Decoding Health Care
Part 11: Value Based Care
When I was in practice, several staff members in my office felt I was too demanding. They didn’t like working with me. But I recruited and retained a very talented team of administrative staff, nurses, and medical technicians who shared my vision and we worked together for several years. We felt that our goal should be to deliver care the way we would want to receive it, even if it meant extra work. We felt that was important for cancer patients. After I left practice, what I missed the most was my team.
My vision was certainly not unique to me and my team. I was a good doctor but I wasn’t that special. In fact, some of the best care in medicine is provided by physicians with the same type of care teams. This care model is absolutely patient centric. And it is at the heart of value-based care. Interestingly, it is also the basis for most alternative payment models (in which we move away from traditional fee for service). And it all started with pediatrics.
Children with special health care needs are a complicated population. Among other things, they consume a lot of time and cost a lot of money. Their care often involves a number of physicians whose efforts need to be coordinated. And no ordinary pediatric practice is really equipped to do what needs to be done. So a care delivery model was developed to meet the needs of these patients, and that model was named the patient centered medical home (PCMH) (https://www.ahrq.gov/ncepcr/research/care-coordination/pcmh/define.html).
The core principles of the patient centered medical home are:
1. Team based care, led by a physician, but multi-disciplinary with each team member practicing to their scope of licensure.
2. Evidence based clinical care.
3. Enhanced patient access to medical services and the care team.
4. Patient centered comprehensive care management.
5. Care coordination especially during transitions in care.
6. Shared decision making.
7. Performance measurement and quality improvement.
All of this needs to be supported by an IT infrastructure that facilitates documentation, communication, and data analysis. And it needs to be supported by appropriate payment to the practice to allow transformation, implementation, and execution. This is an ambitious endeavor. It clearly means more time spent on each patient with the goal of better patient outcomes as well as happier patients and more fulfilled providers.
The amazing thing is that the pediatric community embraced this approach. It went far beyond the special needs patients and was soon adopted by over 20% of all pediatricians. It did not stop there. Other primary care practices, including adult internal medicine, decided it was a good model. And the commercial health plans as well as Medicaid also bought in. Practices that were accredited as a medical home(usually by an organization like the NCQA, the National Committee for Quality Assurance) received an extra management fee per patient per month from the health plan. It was understood that these extra services cost money and took time, and that the number of patients that a practice could accommodate would invariably be fewer.
The enthusiasm for the medical home approach was contagious. In 1985, the federal government, through the Social HMO programs, decided to experiment with PCMH principles in Medicare for the needs of the frail elderly, another challenging patient population. The pilot was a response to the inherent limitations of traditional healthcare models to adequately address the complex needs of these patients. By coordinating medical care with essential social services like meal delivery, transportation assistance, and caregiver support, the program aimed to enhance the overall well-being and quality of life for its participants. This focus on paying for social determinants that influence health outcomes with a special benefits design was novel at the time. Awesome.
This should be the beginning of a beautiful story. But alas it is not. There are many reasons why the initial enthusiasm for the PCMH lagged. The first is that although patients liked this arrangement, objective evidence of improved “quality” of care is hard to come by. Part of this may be because it is so difficult to define what we mean when we say “quality”. If you define quality as more patients getting a mammogram, mission accomplished. But real improved health outcomes is another story.
A second problem is that the amount of work required of the practices was substantial. After considering the cost of model implementation it was tough to even break even. Remember that in general the management fees were layered on top of traditional fee for service. Those fees for service were still the bread and butter of the practice. So after a period of initial enthusiasm most practices just decided it wasn’t worth it. Small practices were particularly frustrated.
A third problem involved what the health plans got out of the model. There was a fair amount of extra work for them as well. For example, “attribution”, the process by which a patient is “assigned” to a specific medical group is a constant headache. Should that patient “count” if they get a significant amount of care outside the medical home? When it comes to dollars and cents, savings could only be generated by reducing ER visits and hospitalizations. These savings were fairly modest and not uniform across practices or patients (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5578895/).
But as we have seen in health care over and over again, never let facts get in the way of a good hypothesis. The Affordable Care Act breathed new life into the PCMH model in the form of ACOs. In an effort to offer alternatives to traditional fee for service and tackle rising health care costs, CMMI, the Center for Medicare and Medicaid Innovation (which had been created by the Affordable Care Act to experiment in alternative payment models) enthusiastically adopted ACO’s (Accountable Care Organizations). And we have seen the application of the medical home model in subspecialty care, specifically oncology.
Now some might object to my lumping ACO’s in with the medical home, but I think it is justified. ACO’s are groups of health care entities including physicians and hospitals that take responsibility for the cost and quality of care delivered to a population. Although ACO’s are not as prescriptive as medical homes about how to get there and appear to be more focused on cost than care delivery, a top performing medical home should be the engine of a successful ACO. And many of the quality measures and health outcomes we are interested in apply to both. And perhaps most importantly, they both aim to achieve the same goals, better quality at lower cost.
ACOs and their derivatives have been around for more than 10 years. They make up the bulk of CMMI’s portfolio (https://www.nejm.org/doi/full/10.1056/NEJMsb2031138). When CMS says that the majority of Medicare beneficiaries will be in an “alternative payment model” in the next couple of years it means they expect everyone to be in ACOs (or closely related programs). CMMI thinks ACOs really work. I am not so certain that I agree and I am not alone (https://www.statnews.com/2022/07/26/value-based-payment-produced-little-value/). My skepticism has to do with how CMS is doing their accounting, and that accounting is exclusively around cost. There is very little data that ACOs improve the quality of care. Let me explain.
First, the programs are voluntary. There is a profound selection bias because the health care provider needs to believe they can succeed. And there is good evidence that entities that do not succeed simply drop out. Attrition in these programs has been substantial. Second, the vast majority of health care entities that participate are hospitals. As I previously discussed hospitals are not in the business of losing money. Hospitals have figured out that certain types of patients are “winners’ in ACOs and they cherry pick these patients for inclusion. Again, data shows you are more likely to succeed if you take care of lower risk patients. Third, the math which determines whether an ACO is successful is based on performance compared to a target price. CMMI calculates what they think the patients would otherwise cost. If the ACO spends less they call that savings. It doesn’t take a genius to realize that whether savings are actually generated is only as good as the model. Yet CMS has trumpeted the success of the ACO approach. Not everyone agrees (Understanding The Latest ACO “Savings”: Curb Your Enthusiasm And Sharpen Your Pencils—Part 1 | Health Affairs). But CMS pushes on.
Cancer patients seem perfect for a medical home approach. And for the last decade, the oncology patient centered medical home has gained traction. This initially started with a single community oncology practice in suburban Philadelphia led by Dr. John Sprandio. John adopted primary care medical home principles, transformed his practice care delivery model, built a tracking system that sat on top of his EMR and showed a substantial reduction in ER and inpatient utilization. This generated a lot of excitement. Building on this, Dr. Barbara McAneny built a consortium of seven geographically dispersed community oncology practices that agreed to adopt this same care delivery model. She obtained a grant from CMMI (at the time CMMI was funding specific pilots) and Come Home was born.
Come Home standardized a number of care processes within practices. And like the original primary care medical home, it was focused on team-based care, enhanced access, comprehensive care delivery, and coordination of care. There have been numerous presentations and publications that discuss Come Home. Let me summarize. Come Home did save money, and the savings were real: cost of care in patients treated in Come Home practices was compared with concurrent propensity matched controls. But the amount of money saved per patient was fairly modest, about $600 every three months, amounting to less than 5% of the total cost of an episode of care (net after considering the amount of the grant award). There were small differences in ER and inpatient utilization (Health Care Innovation Awards Disease-Specific Evaluation (cms.gov)). At least one publication has suggested there were reduced costs in patients at the end of life. A solid single but not a home run.
Several other attempts at implementing the medical home in oncology soon followed. While I was serving as Oncology Medical Director at Aetna, Ira Klein and I piloted the first payer driven oncology medical home program. We recruited about a dozen practices, provided them with a lot of data, and offered them a shared savings model based on drug, ER, and hospital spend compared to concurrent spend in matched practices. We also found encouraging savings (https://ascopubs.org/doi/full/10.1200/JOP.17.00091?role=tab).
All of this ground work was enough to prompt CMMI to launch a national oncology payment pilot based on medical home principles, the Oncology Care Model (OCM) (https://www.cms.gov/priorities/innovation/innovation-models/oncology-care). This model was voluntary; 196 practices signed up. It was built on top of a foundation of Medicare fee for service. Practices received a monthly management fee for patients undergoing systemic therapy for cancer. In addition, practices could earn a performance-based payment if their costs were less than modeled costs for patients with those specific cancers. The practices that participated really liked the model. CMMI hoped that practices would work hard at keeping patients out of the ER and hospital, and that they would make “value-based” decisions when making chemotherapy choices. What I mean is that the hope was that if two drugs were equivalent in terms of efficacy and toxicity for a given cancer that the physician would choose the less expensive one knowing they would be rewarded with a performance-based payment by generating savings. The program was supposed to last 5 years but was extended by another year because Covid provided some unique challenges. The program concluded about 2 years ago.
The results were disappointing, although the spin doctors are still trying to convince us otherwise (https://www.cms.gov/priorities/innovation/data-and-reports/2023/ocm-evaluation-pp1-9-exec-sum). When compared to patients treated concurrently in non-OCM practices there were no net savings. This was because although there may have been some savings in a handful of cancers in OCM practices, the monthly management fee paid to OCM practices exceeded the cost of care savings in participating practices. So the program wound up costing Medicare extra money. It was very troubling that there appeared to be no reduction in ER or hospital utilization and there were no differences in prescribing behavior.
Remaining true to form, CMMI has decided to move on to a modified OCM called the Enhancing Oncology Model (EOM) that includes fewer tumor types, reduces the monthly management fee, and changes the shared savings model by requiring a payback if the practices spends more than the model says they should, so-called downside risk (https://jamanetwork.com/journals/jama-health-forum/fullarticle/2800747). Only about 44 practices signed up for this, far fewer than in OCM. We are a little less than a year into the program so it’s too early to know how this will turn out. Even if savings are generated, and I am certainly not sure this will happen, a critical appraisal would question the problems of practice selection and limited patient eligibility. In other words, it’s not a solution for our cost challenges in cancer care.
Based on what I have just discussed, you might think I really do not like the medical home model. You would be totally wrong. It is an excellent model of care. It’s just a lousy payment model, at least as currently constructed. Because it is so hard to “prove” quality was improved it is tempting to assume that it was not. Wrong. Just ask practices that participated whether the quality of care improved and they will all say yes.
I’ll provide two examples. First, just a few years ago, if you called your oncologist with an acute medical problem they would send you to the ER and more often than not you would get admitted. Not anymore. Every OCM practice implemented urgent visit clinics, often staffed by a physician extender and offering same day care. Second, prior to the OCM only hospitals used patient navigators. Patient navigators were invented to help patients “navigate” the complex health care system. Hospitals had them because they could afford them. But OCM required them and helped pay for them. And everyone, including doctors, staff, and patients loved them.
But by including costs that are difficult or impossible to control it was very difficult to save money. In oncology, those costs are in the drugs. It would be unethical to treat a patient with a less costly but inferior therapy, and to their credit, the oncologists in the OCM did not. So how do we solve this? I propose we build a model that takes the drugs out. It’s not that we should ignore the drugs altogether; I have some ideas which I will write about in a subsequent post. And I have already mentioned one of those ideas, paying for drugs approved by an accelerated FDA process (using surrogate endpoints) at a “conditional” price until confirmatory data is available. These so-called outcomes-based contracts would put the brakes on expensive new-to-market therapies. If we reduce total costs to manageable costs this might just work. The medical home (minus the drugs) will NOT save a ton of money but it will provide a sustainable, patient centric care model.
The other thing we should consider is moving from shared savings models to more global payments like capitation or episode-based payments or bundles. We have learned a lot about what it costs to take care of various cancers. Let’s pay the practice a certain amount to care for an early-stage breast cancer, or a late-stage lung cancer. Again, we would need to carve the drugs out. This kind of approach is already being done in other payment areas. DRGs, the basis for Medicare payment to hospitals for inpatient care, are an episode-based payment. CMMI has also experimented with a bundle for major lower extremity joint replacement. This approach will have practices that are both winners and losers, ie some practices will make money and some will lose money (if for example the costs exceed the bundle or capitated payment), and that scares some people. But there is a school of thought that doctors won’t change their behavior unless they have financial skin in the game, and there is evidence that risk is a powerful motivator.
But is this wise? Can patients be harmed? That is an excellent question. By doing less you can get paid more. Capitation was all the rage in California in the 1990’s. But both patients and doctors felt the quality of care suffered (https://pubmed.ncbi.nlm.nih.gov/9228437/). People were so unhappy capitation was largely abandoned. But I would like to think we have learned our lesson. In order to protect patients, we need to make sure quality metrics are baked into whatever model we move forward with. And underperformance on these quality measures should be penalized aggressively. We just need to figure out what these quality measures ought to be; that will also be the topic of an upcoming post.
For this approach to work there needs to be the ability to obtain data almost real time to allow practices to modify behavior. And of course there needs to be transparency and trust between doctors and payers. But the beauty of the medical home approach is that it can allow us to deliver care in a fashion similar to but potentially better than the one we have become accustomed to. Plus it can allow practices to be creative in adapting the care model to populations of patients based on their needs (including social determinants). First we need to get the doctors to buy in. Then we need to get the payers to agree to a standard model; every payer cannot have their own model. And they need to provide a reasonable management fee to fund these extra services. You might be thinking good luck. But this is precisely what CMMI was created to do. The patient centered medical home should be the clinical engine that drives value-based care.
All doctors want to give their patients the best care possible. I think we may have a road map. Let’s separate the money issues from the quality of care issues. Not that we can ignore the money issues. Value must consider both the quality of care provided and the cost. In our next post we will address the use of evidence-based medicine to promote high value medical care, especially the use of chemotherapy.