Improving patient-centric measurement in value-based careLessons from oncology
The healthcare industry is shifting towards value-based care models, but how we measure value is still stuck in the past. Over the last few decades, healthcare has built up a preponderance of process measures, but remains lacking in effective outcomes measures. Despite widespread adoption of EHRs and the technology needed to make data interoperable, most health systems still fare poorly at measuring anything that happens after the patient leaves their facilities.
At the level of patient care, this means that we’re capturing isolated snapshots based on people’s interactions with the system, most commonly in the forms of claims data and some sporadically collected clinical values. The industry has just now begun to expand its focus beyond these processes of care to connect patients’ day-to-day experiences of health, illness, injury, and the symptoms and behaviors involved therein.
At the level of clinical practice, this means we’re measuring a lot of activity on the part of clinical care staff, but that we’re leaving out factors important to the success and scalability of new care models: sufficient staffing and infrastructure, levels of communication, transitions of care, and more. Such considerations are still process-oriented, but they play an important role in setting up value-driven systems, because they start creating organizational self-awareness around the “how” and the “who”’ of reform, rather than a myopic focus on the “what.”
The Center for Medicare and Medicaid Services’ (CMS) efforts have been getting better in recent years after a rocky start. Most of the measurements baked into the Meaningful Use program were high-level checklists about physicians’ clicks, and new requirements on patient-generated data have simply failed to materialize. The measures to document care coordination in the chronic care management (CCM) were considered so burdensome to complete for clinical practices that they effectively canceled out the value added by any potential reimbursement.
However, more recent efforts, notably the Comprehensive Care for Joint Replacement Model (CJR) and the Diabetes Prevention Program (DPP), are helping introduce new measures around patient follow up and engagement. In an encouraging sign, these federal efforts have begun crossing over into commercial insurance, helping kick start the shift towards outcomes-driven digital health.
Providers are starting to embrace these new measures as part of value-based care models, although they’ve got a lot of work left to do. According to a CMS survey:
Less than 20 percent of hospitals currently follow up on their patients by collecting outcomes data between or after treatment.
In part, this is because the measurements are confined to very specific pockets via one-off payment programs, such as joint replacements. Even the best Accountable Care Organizations (ACOs) are just breaking even financially and seem content to maintain a tight focus on the core process measure sets for which they’re paid as they figure out the challenges of risk adjustment, payment benchmarks, and operational efficiency.
The question arises as 2017 gets underway: Outside of the carrots and sticks approach from Washington, are hospital systems going to start taking a more holistic, self-deterministic approach to value-based measurement and care delivery? Will they overcome traditional, institutional barriers and find new ways to connect data to care?
For insight into what the future may hold, we can learn a lot by looking at cancer care.
Dr. Thomas Feeley, chair of MD Anderson Cancer Center’s Innovation Institute, argues that value-based care requires a different framework for measurement.
This is particularly true in a complex space like cancer care. Many cancers and other genetic diseases prove doubly challenging to measure outcomes, because in addition to changing how and who we pay, we’re changing the what. The very models of care are evolving, with new treatment pathways, medicines, recommendations and guidelines, along with our fundamental understanding of these diseases themselves.
This means that cancer care is a particularly ripe area to introduce a new wave of patient-centered measures that can help guide care delivery towards optimal outcomes for patients and health systems. Moreover, while oncology is a specialized, and in many ways, different field than the broader market of non-specialized acute care hospital systems, similarities are striking. According to Dr. Arlene Chung from the University of North Carolina School of Medicine:
“This vision for a learning cancer care system centers on data coming from individual patients and their cancer care teams. With 90% of oncology practices reporting that they either already have or plan to have an electronic health record (Electronic Health Record (EHR), as defined in Defining Key Health Information Technology Terms (The ... More) and advancements in Electronic Health Record (EHR), as defined in Defining Key Health Information Technology Terms (The ... More functionalities, the necessary technical infrastructure is now available to realize the potential for PGHD to enhance quality cancer care and build learning cancer care systems. Whereas much progress has been made in collecting data from clinical care teams via EHRs and cancer registries, there is still a critical gap in capture of information from patients/caregivers. PGHD integrated into EHRs could address this important gap and make this information available for clinical care, research, and quality improvement.”
If you take out the word cancer, that sounds an awfully lot like the overall healthcare IT landscape. For a closer look at how oncology and cancer treatment systems are approaching value-based care, let’s turn to cancer-specific patient reported outcomes measurement and value-based care planning.
Patient-reported outcomes for value-based cancer care
Dr. Amy Abernethy is one of the pioneers of patient-centered cancer care. In particular, Dr. Abernethy has championed the application of patient-reported outcomes to cancer care and helped define today’s clinical understanding of post-treatment survivorship. She wears multiple hats, including serving as an appointee to the Institute of Medicine’s (IOM) National Cancer Policy Forum and as the Chief Medical/Scientific Officer and SVP of Oncology at Flatiron Health. Last summer I had a chance to ask her for a lay definition of value-based oncology:
“Oncology is not knees and hips. It’s relatively easy to understand how you’re going to operationally improve with hip replacements. Oncology is hard. It’s complex science, complex personal lives, highly variable from individual to individual. How do you translate that into a system of improving value?
The current model du jour that’s being tested, commercially and at CMS — is looking at a number of pretty fundamental things. It’s making sure that doctors and patients — and when I say patients I also mean families — are on the same page as to what the plan’s going to be. So, [we’re talking about] care planning. It seems simple, but it’s still not a standard part of oncology care, despite the research that’s shown it reduces hospitalizations.”
If “oncology is not knees and hips,” then what’s so different about a care plan for cancer patients as compared to a care plan for someone getting their knee replaced?
For one, care plans for cancer patients need to measure different factors at different points of the treatment journey, and those factors may have different target values or ranges at different points in time, from before chemotherapy, during treatment, and into the post-treatment survivorship phase. Clinical trials and other experimental treatments come with a litany of additional biometrics, symptoms, side effects, and so forth.
Beyond the clinical components of managing a particular disease, care plans for any particular patient require contextual grounding in a very practical sense. Cancer patients, for example, are more likely than diabetics to travel to large treatment centers like MD Anderson or Dana Farber for their care. This raises the need for better data sharing, more portability, and the flexibility for numerous clinical experts to co-author and periodically update different parts of the care plan.
Moreover, different diseases (cancer in particular) come with the need for specific education, instructions, goal setting, and personalized support, all of which tend to evolve over time in a way that patients (and their care team) can understand and keep up with. So, there’s much more than measurement involved, which we’ll discuss later.
Federal efforts are driving progress
One area that holds promise for both promulgation of better outcome measures and care planning in cancer more generally is the Oncology Care Model (OCM), a CMS initiative. The OCM requires a documented care plan that contains 13 components outlined by the Institute of Medicine.
Components of a Cancer Care Management Plan from the IOM
- Patient information (e.g., name, date of birth, medication list, and allergies)
- Diagnosis, including specific tissue information, relevant biomarkers, and stage
- Treatment goals (curative, life-prolonging, symptom control, palliative care)
- Initial plan for treatment and proposed duration, including specific chemotherapy drug names, doses, and schedule as well as surgery and radiation therapy (if applicable)
- Expected response to treatment
- Treatment benefits and harms, including common and rare toxicities and how to manage these toxicities, as well as short-term and late effects of treatment
- Information on quality of life and a patient’s likely experience with treatment
- Who will take responsibility for specific aspects of a patient’s care (e.g., the cancer care team, the primary care/geriatrics care team, or other care teams)
- Advance care plans, including advanced directives and other legal documents
- Estimated total and out-of-pocket costs of cancer treatment
- A plan for addressing a patient’s psychosocial health needs, including psychological, vocational, disability, legal, or financial concerns and their management
- Survivorship plan, including a summary of treatment and information on recommended follow- up activities and surveillance, as well as risk reduction and health promotion activities
While these recommendations don’t contain many actual outcome measures, they do harbor provisions for patient engagement, collaborative goal setting between patients and doctors, and other “hooks” where patient-reported outcomes data could help drive value. And importantly, this new model is being deployed at 190 practices around the country.
As Dr. Feeley points out: “The inclusion of the care plan in the OCM does represent CMMI’s (the Center for Medicare and Medicaid Innovation) ability to adopt measures not previously vetted by the NQF and MCS in a pilot program that may set the way for more rapidly testing true outcome measures in upcoming CMMI programs.”
This warrants a quick moment of reflection: A bunch of experts just came together, decided what was important to put in every patient’s care plan, and then CMS just did it. Compared to the usual processes at the National Quality Forum (NQF), and the federal government more broadly, the speed of this approach feels unprecedented. Look at the above list again, and imagine if similar approaches for onboarding patients onto a self-management plan existed more broadly for disease management programs in areas like diabetes or hypertension, or even just primary care in general.
And this is not to take away from the NQF’s dedicated leadership – to their credit, they are continually researching and developing dozens of metrics and measure sets, and have even introduced a “measure incubator” to accelerate measures where there are significant gaps in the field.
It’s important to appreciate the substantial work and effort that has gone into getting to this stage – introducing new measures, tying them into a care plan methodology, and moving them from theory into practice (again, 190 sites). Yet, at a time where the CMMI and other federal initiatives on value-based care face an uncertain future in the new government, it’s equally important to ask whether the private sector is ready to grab the baton and drive the next phase of progress of value-based, patient-centered care planning, measure collection, and engagement.
Private sector: Letting a thousand flowers bloom
Remember “all that stuff” above: The litany of practical and disease-specific considerations that go into making a care plan fit into patients’ lives? The good news is that this is the stuff with which industry, not government, is uniquely suited to drive the next phase of progress.
Since Dr. Feeley leads MD Anderson’s innovation center, let’s start by taking a look at what they’ve been up to. Even just a cursory scan of their activity in the last few months reveals numerous efforts around value-based oncology models with a strong patient-centric engagement component:
- A study found that patients using a digital platform by HealthLoop showed higher levels of education, preparedness, and engagement in their treatment decisions.
- A project to test a 1-year bundled payment model for head and neck cancer includes 28 PRO measures that are administered through an Electronic Health Record (EHR), as defined in Defining Key Health Information Technology Terms (The ... More/patient portal, and covers a window of over two years in patients’ lives.
- A clinical trial in partnership with Vida seeks to provide: “monitoring, guidance, accountability and social support in between doctor visits to help them manage their conditions and achieve lasting behavior change.” Vida’s platform combines mobile coaching with evidence-based clinical programs, devices, and data collection.
- A framework “to rapid-develop, at an accelerated pace, comprehensive disease-specific outcome measure sets, including provider-generated outcomes and PROMs, that could be integrated with electronic health records (EHRs) and incorporated into clinical practice.”
- An inventory of multi-symptom PRO measures with additional clusters of metrics specific to patients being treated for more than a dozen types of cancer.
This is far from an exhaustive list – as might be expected of one of the world’s leading cancer treatment centers, MD Anderson has dozens of research initiatives around some application of PRO, as well as many disease-specific measure development efforts. It’s highly likely that similar lists will be found at cancer treatment centers like Memorial Sloan Kettering, Dana Farber, Fred Hutchinson, and so on.
While the philosophy of letting a thousand flowers bloom sounds great, in practice it amounts to a sort of industry isolationism, as opposed to the level of leadership that numerous groups (IOM, NQF) put into developing the OCM’s framework.
Many of MD Anderson’s initiatives appear distributed across numerous silos of different researchers, pilot projects, and departmental efforts. Some efforts involve working with competing digital health startups. For example, MD Anderson did a small-scale engagement study with Healthloop – do they know, or care, that Healthloop has also deployed a collaborative cancer care planning tool commercially, with Memorial Sloan Kettering in New York? Might this little company know a thing or two that MD Anderson could benefit from, too?
And despite their work with those mobile health startups, MD Anderson persists in their use of EHR-tethered surveys in their innovative payment pilot with United Healthcare, instead of patient-friendly tools available inside of their own walls.
It should be noted, however, that new methods of health data exchange, such as HL7 FHIR and JSON, have made Interoperability refers to the ability of two or more systems or components to exchange information ... More between patient-facing apps and provider-facing EHRs possible. This type of patient-EHR interaction would have been close to impossible as recently as 2015.
If we zoom out, there are an expanding array of new measure sets being issued by consulting firms, a cadre of federal organizations and independent associations, and other groups. Some agencies like ICHOM take an open-source approach, but private sector institutions treat their work on patient outcomes measurement as proprietary.
Moreover, there’s little evidence of collaboration with the outpatient segment of cancer care providers, such as involvement in the American Society of Clinical Oncology’s (ASCO) data-sharing initiative, CancerLinQ. In cancer, including “the other half” of the care continuum becomes particularly important for post-treatment survivorship, re-integration into primary care, mental health care, social services, and so on. It remains true in healthcare, as in many other industries, large organizations who embrace a brand-first mentality are also succumbing to siloed thinking.
Strategic Data Framework Whitepaper
How is it possible to become innovative during such tumultuous times?
The latest white paper from Corepoint Health takes a look at how the changes are impacting IT departments, which are quickly learning that creating a strategic data platform allows the needed versatility to adapt and accept changes quickly.
Conclusion: How healthcare can grab the baton
The federal government has been hard at work over the last several years, standardizing a new wave of measurement to help define value-based care models – in and beyond cancer care. While we’ve undoubtedly made substantial progress in just a few years, there is a lot of work ahead, from measure selection and development, to field-testing and implementation, to technological integrations and product management, and so on and so forth.
Yet, as soon as these measures are deployed in the field in the form of value-based care programs that are funded by public and even some commercial payers, it will be up to the private sector to grab this baton and run with it. It is at this stage of the race that the bigger challenge comes into view: How can industry turn one-off, scientifically validated instruments into comprehensive real-world solutions that generate value for patients, doctors, and system administrators?
- Provider systems can do a better job of streamlining their approach to selecting startups and technology partners. Is it wild to imagine one or two qualified technology partners operating the majority of an institution’s research pilots? Would that make it easier to take findings from those pilots and introduce them into clinical practice more quickly?
- Perhaps this could be a leadership opportunity for the litany of innovation centers that have sprung up around the nation. Issuing RFPs, running hackathons, or reverse pitch events to find qualified help, and rolling up their sleeves to turn qualified companies into strategic partners (perhaps in exchange for equity), rather than disposable, replaceable, temporary pilots.
- Payers play a key role in continuing to work with delivery systems to develop innovative models and measures. One example is the bundled payment pilot described above, which was a partnership between MD Anderson and UHC. Payers can also do a better job of making it easy for startups to find and apply to be a part of these opportunities – an easy example is finding proven startup partners (e.g., DatStat, RoundingWell) to come in and deploy mobile data capture tools.
- Startups innovating in this space should set realistic expectations about working with large medical systems on clinical research pilots. While a small pilot may provide great opportunities for everything from field testing to marketing, these are typically unlikely to result in a scaled up enterprise deployment. Startups (and investors) should plan and deploy resources accordingly.
- Traditional health IT vendors have their work cut out for them as the foundational EHR systems in place across most of the country are figuring out how best to handle the new wave of patient-centric measures. Validated questionnaires like PROMIS are already available within many existing EHRs, but these still tend to lack mobile-friendly design, workflow-friendly integration, and importantly, patient-facing summaries. Without a modern integration platform in place, it is untenable for any Electronic Health Record (EHR), as defined in Defining Key Health Information Technology Terms (The ... More to wrangle the numerous data standards and proprietary architectures, particularly when you add in the non-traditional realm of wearables, behavioral data, and emerging survey tools.
Naveen is the Founder and Managing Partner of Patchwise Labs, a digital health strategy firm dedicated to improving healthcare for the patient using technologies, approaches, and ideas from all sectors. Naveen’s passion for healthcare centers on patient rights and the emerging consumer movement in healthcare. His background spans consulting and policy analysis, patient advocacy, health IT and digital health. He also serves as Managing Editor of Tincture, a thought leadership publication for important ideas in healthcare and medicine. Connect with Naveen via Twitter or LinkedIn.