Blogs

Dr. Ezekiel Emanuel’s article is a description of an inefficient healthcare ecosystem and a plea for better tools to counteract the forces causing inappropriately high health care costs. Dr. Emanuel implicates behaviors of providers and patients as causal mechanisms. Dr. Robert Califf goes even further claiming the ecosystem is imbalanced and this causes harm, primarily by implicating pharmaceutical pricing caused by charging what the market will bear. Dr. Emmanuel embraces the entire health care ecosystem operating to cause high health care costs due to mechanisms collectively deemed behaviors. Of the four main categories of cost, administrative, procedures, imaging, and pharmaceuticals, each is at risk of excessive use because of modifiable behaviors. Exploring the many causes of behaviors, the ecosystem offers an opportunity for asking questions that identifies these causes, regarding the ecosystem and the patient holistically.  On the other hand, Dr. Califf has a reductionist view of suboptimized value. In this example, inappropriate pharmaceutical pricing is a mechanism of deviance from value. The details of cholesterol effects on health, pharmaceutical company business practices, indifference of providers are all details deemed in their collective effect an imbalance of the ecosystem. The effect of the imbalance is the failure to optimize cholesterol levels, a reductionist outcome.  This brings to mind Einstein’s comment, “God does not throw dice.” The dice evoke the potential causes and outcomes of the patient centered ecosystem. The present thrust of value based purchasing is to modify reductionist outcomes, but tools to modify widespread broken behaviors do not exist.  Noting what the dice show before and after they are thrown will suggest these tools; the resting pair of dice show many potential outcomes, the thrown dice narrows to a single outcome. This difference is the key to isolating and modifying behaviors, which depends on integrating reductive and holistic approaches₃, as in the throw of the dice.

  The potential solutions for broad ecosystem balance coming from population policies, if balanced with actionable, specific interventions can succeed where using one approach without the other will fail in achieving population scale cost reduction. Balancing these approaches can benefit from introducing causal paths and counterfactual analysis into health care thinking₄, allowing comparisons of single patient, single event actions to similar actions of larger populations and  can include information based on factors such as provider experience, social determinants of health, and psychosocial effects on outcomes. These factors registered as causal paths will allow all sources of information in the moment to be complementary. With context, there is an inclusion of what did not happen as a counterfactual, which is essential if value is to include avoiding inefficient or more costly paths. The essence of casual paths when compared at the population level is to allow prospectively selection of the most optimal paths. Predictive modelling supports the reductive view, whereas the holistic view can include a much broader and diverse set of variables when relevant at the patient level.

 Because each piece of the healthcare ecosystem exists in relation to a complex web of entities₅, a holistic view of the ecosystem will work better than each entity is isolation, again an example of the difference in the reductive and holistic approach.  For example in a holistic perspective, entities of large scale will not overwhelm or control information of other entities that are of a smaller scale.  When patient centered events are a basic fundamental component of the ecosystem, the granularity and sheer number of events isolates them individually from the ecosystem in its largest sense. A holistic view equates scale as a fundamental property of entities that will always exist where the small and large scales will be visible to other entities in the ecosystem. To illustrate this difference of scale is the use of population codes for event billing where plans determine reimbursement from the codes. Rarely do diagnostic codes totally encapsulate the reality and completeness of the event, because of the reductionist flavor of the coding methodology.  This difference between the large plan scale and the small patient-provider scale at the level of single events encapsulates the difference in what is real, known or unknown, certain or uncertain, for entities regardless of scale. For a patient centered ecosystem, which allows the flavor of holism to be recognized, the challenge is to recognize as well all factors that exist in the multiscale ecosystem relevant to the fundamental level of the event. Therefore, the context of the event is any factor, or in a technical sense any variable, that is background for the event.  To add a notation to this complexity at any scale is the objective of creating context of granular events one event at a time, even accounting for background factors that are measurable or not.

  A comparison can be made of Dr. Emanuel’s holistic objective to Dr. Califf’s reductionist approach in showing how context for events differ and color what is considered known and real as a function of scale of the entity. Optimizing the ecosystem for value starts with accepting the potential complexity of value if optimization in the standard. Clearly, value production for the reductionist system of reimbursement coding is a simplification of context but is effective in authenticating services of a single event nature. However, adding context holistically to events allows a more complete version of value framed by a selected version of the ecosystem. In fact, there can be a variety of versions of the ecosystem, fitting the needs at hand to target communication of value to those parts of the ecosystem of particular relevance. This more complete version of value can address Dr. Emanuel’s vision of looking at behaviors that are provider and systemic levels of conglomerations of mechanisms that drive overutilization of selected specific components of the ecosystem amenable to value based scenarios.  

                This discussion leads to a consideration of what causes behaviors. As alluded to above, this requires a causal interpretation. With the fundamental unit of healthcare established as event based, with the contextual background variables existing from any source at any scale, then application of causal path analysis is possible. The paths one will note will wind its way through the ecosystem. When patient-centered they include complex journeys with clusters of providers over time, varieties of diagnoses, risk status for diseases, and so forth. Paths that are provider centered tend to be selected populations from which medical evidence is determined. Motivations for provider clustering such as market power from vertical integration, informal referral networks, coordination of care, and collaboration will exist as mechanisms that affect clinical choices with cost impact. Plan centering reflects population metrics of all types, particularly global costs and patient access to care. Thus, value optimization depends on optimization centering. Path analysis has the potential to put all this together simply by deriving all quantitative tools from the fundamental event base, which allows full description of holistic, and reductionist complexity.  

                The challenge of the health care ecosystem is to pull back the obscurity and promote transparency. When framing value not as a highly reduced objective, but as a holistic objective, then the wisdom of Einstein in exploring the real and unknown world of quantum physics can help us to understand new ways to approach a vast world of the healthcare ecosystem.

!.  Glickman A, DiMagno SSP, Emanuel EJ. Next Phase in Effective Cost Control in Health Care. JAMA. 2019; 321(12): 1151-1152

2. Califf RM. Balanced Dysfunction in the Healthcare Ecosystem Harms Patients. Circulation. 2019; 140(23): 1860-1864

3. Greene JA, Loscalzo J. Putting the Patient Back Together – Social Medicine, Network Medicine and the Limits of Reductionism. NEJM. 2017; 377(25): 2493 – 2499

4. Pearl J. The Logic of Structure-Based Counterfactuals Chapter 7. Causality, 2014; Cambridge University Press: 201-257

5. Lipsitz LA. Understanding Health Care as a Complex System. JAMA. 2012; 308(3): 243-244

6. Berwick DM. Elusive Waste The Fermi Paradox in US Health Care. JAMA. 2019; 322(15): 1458-1459

7. Schwarze ML, Taylor LJ. Managing Uncertainty – Harnessing the Power of Scenario Planning. NEJM. 2017 377(3): 206-208