The affordability of healthcare in an era of explosive technological innovation compounded by exhaustion of financial resources is arguably the No. 1 problem of the industrialized world. Presently, greatly beneficial care is under-delivered, while unneccessary or harmful care is over-delivered, leading to the inevitable unaffordability of all care. Verily, it is likely that modern societies are facing a healthcare financing bubble on a global basis, due to the inability to clearly distinguish between such decisions.
A New Approach
The optimal allocation of resources for the greatest benefit to a population requires a completely new approach. Innovative polices like the CMS (U.S. Medicare + Medicaid) Coverage with Evidence Development are practically unscalable because requisite information technology tools are deployed on a case-by-case basis, and re-invented each time. Indeed, healthcare IT today is a set of highly fragmented applications which do not contribute to the topology of a complex healthcare system, where micro decisions at the patient-physician level are directly connected to a macro view of the entire system as a whole.
The present approach to healthcare financing, from a business intelligence perspective, is computationally intractable and leads to arbitrary coverage and health policies in the absence of data that cannot possibly be obtained in the time frame it is needed. Meanwhile, technological developments march inexorably forward, creating a culture of endless demand for lifesaving as well as utterly futile treatments, without clear distinction, at esoteric costs. In contrast to the disorganization of healthcare, modern Internet search engines have a macro view of billions of pages of data contained in the World Wide Web, and update this world-view interpretation on a continuous (hourly) basis. Importantly, none of the millions of individual website creators designed their own content to be interpreted by a global system. Regardless, an individual web user can rapidly search and find crucial information extremely rapidly (essentially in real-time) across a disparate universe of massively heterogeneous datasets, ranging from an airline boarding pass to streaming video. Additionally, virtually all users, from public officials to grade school students to astronauts, tap into the same engine.
A macro view of healthcare, continually updated by Point-of-Care decisions now technologically exists and can be inexpensively deployed. Its availability would have immediate and profound implications on allocation of resources, patient safety, and knowledge discovery. We call this solution a Medical Operating System, to distinguish it from applications which are stakeholder-specific, microview tools.
Knowledge and Learning Architecture
In this Operating System there is realtime coupling of each of the following attributes, creating a singularly powerful framework:
(1) An open, nonproprietary thesaurus of terms which utilizes and leverages all existing standards, but serves to bridge heterogeneous ones that might never harmonize due to political or financial reasons,
(2) an open rules-authoring environment where a medical societies can publish computable guidelines related to clinical care,
(3) the same rules-authoring environment where a payer organization can publish computable guidelines for financial coverage (e.g., step therapy or prior-authorization),
(3) the same rules-authoring environment where regulatory agencies can create computable quality performance measures, that have a zero net-cost to implement, and require no new engineering effort when requirements change,
(4) a real-time rules execution engine that is publicly available, so that any patient, caregiver, or physician can obtain personalized health guidance at the Point of Care based upon the collective knowledge of the entire system at that moment in time, intersected with patient-specific data,
(5) a macroscopic outcomes transparency engine that runs continuously to measure comparative efficacy of treatments, and is made available to the public in a transparent way when difficult choices are required,
(6) symmetrical consumer-level advice to patients based on the same algorithms utilized at the POC, to allow for shared decision making on risk, benefit, and costs, and
(7) a service-oriented architecture (SOA) that can be delivered to virtually any device, application, or end-user using a simple XML protocol.
Multiple synergistic applications can be built on top of this Operating System, in distinction to an ever-increasing number of single-purpose applications that are non-additive today. Most crucially, a Medical Operating System can be incrementally adopted without wholesale replacement of legacy systems, using commodity technology and a webservices API.
Additionallly, the power of the system grows exponentially with the number of users, in accordance with Metcalfe's law. A Medical Operating System would also foster competition for healthcare outcomes (rather than care volume) among virtually every manufacturer and service provider, but be transparent and allow for unambiguous comparisons that are largely missing today. The lack of a transparent knowledge infrastructure has created a culture of distrust among stakeholders - and frank paralysis of desperately needed innovation to avert financial crises - because society has few consistently meaningful yardsticks to prioritize decisions beyond the political pressures of the moment.
Healthcare's GPS - An Analogy
The following table illustrates how just-in-time decision making is transformed using a Medical Operating System, similar to a Global Positioning System:
GPS Characteristic/Medical Operating System Analogue
Present Location
(Latitude, Longitude)
Patient Longitudinal Health Record
(Age, Gender, Weight, Procedures, Medications, Diagnoses, Lab Results)
Destination
(Address)
Effective Treatment Options for Problem X
Patient Quality of Life Preferences
Routes
(Maps and Traffic)
Marginal Costs and Efficacy of Competing Therapies
Realtime Financial Approval
Satellities
(Topology)
Quantitative World View of Like-Patients in Realtime
Continuously Updated and Refined
Implications
With a dedicated staff of 270 members and 2000 outside experts, U.K. NICE has been able to publish an average of 17.5 technology assessments per year during the past decade, while the number of new therapeutic protocols introduced during this time period was in the thousands. While insightful and commendable first-generation approaches such as NICE technology appraisals (and similar models in the U.S. such as Coverage with Evidence Development, and Congressional Bill S.3408) are very important, these strategies are not scalable or generalizable without tools that provide continuous, automated world-views of treatment value.
A world view also serves to make quantifiable, instead of arbitrary, choices on thresholds for coverage using a Quality-Adjusted Life Year (QALY) cost model. Instead of choosing arbitrary values (e.g., a $50,000 QALY threshold for coverage of a handful of formally-reviewed treatments), it is possible to determine, given a specific amount of total healthcare capital, what the consequences are, for the macro system in terms of outcomes, as the QALY threshold is changed for any treatment.
This "What-If" analysis, fundamental to virtually every traditional business, is not done today in healthcare because there is no world view system in place that can interpret the semantics of extraordinarily complex data. However, the same could have been said of the billions of terminologically inconsistent and unstructured webpages in the late 1990s.
Importantly, the tools to create this world view now exist, and can be inexpensively deployed, making public and private choices fundamentally understandable, transparent, and defensible. Only leadership and will are required to move forward.
No comments:
Post a Comment