Do you remember the discussions about the medical home? The patient-centered medical home was going to be the answer to the problems in primary care. I never bought it. I worked in a large primary care practice in East Tennessee that was a level three medical home. It seemed to me that it was a matter of checking boxes. It is an administrative solution for a clinical problem. I never saw that it changed what happened between the clinicians and patients in any way.
The multipayer medical home project started with 32 practices in Southeastern Pennsylvania in 2008. Practices were invited to participate by 6 health plans (3 commercial and 3 Medicaid managed care plans). There were 64,000 patients involved. The improvement intervention included technical assistance, disease registries, coaching to achieve practice transformation and medical home recognition. Performance improvement efforts targeted asthma for pediatric practices and diabetes for practices serving adults. There were just over $30,000 in incentive benefits for each participating physician.
This project failed. “There were no reductions in health care utilization of hospital, emergency department, or ambulatory care services or total costs, and there was improvement in only 1 of 11 quality measures of chronic disease management, nephropathy monitoring in diabetes.”
These are expensive projects. It is important that we learn the lessons that they teach. Implementing the medical home is expensive for a practice. It involves a lot of effort. The patient-centered medical home, like many other clinical interventions, has been promoted for widespread adoption of a number of structural changes in practices before understanding which features really matter. It is past time to evaluate what works and what does not.
They Fail Because They Don’t Follow the Evidence
An editorial accompanied the CPC+ failure article in the Journal of the American Medical Association on Dec 15, 2023. It pointed to factors that led to minor increases in success. “…spending growth was lower among CPC Plus practices that also participated in Medicare Shared Savings Program ACOs compared with those participating only in CPC Plus, suggesting potential synergies between primary care transformation and value-based payments incentives at a broader organizational level. Singh and colleagues noted that the least favorable spending and utilization outcomes occurred in hospital-owned or system-owned practices that were not participating in Medicare ACOs.” So, compensation that was less fee-for-service and more value based made a difference. Hospital or system-owned practices not participating in valued based payment models had the worst performance in generally poor performance. Independent primary care practices performed best and that was absolutely our experience in the Quality Blue Primary Care Program in Louisiana. If you depend on heads in beds, expensive procedures, and expensive tests improving care while reducing costs cannot happen. Large hospital systems lose revenue when we keep people healthy. This editorial did not really point to major opportunities to improve primary care in ways that really matter.
That brings us back to Dorothy. Dorothy is primary care on the yellow brick road to improvement. The Scarecrow, the Tin Man, and the Cowardly Lion are the health leaders working to help her get to her goal, but they all have something missing and that is the problem. Eighty-six percent of healthcare costs come from patients with chronic diseases. Successfully managing patients with chronic diseases requires a comprehensive solution. The Comprehensive Primary Care programs had pieces of a good program, but they did not fully address the new science, new systems, and new payment models that are required for the best results. That is why nearly all quality improvement efforts fail to improve quality and reduce costs. The people running these programs think they are doing as well as anyone could do, but they are wrong. They are doing things that make sense, but they are mostly ignoring the evidence.
This next point is extremely important. Interventions that make sense frequently make no difference. The target condition for improvement in the medical home adult patients was diabetes. It is easy to tell how well you are doing in diabetes by measuring how often you achieve optimal medical therapy which is defined by the concurrent achievement of five variables:
1. Blood pressure of 140/90 or less
2. Hemoglobin A1c of 8 or less
3. LDL cholesterol of 100 or less and on a statin
4. Aspirin in patients with arterial or chronic kidney disease
5. Not smoking cigarettes
You cannot know how well you are doing with concurrently achieving these 5 variables if you are not measuring them. When you do measure it, it is always shocking how bad the numbers are when you look for the first time. The practices that think they are doing as well as anyone could are wrong if they are not even measuring these targets and I have seen no evidence this was done in any of these programs.
Here is some critical evidence. Diabetes management represents a great opportunity to improve health and reduce costs, but performance in our country is horrible on the measures that improve outcomes. “Many approaches have been tried to improve diabetes care but, with one exception, have been mostly ineffective. These include simply reminding patients about appointments; providing laboratory information on the patient to the physician, even when specific treatment recommendations for the individual patient were included; case management when the case manager could not make independent treatment decisions; education of physicians; and multifaceted quality improvement interventions in the practice setting.” None of that works and many organizations continue to waste money and time doing it. “The one approach that has proven to be effective is using specially trained nurses or pharmacists, under appropriate supervision, with authority to make medication changes without consulting the physician as long as the changes fell within approved treatment algorithms.” That’s it. That’s what works, but very few practices do that. I can see no evidence that the multipayer medical home, CPC or CPC+ used that approach and that is one big reason they failed.
Optimal medical therapy achievement for diabetes is the critical quality measure for that condition. If you look at the data, only about 20% of people with diabetes achieve those goals concurrently which make all the difference. It is not just achieving the goals, it is how you achieve them. Lisinopril or losartan and spironolactone or eplerenone for high blood pressure, statins for cholesterol, and metformin for diabetes all have antioxidant and anti-inflammatory effects that protect cells and organs more than other methods of lowering the target risk factor.
I have broad experience using algorithms for diabetes and related conditions. I published accomplishments using a protocol and a systematic approach in work going back to the year 2000. I worked with a team of three pharmacists at Holston Medical Group in East Tennessee. We worked with high-risk patients with diabetes and hypertension for Eastman Chemical Company. The pharmacists were all University of Tennessee faculty members and we trained a pharmacist annually in an ambulatory care fellowship. They saw patients just as I did under a collaborative practice agreement. We really moved the numbers as you can see in the slide below.
When I started working for the Quality Blue Primary Care program at BCBS of Louisiana, average concurrent achievement of the five goals for optimal medical therapy was about 14%. Within a couple of years we had improved that number to the 30-40% range and some individuals were achieving those goals in 60% if their patients. This program is listed in the references of the CPC+ failure article in JAMA. We reduced net costs by $27 per member per month within a year by reducing hospital admissions and length of stay. I have worked with Vestra Health for 5 years. Patients who are seen in their worksite clinic cost half as much and are in the hospital one fifth as often and the ER one third as often compared with patients seen in the community. That is a massive difference in cost driven by a robust change in clinical outcomes. The Vestra clinic has one family practice physician working with 4 nurse practitioners and they all use optimal medical therapy protocols. They use other levers to reduce cost like direct contracting. Over more than two decades, the initiatives that I worked with that used protocols and a systematic approach moved the cost and quality numbers. I have never seen anyone move these numbers significantly or persistently unless they were using algorithms or protocols. That was a major defect in the CPC and multipayer medical home efforts. They did not use protocols. It could not work!
The Institute of Medicine (now the National Academy of Medicine) provided a manual to guide us in improving health and saving money. They did not just point to a horrible problem; they provided a roadmap to fix it. Their analysis? The US health system performs poorly for a very simple reason. Chronic diseases cause most premature disability, death, and cost and we have not done the extensive system re-engineering required to address that reality. Tweaking around the edges will not work. They also gave us a simple step-by-step guide to use and they were very specific in their recommendations:
“To initiate the process of change, the committee believes the health care system must focus greater attention on the development of care processes for the common conditions that afflict many people. A limited number of such conditions, about 15 to 25, account for the majority of health care services… Nearly all of these conditions are chronic. By focusing attention on a limited number of common conditions, the committee believes it will be possible to make sizable improvements in the quality of care received by many individuals within the coming decade.”
“Health care for chronic conditions is very different from care for acute episodic illnesses. Care for the chronically ill needs to be a collaborative, multidisciplinary process. Effective methods of communication, both among caregivers and between caregivers and patients, are critical to providing high-quality care.” They are describing the outpatient advanced primary care team focused on cardiometabolic disease.
Now they get really specific. “To facilitate this process, the Agency for Healthcare Research and Quality should identify a limited number of priority conditions that affect many people and account for a sizable portion of the national health burden and associated expenditures. In identifying these priority conditions, the agency should consider using the list of conditions identified through the Medical Expenditure Panel Survey (2000). According to the most recent survey data, the top 15 priority conditions are cancer, diabetes, emphysema, high cholesterol, HIV/AIDS, hypertension, ischemic heart disease, stroke, arthritis, asthma, gall bladder disease, stomach ulcers, back problems, Alzheimer's disease and other dementias, and depression and anxiety disorders.
Health care organizations, clinicians, purchasers, and other stakeholders should then work together to:
(1) organize evidence-based care processes consistent with best practices, This is a specific recommendation to use evidence-based protocols and systems specific to the priority chronic condition. Most of this work can be done remotely.
(2) organize major prevention programs to target key health risk behaviors associated with the onset or progression of these conditions, Patients need coaching to understand their disease, what can be done about it, and why it is worth their time.
(3) develop the information infrastructure needed to support the provision of care and the ongoing measurement of care processes and patient outcomes, and an EMR does not meet this requirement. The population health tool must identify the high-risk, high-cost conditions, which patients have achieved their goals and had their tests, who is on the protocol medications and taking them, and who has not been seen, to feed gaps to nurse coaches for closure.
(4) align the incentives inherent in payment and accountability processes with the goal of quality improvement.” The best payment model is full risk. Pay the team to do the work. Fee for service is a massive barrier.
Diabetes Care called for teams of nurses and pharmacists teamed with a physician and using protocols that are authorized to make independent treatment decisions in real time. The National Academy of Medicine provided a guide for implementation. Both have been ignored here in the US but adopted in other countries. Until we adopt these recommendations and follow this evidence we will never have better health at lower cost.
There is another side to this failure. Five percent of patients generate half of healthcare costs. Adding the next 15% of high-risk, high-cost patients with chronic conditions brings that number up to 86%. Focusing your efforts on 20% of the patients with the fifteen priority conditions provides the biggest gains in clinical and financial outcomes most quickly. Half of the patients cost almost nothing because they are healthy. Applying extra resources on them accomplishes nothing. It is really as simple as one and one are two.
You have seen my work as an individual, at Holston Medical group, with BCBS of Louisiana, and with Vestra Health. The initiatives were all highly successful in improving clinical results and/or reducing costs in multiple settings. But these are more modest accomplishments compared to the nationwide model in Singapore, a nation of about 5 million people. They employ evidence-based processes consistent with best practices in many of their national policies including education and healthcare. Their system is rooted in primary care and they provide it for everyone. They have 20 one stop advanced primary care practices scattered across the country and within the clinics they have pharmacist teams treating the priority chronic conditions hypertension, diabetes, and cholesterol which reduce the toll of heart failure, chronic kidney disease, heart attack, and stroke. The result? We spend 18% of gross domestic product (GDP) on healthcare. They spend less than 5% and they live longer. They are where we need to go. England provides primary care for everyone, primary care manages heart artery disease using optimal medical therapy and they live longer for half the impact on GDP.
Despite the failures of the federal primary care improvement initiatives and most quality improvement efforts to improve quality and reduce costs in this country, there have been massive improvements at scale in some practices. The Kaiser Permanente system has done it best. In Northern California, nurse teams using protocols lowered blood pressure to less than 140/90 in 90% of 750,000 people. In Colorado, their Collaborative Coronary Care service is even more impressive. Nurse and pharmacists do most of the work. Patients who have heart artery disease in the program cost $21,900 less per member per year than those who are not. They are ten times more likely to be alive in 5 years compared those who are not in the program There is a ten-fold reduction in cardiovascular deaths. Even more impressively, there is a ten-fold reduction in ALL-CAUSE MORTALITY! At latest report there were 12,000 people in the program. 12,000 X $21,900 = $262,800,000. That is a quarter of a billion dollars saved for one condition in one state and even that could be better. The Kaiser heart program is rooted in their cardiac rehab program. Move it to primary care, include the priority conditions heart artery disease, other artery disease, chronic kidney disease, hypertension, diabetes, high cholesterol, and stroke to be cared for by the same team. Use telemedicine and home testing to make the care convenient and reduce costs even more. Use information technology to automate the whole process as much as possible. Make it an urgent priority to expand the model to the whole system.
Steno Diabetes Clinic in Denmark used a similar program to treat patients with high risk type 2 diabetes, compared with usual care patients managed by nurses and protocols had one fourth as many heart attacks, one fifth as many strokes, and one sixth as many people went on dialysis. All other diabetes outcomes were similarly improved. The intensively managed patients live eight years longer without heart attacks and strokes. Hospitalizations for heart failure were 70% reduced. These are documented, proven results. The path forward is clear.
The reason CPC+ failed is perfectly clear. They failed because they did not follow the massive body of evidence that proves a comprehensive solution can succeed. They failed because key components were missing. More on that in the next post.
https://ourworldindata.org/grapher/united-states-rates-of-covid-19-deaths-by-vaccination-status
This data comes from a not for profit organization in Great Britain. At the peak of the pandemic, ten times as many unvaccinated people were dying of COVID compared with vaccinated. I just call balls and strikes. We have a healthcare system that does not serve us well to preserve the money, power, and influence of very wealthy Americans. There are plenty of facts that support that position and we must stick to the facts to have credibility.
One of your best posts, Bill. I know, I've said that before but in all honesty I cannot envision a more thorough, clearly articulated, complete, data-based treatise on exactly the solution to the healthcare crisis. Period. This absolutely needs to reach the top echelon of our national healthcare policy makers, now. Maybe Scott Connard can help with his new organization. I would just add one little thought, and that is how/where should all these protocols be "housed". Medical Homes didn't work but they could if they were reconfigured into Porter and Teisberg's "integrated practice units" described 17 years ago in "Redefining Healthcare". Let's reinvigorate IPUs. They've worked in Europe for years.