Welcome to Value-Based Care, Outcomes and Reimbursements. This is lecture a, Foundations of Outcomes and Reimbursements. This lecture provides background information about the old HMO model. How HMOs compare with ACOs, and how ACOs are designed to ensure quality and reduce costs. This lecture also takes a close look at quality measurement, attribution, and risk stratification, and how they work together. The lecture concludes by looking at challenges in measurement and reporting. The learning objectives for this lecture are to compare the strategies of health maintenance organizations or HMOs and current accountable care organizations or ACOs. So we understand how the lessons learned from the experience of HMOs in the 1990s inform today's shared risk contracts and quality measurements. Describe how quality measures, attribution, and risk stratification work together to drive quality while reducing costs. Outline common issues related to collecting quality and cost measures. And discuss the administrative burden of reporting on various CMS payment models, and insight on how practices may reduce this burden. The models described in this lecture include a broad set of performance based payment strategies that link financial incentives to providers' performance using defined measures. To achieve better value through improvements in quality and slowing the growth, decreasing healthcare costs in healthcare spending. As a reminder from earlier lectures, ACOs and other value-based care models are meant to reward value. And to achieve the triple aim of better population health, better patient experience, and lower costs. To achieve that aim, we need to measure and reward high quality care. As the IOM and ARHQ definitions reflect, quality is about getting patients the care they need. The problems in quality often are broken down into the categories of overuse, underuse, and misuse. Overuse means that a patient gets services that don't provide a benefit or where the harm is greater than the benefit. An example of overuse would be a patient getting a prescription for an antibiotic when the patient has a cold. Or a patient getting a test repeated because the physician didn't have the results from the earlier test. Underuse means a patient doesn't get care that is needed. For example, if a patient didn't get needed physical therapy, that would be underuse. And misuse means that a patient gets the wrong care. For example, there's an error and a patient is prescribed the wrong medication. With that understanding of healthcare quality issues, how should our approaches to outcomes and reimbursement address quality? How can we measure quality and structure payments to encourage the right level of evidence-based care? Let's begin by looking at some history and lessons learned from earlier payment models. To provide context about outcomes and reimbursement in today's value-based payment models, let's look back a couple of decades at the experience of HMOs. In the 1990s, HMOs became a popular approach to rein in healthcare costs, which had been growing quickly. In this form of managed care, HMOs received capitated payments. In other words, the HMO got paid a fixed amount per member. And in return, the HMO provided or arranged for all covered services for its members. An HMO's profit or loss depended on whether it could deliver those services at a cost that was less than the capitated payments it received. This meant the HMO had a strong incentive to squeeze costs. The HMOs tried to control costs by negotiating low rates from providers, and by restricting their members' use of services. For many providers, the negotiated rate was lower than the cost of providing care. So those providers lost money. For members, costs generally were good, but services suffered. Members usually had low co-pays and other costs. But they had to get a referral from their primary care provider in order to see a specialist. If members got care outside the HMO, they weren't covered for that service. Over time, the HMO model faced a backlash from providers and from the general public. As Atul Gawande explains, the HMO approach to limiting unnecessary treatment was effective in reducing health care costs. But also had negative consequences that led to a backlash. Faceless corporate bureaucrats second guessing medical decisions from afar created an infuriating amount of hassle for physicians and patients trying to orchestrate necessary care. And sometimes led to outrageous mistakes. Insurance executives were accused of killing people. Facing a public outcry, they backed off and healthcare costs resumed their climb. Given the history of HMOs, many people worry about ACOs repeating the past. Like HMOs, ACOs involve providers taking on financial risk, rather than simply being paid for each service provided to patients. How do we use lessons learned from the HMO models to positively inform the design of ACOs? ACOs include some design elements that are meant to prevent the problems and encourage high quality care. Unlike HMO members, Medicare ACO beneficiaries may choose to obtain care outside the ACO and still have coverage for those services. The ACO is responsible for the cost of care for its assigned beneficiaries regardless of who provides the particular service. This provides greater coverage for beneficiaries and it reduces the incentive for an ACO to refer patients away for high cost services. The approach to payment also is different. Instead of full capitation, the medicare shared savings program ACO models use shared risk. Because providers are at risk for only a portion of the total cost of care. They do not face the same level of potential loss as with the HMO model. In addition, ACOs have to meet quality measures. These quality measures are intended to reward an ACO if it provides good care. And penalize an ACO if it fails to provide the right care. Quality measures were not part of the HMO model. As a side note, part of the reason that quality measures can be used in the ACO model has to do with technology changes between the 1990s and today. Providers and hospitals are much more likely to have electronic health records or EHRs. Providing quality measures in an EHR is much less time consuming and labor intensive than having people work through paper records to put together quality data. At the beginning of this lecture, we noted that many quality problems relate to overuse or underuse. Let's consider how the ACO model addresses those issues. It's important to note that in addition to driving up healthcare spending, overuse can cause serious harm to patients. Overuse of antibiotics leads to the development of drug resistant bacteria. Overtesting can lead to unnecessary exposure to radiation, and thus to increased rates of cancer. It also can cause needless stress and expense to patients. As well as overdiagnosis and treatment of diseases when the disease poses no real risk to the patient and treatment increases the risk of other harms. For example, in the United States Rates of thyroid cancer detection and removal have increased by three times in the past 20 years. Despite that dramatic increase in detection and treatment, the thyroid cancer death rate has not decreased. The number of patients experiencing permanent complications from thyroid surgery however, has dramatically increased. By creating financial incentives to avoid overuse, the ACO model can help address this issue. Patients, as consumers of healthcare, also can be encouraged to avoid overuse and choose high-value services. We'll look at the role of consumers more in the next lecture. Underuse was one of the key problems with the HMO model in the 1990s. Financial incentives encouraged denial of necessary care, as well as unnecessary or low value care. By including quality measures as a condition, the ACO model creates a different set of incentives. Let's look at the checks and balances built into the ACO model. The figure shown here illustrates how the three competing elements of quality, cost, and healthcare experience provide these checks and balances. In the first check and balance image, image one, ACOs could exert downward pressure by attempting to meet cost goals by restricting care to individuals. For example, a provider could give less care and not charge as much in order to achieve a lower cost of care. However, there is upward pressure also from patients because if they are not happy with the amount or quality of care, they have the ability to go somewhere else. The ACO, however, is still assigned the accumulated cost for these patients. Therefore, there isn't much reason to restrict care. Similarly in image two, it is possible that in ACO could meet the quality standards by over treating and spending huge amounts of money. However, the payment model has cost measures in place that discourage providers from over treating or using expensive treatments that don't improve health. Finally, in image three it is possible that the ACO could try to meet cost goals by reducing care. There is another check and balance here because the quality goals of the model encourage comprehensive care. These goals for providing care exist to assure the providers deliver care that promotes quality outcomes. With that background, let's define some key concepts and take a closer look at how they work in value based models to align payments and outcomes, so that quality is maintained or improved while costs are reduced. Quality measures are a key part to value-based models. Before Medicare ACO can share in any savings, it has to show that it met required quality standards. We'll look at some of the measures later in this lecture. To apply quality measurement, it's important to know who was responsible for meeting the measure. Attribution refers to the method used to decide who is responsible for the cost or quality of a patient's care. For Medicare ACOs the Centers for Medicare and Medicaid Services, or CMS, refers to attribution as assignment. But regardless of which word is used, the idea is to identify which providers were involved in caring for a patient and should be held accountable for cost or quality measures. For ACOs, attribution factors into calculation of benchmarks, financial performance, and quality measures. CMS publishes very detailed specifications for its methods of assigning beneficiaries. And those specifications are established and updated through the shared savings program regulations. Generally, a beneficiary can be assigned to a Medicare shared savings program ACO if the beneficiary's primary care physician is in the ACO. Or if the beneficiary goes to a specialist in the ACO for most of the beneficiary's primary care. Since beneficiaries are assigned based on their use of services during the performance year, the ACO doesn't get the finalized list of beneficiaries until after the end of the performance year. The assignment rules are different in the next generation ACO model. Those ACOs are told in advance which beneficiaries are assigned to them. If we know what the ACO is being measured on and who is responsible for meeting the measures, what else do we need to know? A fair assessment of quality also takes into account how sick or healthy the patient was to start with. This brings us to the idea of risk stratification. Risk stratification involves identifying the likely levels of patient health and care needs based on characteristics of the patient population. For example, patients in the age range of 65 to 70 are likely to be healthier than patients in the age range of 85 to 90. Patients with diabetes are likely to have more care needs than patients with no chronic diseases. These different risks are important in many aspects of the ACO model. Let's take a closer look at quality measurement and then dig a little deeper in to risks stratification. We'll begin with a big-picture look at quality. The Affordable Care Act, or ACA, directed the US Department of Health and Human Services to establish a national strategy for improving health care. The National Quality Strategy, or NQS, sets six priority areas. One, making care safer by reducing harm caused in the delivery of care. Two, ensuring that each person and family is engaged as partners in their care. Three, promoting effective communication and coordination of care. Four, promoting the most effective prevention and treatment practices for the leading causes of mortality, starting with cardiovascular disease. Five, working with communities to promote wide use of best practices to enable healthy living. Six, making quality care more affordable for individuals, families, employers, and the governments by developing and spreading new healthcare delivery models. The priority areas are sometimes called NQS domains. CMS's value-based payment models explicitly align quality measurement to those domains. Here's an example of alignment with a national quality strategy. For 2016, the quality metrics for Medicare ACOs include 34 measures. These measures fall into four of the six NQS domains. One, patient and or caregiver experience. Two, care coordination and, or safety. Three, preventive health. And four, care for at risk populations. As this table reflects, CMS's expectations for quality measurement become more demanding over time. During an ACOs first year of participation the ACO meets the standards by providing an accurate and complete report of quality measures. This type of measurement is called pay for reporting. In following years, the ACO also will be accessed on how well it performs on certain measures. This type of measurement is called pay for performance. In unit five, we learned about different types of measures. As a reminder, process measures look at whether a particular action was performed. For example, a process measure might assess whether a provider tested patient's blood sugar levels as recommended. Historically, process measures have been more common, because they could be more easily produced from the available data. Outcomes measures, on the other hand, look at how the patient fared. For example, an outcome measure might access whether a patients blood sugar level was under good control during the measurement period. Outcomes measures require more clinical data and are difficult to produce in a reasonable time and without a labor intensive chart review. Unless data is being captured in a EHR. Currently, CMS is working to increase it's use of outcomes measures. We should note that private payers who operate ACO models may not use the same set of measures required for the CMS models. However, in general, these private payers maintain a set of quality measures that work in conjunction with the attributed patient population and risk stratification factors to set pay for performance targets. Let's take a closer look at risk stratification. To understand the importance of risk stratification consider that a severely ill patient is likely to require more services and that even with excellent care. The patient may still have worst outcomes than a patient who is relatively healthy. If the value based arrangement fails to account for those differences it can create an unintended consequence, where providers have an incentive to avoid caring for higher needs patients. Taking risk factors into account makes measurement and payment more accurate and fair. In quality measurement, risk stratification effects scores. So, providers who care for sicker patients, aren't unfairly compared with providers who care for healthier patients. This is especially important when providers payments are linked to quality performance or when providers performance is going to be publicly reported. In a risk stratification approach, patients can be divided into different Different quality measurement groups. High-risk patients could be counted in one measure and low-risk patient in another. Risk stratification can help identify places where there are disparities in the care that different groups received. A related approach is to define the quality measures so some patients are excluded. For example a surgical care measure may be defined so it does not count patients who would not benefit from surgery. In payments, risk stratification reflects the additional cost of caring for higher risk patients. Imagine if an ACO got paid the same amount to care for a very sick patient who was in the physician's office every other week, as it got paid to care for a healthy patient who just needed an annual wellness exam. That would set up a pretty strong incentive for the ACO to try to serve only the healthy patients. To prevent that problem, CMS includes risk adjustments in its payment model and it monitors the medical ACOs to make sure they aren't avoiding at-risk patients. For reimbursement, CMS uses a model called the CMS Hierarchical Condition Category or HCC to calculate ACO beneficiary risk scores. The system uses diagnosis Information from claims data for individual patients to create an individual risk score. Which is then averaged across the assigned beneficiaries to create a risk adjustment factor for the beneficiary group. In addition, providers can use risk stratification data to target carers who meet patients' needs, especially for patients who are at high risk of complications if their care isn't carefully coordinated. For example, a safety net provider organization in Colorado analyzed its data to identify different levels of patient needs and then matched services to the needs. All patients could choose to receive text messages, reminding them of appointments and recommendations for preventive care. While higher risk patients also would receive complex care coordination support with a care team that included patient navigators. In defining different risk levels, the organization thought about the patients currently being seen in primary care clinics and the patients who needed services, but had not been presented at the clinics and had lacked a medical home. To help ACOs with their data needs, CMS pulls together information from Medicare enrollment and claims data, and provides each ACO with aggregate information about its assigned population, and financial performance, on a quarterly basis during the performance year. For CMS, access to past history can be easier as CMS is in essence a single payer of healthcare for patient populations that are 65 and older. Private insurers maybe challenged in gaining access to equal levels of information for their patient populations. Now that we understand why risk stratification is important let's look at an area where it gets difficult. One of the challenges in risk stratification involves social determinants health. These are the many environmental factors that affect a person's health. These factors have a big impact on health but they are difficult to address in current quality measurement and reporting. Health care systems and quality measures generally track information about medical care. As this diagram shows, medical care is one factor in health outcomes, but it is not the largest factor. Providers can control the medical care they provide to patients. They can ask about and try to affect patient's health behaviors, for example by counseling patients on a balanced diet or offering smoking cessation programs. Factors such as social and societal characteristics and total ecology however, have a greater overall effect on health outcomes and most healthcare organizations aren't well equipped to address those factors. A healthcare provider has very little ability to influence whether a patient lives in a safe home, for example. The providers EHR probably has data feeds for the patients current address, but not for the details of the patients housing situation. Quality measures generally don't take housing status into account. But housing has an enormous impact on the patient's health. In designing value based outcomes and reimbursement models, accounting for social determinants of health is difficult. CMS has been criticized for using quality metrics that may unfairly penalize hospitals and providers in low income areas. Recently, CMS announced an initiative. The accountable health community model meant to improve health outcomes by promoting collaboration between clinical healthcare and community social support organizations. In addition, the National Quality Forum, or NQF, is examining the role of socioeconomic factors in quality measures. For those who are interested in seeing social determinants data at a state and county level. The county health rankings and road maps program, which is a collaboration between the Robert Wood Johnson Foundation and the University of Wisconsin population health institute provides a wealth of information. Let's look at some of the other challenges in collecting measures of cost and quality. You may recall from unit five that there may be challenges in data collection for measurement. For example, do providers consistently record information? Is the data in the same place and consistently defined? Some of these challenges relates to the use of EHRs. Health information technology systems, must be designed and implemented correctly. Although certified EHR technology requirements include the ability to calculate some quality measures, the certification testing doesn't always find all the bugs in a system. These bugs can cause issues in the logic of quality reports, that must be corrected to meet the requirements. And to be fair, sometimes the reporting logic of the measure specification has errors that needs to be resolved. In addition, if the EHR isn't implemented and configured correctly, there could be inaccurate reports. For example, an incorrectly configured EHR might pull data from the wrong fields, resulting in inaccurate quality measures. Workflow is the most common element overlooked in preparing to report on quality outcomes. Their provider, or other person using the EHR, might not understand the workflow steps, and may consequently miss important clicks or documentation needed in the electronic record, in order for the quality reporting measures to be accurate. Sometimes a provider who has had the appropriate training may be unwilling or unable to follow the workflow. This may result in them not completing the required steps in the EHR, and thus producing inaccurate quality measures. As we've discussed, quality measurement is important. Currently, however, it also places huge administrative burdens on providers and their staff. According to one survey of physician practices, practices reported that their physicians and staff spend 15.1 hours per physician per week dealing with external quality measures, including the following. Tracking quality measure specifications, developing and implementing data collection processes, entering information into the medical record, and collecting and transmitting data. This is equivalent to 785.2 staff and physician hours per physician per year. The average physician spent 2.6 hours per week, enough time to care for approximately nine additional patients dealing with quality measures. Staff other than physicians spent 12.5 hours per physician per week dealing with quality measures, with the largest proportion, 6.6 hours, by licensed practical nurses and medical assistants. This happens for a number of reasons. Different payers and oversight bodies often require different measures. Measures may be required for different purposes, such as quality improvement, public health goals, transparency on costs and outcomes, regulatory requirements related to health and safety or accreditation, payment and purchasing decisions, or other needs. Even if each organization requires a reasonable number of measures, the total number of measures quickly adds up. Making the problem worse, payers may require similar measures with slightly different specifications. So the same quality reporting can't be reused across the payers. Ideally, EHRs would simplify reporting by using the data that providers are capturing as part of providing care, and calculating quality measures without additional effort. That is not the current state, however. Instead, the average physician in the survey was reported to spend 2.3 hours per week entering information into the medical record only for the purpose of reporting for quality measures from external entities. Licensed practical nurses and medical assistants spent even more time on this information entry, 6.1 hours per practitioner per week. More work is needed, so that measures can be produced from EHRs without this additional burden. Finally, there is also a lot of dissatisfaction with many of the measures currently in use. Providers often are frustrated with measures that don't reflect their particular specialty, or that aren't perceived as accurately representing quality. Other critics point to the lack of good measures of value, addressing both outcomes and cost. So while there is widespread agreement that quality measurement is important, there is still plenty of work to be done to improve measurement itself. In the movement toward value-based care and payments, there is a recognition that the current challenges around reporting need to be addressed. There is a lot of discussion about the need for metric alignment. And in early 2016, the Core Quality Measures Collaborative, which includes CMS, commercial health plans, physician groups, and other stakeholders, announced a set of seven clinical measure sets that would support multi-payer alignment of quality measures for physician quality programs. Large healthcare organizations often work closely with their IT vendors or internal IT staff to optimize their EHR, to address issues around quality reporting. Smaller organizations and provider practices often don't have the staff or financial resources to focus on that work. However, in the Comprehensive Primary Care plus, or CPC+ program, announced by CMS in 2016, there is a requirement for Track 2 practices participating in alternative payment models to have a letter of commitment from their health IT vendor, that they will be partners in achieving the requirements of the program, including risk stratification and reporting. We have talked about resistance or lack of training to complete designed workflows that capture the information in the required format to generate accurate reports. There may be different solutions for different settings. Sometimes additional provider training is necessary. Sometimes care team roles can be defined to maximize the efficiency of capturing the necessary data from the patient visit. Sometimes workflows need to be redesigned. Many healthcare delivery organizations are also investing in health IT solutions beyond their EHRs, including data aggregation tools. These can help not only with the measurement and reporting required through value-based contracts, but also with the benchmarking and quality improvement projects that are an important element of moving the needle. And making sure that individual providers and clinic teams are achieving quality outcomes for both individuals patients, and patient panels attributed or assigned to them. This concludes Lecture a of Outcomes and Reimbursements. In summary, this lecture provided background information about the old HMO model, how HMOs compare with ACOs, and how ACOs are designed to ensure quality and reduce costs. Then we took a closer look at how quality measurement, attribution, and risk stratification work together. Finally, we looked at challenges in measurement and reporting.
