences decreases quality-adjusted life-years of intensive glycemic control with insulin and makes treatment more harmful than beneficial. Factoring in patient preference regarding oral agents reduced cost-effectiveness. This type of study demonstrates how significantly patient experience can influence value.

Measuring Quality of Care for Patients With Multiple Morbidities

Cynthia Boyd, MD, MPH, associate professor of medicine, division of geriatric medicine and gerontology, Johns Hopkins University School of Medicine, presented “Measuring Quality of Care in Complex Patients.” Multimorbidity is common—43% of Medicare enrollees 65 years or older have 3 or more chronic conditions.

 Decisions for patients with multimorbidity are complex. Most clinical practice guidelines are developed from a single-disease perspective, making them difficult to apply to patients with complex medical histories.

 Dr Boyd discussed how clinicians need guidelines that help them navigate the treatment of patients living with multiple diseases or chronic conditions. Boyd explained that to best care for people with multimorbidities, clinicians need guidance on how to: maximize use of therapies likely to benefi t patients with multimorbidity, minimize use of therapies unlikely to benefit or likely to harm patients with multimorbidity, and incorporate patient preferences and values regarding burdens, risks, and benefits.

The National Quality Forum has engaged its Consensus Development Process (CDP) toward a measurement framework that will assess the efficiency of care (defined as quality and cost) provided to individuals with multiple chronic conditions (MCCs).

 The framework defines MCCs, identifies high-leverage measurement areas for the population with MCCs (to mitigate unintended consequences and measurement burden), presents a conceptual model that serves as an organizing structure for identifying and prioritizing quality measures, and offers guidance on methodological and practical measurement issues.

The framework is centered on patient and family goals and preferences for care and will allow for patient prioritization of disease-specific measures based on individual needs, preferences, and discussions with healthcare providers. Non—disease-specific measures that apply to all patients regardless of condition also factor into the model. Measurements fall under specific domains that are aligned with priorities of the US Department of Health and Human Services National Quality Strategy: person- and family-centered care, health and well-being, patient safety, effective communication and care, effective prevention and treatment, and affordable care.

Patient-Centered Outcomes Research Institute

In recognition of the need to integrate patient preferences into the planning and execution of health research, the American Recovery and Reinvestment Act implemented the Patient-Centered Outcomes Research Institute (PCORI). According to its mission statement, PCORI “helps people make informed healthcare decisions—and improves healthcare delivery and outcomes&mdash;by producing and promoting high integrity, evidence-based information&mdash;that comes from research 

, caregivers, and the broader health care community.”

In May 2012, PCORI issued its National Priorities and Research Agenda, which lists 5 National Priorities: assessment of options for prevention, diagnosis, and treatment; improvement of healthcare systems; research on communication and dissemination; reduction of disparities; and acceleration of PCOR and methodological research (

 That same month, PCORI issued its first 4 funding announcements.

Through its patient-driven research funding programs, PCORI aims to align research questions and methods with patient needs. According to Joe Selby, MD, MPH, of PCORI, who presented “Insights on Quality of Care From Institutional Decision Makers,” PCORI stipulates patient and stakeholder engagement from the earliest phases of study design through the final stages of research to gain better understanding of the choices patients face, ensure that all elements of study design are truly relevant, keep research on track, and avoid changes in course that undermine the utility of the study to patients. Results of PCORI-funded studies will be disseminated to provide patients and providers with supplementary information to improve decision making.

Articles

Conference coverage of the 2012 meeting of the American Diabetes Association.

New Approaches to Defining and Measuring the Quality of Healthcare

Applying Public Comparative Effectiveness Research: How States Are Facing the Challenges

The comparative effectiveness research (CER) landscape has ballooned over the last decade (

). In 2005, the Agency for Healthcare Research and Quality (AHRQ) launched the Effective Health Care Program, in part, to produce, translate, and disseminate CER. The 2009 American Recovery and Reinvestment Act expanded government spending on CER to $1.1 billion over a 2-year period. Subsequently, the passage of the Affordable Care Act established the Patient-Centered Outcomes Research Institute (PCORI) and earmarked $6 billion for CER over the next 10 years.

A variety of different stakeholders turn to CER to meet their distinct needs. Patients and clinicians see CER as a way to inform their decision making with clinicians, and payers see CER as a resource for their coverage and reimbursement decisions.

Two primary providers of public CER to date have been AHRQ and the Drug Effectiveness Review Project (DERP). In 2011, more than 1300 CER reports were published (

). However, despite the volume and breadth of CER that exists and the number of potential users, no clear mechanism exists to ensure that the information is being translated and communicated in a way that actually meets stakeholders’ needs.

At the 2012 annual meeting of the Academy of Managed Care Pharmacy, Jenny Gaffney, senior manager, Avalere Health, discussed the challenges faced by state Medicaid programs in using public CER. Steven Pearson, MD, founder and president, Institute for Clinical and Economic Review (ICER) at Massachusetts General Hospital and Harvard Medical School, described a model that is enabling several New England states to overcome these barriers.

Last year Avalere and the Kaiser Family Foundation (KFF) initiated a study to look at how state Medicaid programs are using CER to determine their pharmacy benefits. “Now, more than ever,” said Gaffney, “Medicaid programs face a delicate balance in providing quality healthcare access to vulnerable individuals, while also being prudent with taxpayer dollars.” The report of their findings, 

The Role of Clinical and Cost Information in Medicaid Pharmacy Decisions: Experience in Seven States

, was published in September 2011. Their study involved primary and secondary research on 7 state Medicaid programs (Florida, Minnesota, Nevada, Washington, Maryland, Louisiana, and Massachusetts) and 3 Medicaid managed care plans (Amerigroup, Molina, and United Healthcare).

Pharmacy directors from both Medicaid fee-for-service and managed care plans prioritize clinical effectiveness in their preferred drug list decision making. All too often, however, evidence is not suffi cient for determining the equivalency or superiority of one drug versus another in terms of efficacy or safety.

In these cases, Medicaid programs default to comparisons of drug price to make their ultimate coverage decision. One state pharmacy director interviewed said, “We often have to line up the drugs and choose what drug to prefer based on the cost of the drug alone.”

Specialty pharmacy and therapeutic areas for which multiple treatment options exist are a priority of pharmacy directors from both the state Medicaid programs and managed care organizations (MCOs). These individuals also expressed a need for research in areas of high cost or prevalence. Specifically, they mentioned a need for CER on hepatitis C, multiple sclerosis, biologics, and atypical antipsychotics. “What is interesting,” Gaffney pointed out, “is that there are public CER reports in these areas, raising the question: Why aren’t they using this research?”

The Avalere/KFF report identifi ed challenges faced by states seeking to use public CER in their decision making: lag time between updates to public reports, lengthy format, and lack of cost information. Furthermore, the reviews are not customized to individual states’ needs and research priorities.

Pearson described some of the specific barriers faced by the New England Comparative Effectiveness Public Advisory Council (CEPAC), an independent evidence review body that serves to enhance the applicability of CER for its 6 constituent New England states and other stakeholders. States want to know more than just acquisition cost, he explained. They are looking to answer questions about budget impact, downstream effects, and cost offset (eg, reduced hospitalizations or doctor visits). The focus needs to go beyond next year’s budget to a demonstration of value. In other words, whether inclusion of something that is more costly in the short term will save money and provide clinical benefits in the long term.

Timing of report release in relation to the state drug review process is also a major issue. Medicare regional directors have to issue formal public notifications prior to drug coverage reviews and offer a public comment period as well. It is diffi cult to integrate public CER into this process because the report release dates are unpredictable.

Study investigators found 4 different models being employed by Medicaid programs (

In the first model, the Medicaid program partners with a local academic institution. For example, Massachusetts hires its local medical school to conduct clinical and cost reviews.

In the second model, the state relies on the clinical reviews of 1 or more public health technology assessment organizations. Only the state of Washington relies completely on public health technology assessment.

To make spending more predictable, other state Medicaid programs are turning over their pharmacy benefits to MCOs. Thirty-fi ve states currently contract with MCOs to provide services for almost half of their Medicaid beneficiaries. According to Gaffney, this is an increasing trend. Compared with the state Medicaid programs included in the study, the coverage decision— making processes of Amerigroup, Molina, and United Healthcare are less transparent—MCOs do not use public CER or stakeholder input during formulary development.

Finally, in the most popular model of the study, states contract with a pharmacy benefi ts manager (PBM). PBMs are able to cater specifically to the scope of each state’s preferred drug list (PDL) review. The drug review reports of one popular PBM address all of the aforementioned barriers imposed by public CER, which is one of the reasons the interviewed states have come to rely so heavily on them.

States see PBMs as a 1-stop shop for claims processing, supplemental rebate negotiations, and clinical and cost review preparation. The pharm acy directors from each of these states explained that their PBM provides a consolidated, up-to-date drug review that includes both clinical and cost information for the specifi c drug or class they are considering. And, best of all, the reviews conclude with specific PDL recommendations. Florida, Louisiana, Nevada, Maryland, and Minnesota rely on the same PBM for drug reviews.

The study identified 5 commonalities in the type of information states are seeking to inform their coverage decisions:

2. Relevant to local needs and local contacts.

Ultimately, Gaffney said, “It is going to be important to have a dialogue about how to reconcile the economic pressures states are facing with the sensitivity to cost and have a discussion around how we can explicitly include cost in these reports.”

Enhancing the Applicability of Federal CER

Steven Pearson talked about the process being employed by CEPAC to incorporate federal CER into state coverage and reimbursement decisions in New England. In describing CEPAC’s model, he broke it down into 3 components: structure, content, and process.

CEPAC comprises 19 members (minimum of 2 per state) from all 6 New England states. The council consists of clinicians and methodologists, as well as health policy experts. Two ex officio nonvoting members represent the payer perspective: 1 public and 1 private. Members of CEPAC are expected to be experienced with evidencebased medicine.

The day-to-day operations of CEPAC are overseen by ICER as part of an AHRQ-funded initiative to develop and test new ways to improve the usefulness of federal evidence reviews for patients, clinicians, and payers.

CEPAC seeks guidance from an advisory board comprising state Medicaid directors, medical society representatives, patient advocates, and regional private insurers. The board helps with topic selection and offers advisement on CEPAC products and processes. But more importantly, said Pearson, board members engage in rich dialogue about the use of evidence reviews in policy. They have learned from one another, not just about the evidence, but about what they are going to do with it and some of the barriers to actually using it.

At its core, the process relies on the executive summary of the AHRQ review itself. To make it more applicable to CEPAC’s needs, ICER provides supplementary content.

Prior to topic review, ICER improves the timeliness and relevance of the AHRQ review by conducting a follow-up search—using the search terms of the original report&mdash;to identify late-breaking studies published after the report’s fi nal search date. ICER also provides CEPAC with state-specifi c data; for example, prevalence, utilization, and practice patterns.

To address the need for cost information, ICER prepares a comparative value report. Different aspects of value are included: cost, budget impact analyses (shortand long-term), and cost-effectiveness analyses. 

After CEPAC receives the AHRQ review and supplementary content, a preliminary teleconference is held with regional clinical experts to discuss any questions they might have about the evidence.

CEPAC holds an average of 2 day-long public meetings each year during which select comparative effectiveness topics are reviewed. Toward the end of each public meeting, CEPAC interfaces with payers and clinical experts during a policy round table. “This has been very, very valuable for driving an alignment between health technology assessment and what is going on in the clinical community,” said Pearson.

CEPAC does not make formal coverage decisions. Rather, it mimics the model used by the FDA advisory committees in which recommendations are based on committee members’ up or down vote regarding the adequacy of the evidence.

The votes taken by the CEPAC advisory board include: “Is the evidence adequate to demonstrate that drug A is equivalent or superior to usual care or drug B for patients with condition x?” It is to some extent, admits Pearson, a judgment: “There is no perfect algorithm.” CEPAC members are also asked to vote on whether each care option is “high,” “reasonable,” or “low” value and to describe, for the public record, their reasoning for their vote (

CEPAC sees this process as a way to engage not just payers but also the clinical community and the public in wrestling with these issues and in making votes about value.

A final report is prepared that includes policy recommendations regarding actions desired by societies, hospitals, or other stakeholders; comments on coverage options; and future research recommendations. The report is posted to the CEPAC website.

Among the highlights: presentations on the challenges faced by states in utilizing comparative effectiveness research, and on the significance of hypoglycemia in diabetes therapy.