Partnering Care Delivery and Research to Optimize Health

Tom Kottke

The five principles for optimizing care through research are summarized in table 1⇓. The first key principle is that the needs of patients and populations should determine the research agenda. Heart disease serves as an example of the importance of asking whether a particular line of research is likely to benefit our patients. Research has shown that for heart disease, the way in which health plans can have the greatest impact on preventing or postponing deaths is by reducing the prevalence of disease via nutrition, physical activity, and adherence to medications.² The evidence is clear, for example, that if we maintain patients on aspirin, beta blockers and statins, we can reduce mortality by 85% to 90%.³ The majority of deaths are caused by a few common behaviors (poor diet, inadequate physical activity, tobacco use and exposure and hazardous drinking); even the most comprehensive research programs into rare causes of disease or rare diseases will have only limited impact. If researchers are going to have an impact, we have to conduct research in nutrition, physical activity, and smoking cessation. Going forward, descriptive epidemiology should set the agenda for basic biomedical research.

The second principle for optimizing practice through research is that the research agenda must address contextual and implementation issues, including the development of delivery and accountability systems. As an example, many people think that the rise of surgery was due to anesthesia, but it was not. Rather, the rise in surgical success was the result of implementing delivery and accountability systems for maintaining good hygiene. Achieving quality improvement involves setting it as a priority, followed by changing systems to make it happen.

A third key principle is that the research agenda must determine the research methods, rather than the methods determining the agenda. We have to decide, as a country, whether a particular research area is a priority. As an example, work by Ballard et al⁴ on improving the delivery of clinical preventative services led to the implementation of multiple initiatives in Texas to increase clinical preventative service delivery, and as a result, they were able to increase “perfect” care from 20% to 50% over a period of 5 years.

A fourth important principle is that researchers and clinicians must collaborate to define the research agenda and resources, and implement the findings. A strategy that works is for researchers and clinicians to sit down beforehand and discuss what evidence is needed in order to implement an intervention. Moreover, a level of funding is needed for implementation research that is commensurate and proportional to the task. The task needs to be reframed to keep the research agenda focused on improving health and on acknowledging that the effort needs to be more comprehensive than translating research into practice. We suggest that the task be reframed as “optimizing practice through research” so that clinical practice is viewed as the client and research is viewed as the tool.

And finally, the last principle is that the level of funding for implementation research needs to be commensurate and proportional to the magnitude of the task. In order to get full value from the research that has already been completed, we need the resources to take that work one step further and incorporate innovation in clinical practice to improve patient care.

Mark Selna

I agree whole-heartedly with principle one, that the research agenda needs to reflect the needs of patients and populations. The first of three challenges that I see with this principle is that a patient’s true needs are generally not known – whether it be the usual paternalistic symptom-focused interaction between patients with their physicians or the even further academic discussions among health services researchers. A second challenge is refining those needs in a way that they actually reflect the broader needs that the payers have as well. Patient needs have to be met in a way that is economically sustainable. The third challenge is delivering against the disparate needs of individual patients. We want to avoid the experience of optimizing for populations, while inadvertently sub-optimizing for every patient in the population.

Principle two, the notion that we should address both the context and implementation issues of research, should perhaps be our first objective. The challenge lies in determining how to accomplish this. Perhaps we should focus more on operations than on biology. The notion that care management is not standardized, and as a result, is not amenable to research, I believe is a fallacy. Care management, while often very dynamic and requiring frequent and intense monitoring, can be standardized and, as such, can be studied. A primary operational issue and therefore research design consideration, is viewing the clinical care process as one of perpetual care as opposed to unrelated episodes of care; unfortunately, the level and scope of care plan development execution and oversight in our industry is woefully inadequate.

Principle three, that research agendas need to dictate the methodologies and not vice versa, is absolutely true in order to conduct true effectiveness research (as opposed to efficacy research). Important challenges in this regard are to avoid relying too narrowly on literature-based findings, and on not doing enough pre-implementation operational due-diligence. As such, it is practical, pragmatic, and ultimately less expensive to take the extra time to conduct some rapid-cycle prototyping, in the operational environment, in order to refine the operational/study design and to co-opt the necessary operational and clinical stakeholders. Ultimately, we get better research as a result.

Principle four relates to the importance of collaboration between researchers and clinicians to develop a research agenda. There is a lot of merit to moving away from a purely government-defined research agenda to one that involves clinicians, but I would submit that we as clinicians are not terribly motivated to innovate, nor are we motivated to rationalize the operational infrastructure enabling what we do. The “end user” from a translational research standpoint is not the clinician, but rather, is the patient, employer and payer. Consequently, the collaboration should be broader than simply being between researchers and clinicians.

Lastly, principle five – that funding for implementation research be commensurate and proportional with the magnitude of the task – I agree whole-heartedly. There are, perhaps, a few caveats, one being that we need to challenge people to examine the financial rationale, or economic return that justifies making an investment. This determination, for a provider organization, is not as simple as it is for health plans, where it is a dollar for dollar return on investment. Instead, we need to consider the measurable returns on the dollar investment, which can include health benefits, dollar benefits, or some other type of benefits. It is a challenge for people to think about benefits quantitatively.

Another caveat is the need to build the data capture process and the data analytic process into the operational redesign effort. Without fully knowing what research agenda might result from the redesign work, we know that better data sets, both in scope and integrity, can help drive research.

The last caveat is the importance of conducting research at a lower per-event acquisition cost by taking advantage of natural platforms, natural stakeholder groups, funding opportunities, and pay-for-performance opportunities that are inherent to the health care delivery system. Decreasing reliance on governmental sources of funding and diversifying funding will be an advantage in the future.

In closing, ideally, research should be focused on eliciting, understanding, and prioritizing a person’s needs, risks and preferences, as well as their capacity to engage over the long term. We need to understand how to obtain the data, how to capture and interrogate it in real-time, how to do so at minimal cost, how to programmatically generate novel associations in the data, and then how to deploy decision support. Lastly, while controversial within our industry but intuitively obvious to outsiders, is the need to re-design high-cost components out of our core processes.

Doug Reding

While the five principles are very thought provoking, there are some inconsistencies with how clinical practice actually functions. This difference presents some challenges when trying to optimize clinical care through research.

For example, early on in our training, scientists are taught about the scientific method, with a focus on conducting hypothesis-driven research. In practice, however, the scientific world is extremely competitive with <15% of all grant applications funded, and with the types of research that are funded largely dictated by the funding agency. For this reason, researchers should ideally pursue resources that will allow us to set our own agenda. Basic science research may be the link to true discovery going forward; this can be thought of as ‘incremental research discovery.’ Rarely does a single scientific experiment lead to significant cures in the health care industry. In clinical practice, in order for a new process to be implemented, findings from well-conducted scientific studies are needed.

Regarding the notion that the agenda dictate the analysis, this is also an interesting point. Consider the fact that 80% of health care expenditures are spent on 10% of people in the last 6 months of their lives. The scientific research agenda in the U.S. does not allocate a proportionate amount of money, time or effort on those last 6 months of life. We rarely ask clinical questions that are important to manage that 80% of health care at the end of lives. Providers have an obligation to address some of the chronic health problems, particularly if 25% of excessive care being delivered is futile.

In a patient-centric model, chronic disease management is the priority, with a particular focus on delivery systems and methodologies in order to improve quality and process for health care. If this focus is going to be our priority going forward, there are some important questions that need to be addressed: what partnerships are important? How should the system be aligned, both at an institutional as well as a national level? To answer these questions, the health care delivery system needs to be viewed as a science experiment, emphasizing principles of standardization, quality control, cost containment, and improved value to patients and clients.

At the present time, we do not have a good system in place for translating scientific discoveries and results to patients. Information systems provide the opportunity for seamless integration of information into practice. The notion that electronic medical records will increase efficient delivery of care, standardize the content of care, and improve the messages to the patients is vital and crucial. Information systems can also aid in decision support. For these reasons, we need to align the interests of information systems, medical genetics, and personalized health care in order to create a medical system that can meet the needs of patients.

Brian Rank

As care systems, our goal is to deliver ‘perfect care’ to each of our patients and hence ‘perfect care’ to the populations we serve. ‘Perfect care’ refers to delivering all the care that can help – and only the care that can help – customized to the needs, preferences, and values of each patient.

The Institute of Medicine (IOM) defines health as the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge. The World Health Organization (WHO) defines health as a state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity. In the IOM report on crossing the quality chasm, which documents many of the deficiencies of the health care of Americans today, it notes a peculiar statistic.⁵ There is a 17-year gap between when new knowledge is created in randomized trials and when the average American finally receives the benefit of that knowledge. And even then, application of that knowledge is highly unreliable.

Research by McGlynn et al⁶ has suggested that Americans receive, on average, approximately one-half the amount of health care that can help them, including preventative services, acute care, or chronic conditions. On the other hand, our present system rewards overuse of potentially harmful and dangerous health care services. Clearly the present health care system in the U.S. is not designed to reliably bring the right care and support to the right patient at the right time. That is both an indictment and a promise for research on health care delivery. There is evidence that organized and integrated care systems such as HealthPartners, Kaiser Permanente, Geisinger, and Marshfield Clinic do a better job than most; there still exists a huge chasm between the care that we deliver today and that which we could and want to deliver to our patients.

Reasons for this chasm are well-documented. A huge number of randomized clinical trials will be published this year, and to incorporate those into the health care system requires a great deal of knowledge and an ability to understand the interfaces between systems. At present, our health care systems are poorly organized to meet the health care needs of individuals and populations. Care is overly complex, consisting of too many valueless steps and uncoordinated handovers, which slow down appropriate care and make care more unsafe.

Our care system today remains an antiquated acute care model which requires visits to the clinic or hospital in which there is little knowledge of patient needs. Ideally, our health care system would include continuous, proactive outreach, based on best science, along with a patient’s biology, preferences, and values.

In the past, performance measurement in health care systems has primarily been based on retrospective measures from random chart samples from a paper medical record. This information was almost invariably untimely, inaccurate, and not granulose enough to make decisions or to influence and guide behavior. Only recently have electronic health records allowed us to report timely and accurate data for improvement purposes.

Until recently, we spent most of our research agenda on generating new knowledge, rather than on figuring out how to reliably implement new knowledge. We are at the infancy of understanding how to reliably and systematically implement care consistent with the IOM’s six aims for improvement.⁵ In order to accomplish this, the research agenda now must include not only traditional research techniques, but also an understanding of care process variation, along with reliability, science, and human systems engineering. Unreliability means failure to meet processes over time; our ability to do that is the difference between the health care that people could receive and that which they are receiving today.

A research agenda also needs to encompass an understanding of how human biological and cultural diversity affects care. We know that measured quality of care and satisfaction with the U.S. health care service vary by race, ethnicity, language, and socioeconomic status. In our care system, we have obtained language preferences for 99% of our 400,000 patients, and 84% have trusted us with their race and ethnicity data. This database can help us to reliably understand how to customize care to the preferences and values of patients, informed by best evidence. It recognizes the deep and respectful relationship between a patient and clinician, with an appreciation of how complex care is at the point of care. Understanding why things happen the way they do at the point of care can bring researchers and clinicians together to advance a new research agenda.

The challenge then is to understand how to redesign our systems to support patients and clinicians in an effective continuous healing relationship. When patients, clinicians and researchers effectively collaborate, we better understand how to optimize practice, to create high performing teams and thus a high performing health care system. For our organization, the close integration of care and research comes from our history and is hardwired into our organizational structure. Our health plan, research foundation, and care system are cross-pollinated by researchers, clinicians, and leaders across the hospital, medical group, dental group and health plan, and our institute for medical education.

Well-respected clinicians serve in senior leadership positions within the medical group and research foundation, and senior researchers recruit and mentor young clinicians in clinical care and care improvement research, creating a deeper understanding of the pressures facing care teams as they seek to deliver ‘perfect’ care to their patients. The close relationship between clinical care and research not only accelerates improved care for our patients, it also brings in tangible benefits to the organization. It clearly helps recruit excellent clinicians and helps build more professionally rewarding careers based on outcomes for patients. The ability to improve care, outcomes, and practice efficiency is seen as an important benefit to being a physician or dentist in our care system. For us, the challenges of linking care and research more closely are probably the same as for others, and these challenges primarily involve time and money. Clinicians in efficient practices today spend most of their time and energy in clinical care, and to incorporate clinical research into an already stressed life takes both effective mentorship and financial support. To that end, in our organization we have created partnership grants to support that relationship between clinicians and researchers. To truly succeed, however, and to change the way in which health care is delivered, we need input from patients.

Appendix: Question and Answer Session

Acknowledgments

Funding for this conference was made possible in part from the Agency for Healthcare Research and Quality (1R13HS017518-01) and the National Cancer Institute (1R13CA128384-02), and by HealthPartners, HealthPartners Research Foundation, HealthPartners Medical Group & Clinics, Marshfield Clinic Research Foundation, Marshfield Clinic, and Security Health Plan. The views expressed in written conference materials or publications and by speakers and moderators do not necessarily reflect the official policies of the Department of Health and Human Services, nor does mention by trade names, commercial practices, or organizations imply endorsement by the U.S. Government.