In 2015, then-President Barack Obama launched the Precision Medicine Initiative,1 establishing ethical and technical rules of engagement for using health-related data to drive more effective and individualized care. The goal of the initiative is to turn massive amounts of healthcare data into actionable information. In theory, knowing more about a patient, including one’s genetic predispositions, may help us understand who will develop certain ailments, who is likely to have complications, and who may benefit from particular therapies. Getting it right can aid earlier intervention, improve outcomes, and contain costs. Collecting and generalizing the data, however, are not easy.
The data must come from disparate and competing entities, including insurance companies and large health care networks. Amassing it may prove too onerous without hefty government incentives. In addition, the bulk of the available data applies to the management of chronic diseases, and may not be applicable to our relatively young, predominately healthy, military patient population. If precision medicine is to be applied to the warfighter, the bulk of the data supporting it must come from within the Department of Defense (DOD).
Though some naval ships have large medical facilities, we cannot rely on a robust network of medical clinics and hospitals like those that existed ashore in Iraq and Afghanistan. While supporting U.S. interests worldwide, deployed forces can expect longer evacuation times—problematic for burn, blast, and blunt trauma victims, the most common serious shipboard injuries. To deliver care, we must foster innovation and develop adaptive capabilities designed to change outcomes and save lives.
To turn clinical data into decisions, we must recognize patterns within healthcare data and create clinical decision support (CDS) tools. As the name implies, these are designed to aid clinicians in the decision-making process. For instance, data can help estimate the likelihood of bad outcomes (like infection) so preemptive treatments may be given. This allows a shift from reactive medicine a more proactive approach.2
Predictive medicine is similar to intelligence work, interpreting data from a variety of sources. The goal is to mitigate risk, take action when necessary, and save lives. Just as intelligence officers of the past sifted through all available information to answer questions about adversaries, clinicians once observed, examined, and tabulated as much information as possible, poking and prodding their way to a diagnosis. Success depended almost entirely on the human brain’s ability to assimilate scarce information. Today, however, the opposite scenario applies. The volume of intelligence and health-related information has become too vast even for teams of analysts and clinicians to distill.
As the ability to sense and image the world improved, the intelligence community embraced computer-intensive methods to tame vast amounts of data. Predictive models became a new tool in the intelligence tool kit, and decision support became a strategic capability.
The human body has many adversaries that can emerge after trauma or illness, such as insulin resistance in Type II diabetes or antibiotic resistance in bacteria. The ability to measure and visualize human physiology, immunology, and pathology far surpasses one person’s abilities. Intuition alone often makes the difference between success and failure, diagnosis or not, and in some cases, life or death. The best care often comes from multidisciplinary teams, but even our best teams have difficulty collecting—let alone assimilating—all available data, in part because the manner in which we display and share it is obsolete. As we implement “Precision Medicine” within the Navy, the medical corps can take a lesson from the intelligence community.
Clinical Decision Support as an Operational Capability
The process of turning medical data into decisions is within our grasp, using genomic and other data to the fullest extent. Knowing one’s genotype, however, is far less important in battalion aid stations (BAS) or on board casualty receiving ships than other parameters. Interventions in the acute setting are guided by the patient’s physiologic reaction to combat trauma, disease, or non-battle injury. We are unlikely to consult the genome to stop bleeding, triage patients, or prioritize medical evacuations.
This is good news for the Navy. Sequencing the DNA of all beneficiaries is untenable at present, but there are ample physiologic, treatment-related, and experimental data to be had. Some guide treatment after blast injuries, determine whether to give costly but life-saving blood transfusions, or predict which wounds are at risk for nasty complications. Another tool is designed to lend objectivity to the difficult question of whether a patient is better served by saving a limb or amputating it. Finally, a suite of tools is being developed to predict life-threatening conditions such as sepsis and heart failure, which are common after wartime injuries. DNA plays important roles in some conditions—like cancer, heart disease, and dementia—so knowing when and when not to focus on DNA is important. Investing in CDS as an operational capability would benefit the entire spectrum of care “from the BAS to the ICU.”
A Lack of Standardization Impedes Implementation
Various groups within DOD are developing CDS tools. Though some are better at extracting data than others, the usual method involves expensive teams of researchers who pore over medical records, laboratory information, images, and operative reports to fill in the blanks. Scavenging data from medical records is akin to a dumpster dive for information. The process is time consuming and error prone, requiring sophisticated quality control measures to separate good data from trash. Furthermore, each current DOD program invests a considerable amount of time, energy, and money in proprietary methods of data collection and storage. As the cost of managing and securing data rises, efficiency suffers. Even DOD-funded programs can spend in excess of $10,000 per patient enrolled, and development costs are prohibitive.
To compound the issue, military CDS developers are entrenched in a research-and-development (R&D) paradigm. Groups who compete for funding fight to protect their hard-won intellectual property. It is difficult to find fault in this, but none share data on any meaningful level, or agree on how CDS tools should be vetted. Building cohesion will not be easy without strong, central leadership and creative incentives.
DOD CDS programs have the potential to give tremendous value to military medicine. They exist on the frontlines of research and are led by seasoned clinician scientists. Investigator-initiated research is innovative by design, and a diversified approach to problem solving can be quite effective. However, if the goal is to bring CDS to the warfighter, the Navy must standardize certain processes and ensure the evaluation, regulation, and implementation of the tools remain consistent.
“Quality Improvement,” Not “Research”
The Navy cannot afford to rely solely on the medical R&D community to develop CDS tools. The problem is time. It takes nearly two years to obtain funding for a research protocol, up to one year to obtain institutional approval to collect data, and about three years to conduct a study of sufficient size. Under these constraints even “new” CDS tools are based on ideas and data that are several years old. Furthermore, each tool would still require “field testing” prior to implementation. Acting on outdated intelligence is risky. Trusting the health of our sailors, soldiers, airmen, or Marines to an outdated treatment paradigm is not an option.
We must quicken this process. Decoupling CDS from the research environment will allow the Navy to redefine it as a quality improvement solution, ease regulatory restrictions, and make the process of turning data into decisions more efficient. The standardized development environment shown in Figure 1 includes a common data repository analogous to the “Information Commons” advocated by the National Research Council.3 This quality improvement database may be designed to accept data from a variety of sources including the electronic health record (EHR), clinical trials, and physiologic monitors. Standardizing the manner in which data are de-identified and stored is analogous to the open systems architecture used by the Navy. The approach is known to stimulate innovation and competition while alleviating challenges associated with home-grown or proprietary data management schemes. Fast Healthcare Interoperability Resources (FHIR, pronounced “fire”) is a capability already supported by DOD and the Veterans Administration, and its level of standardization may not be a bridge too far. By condensing the process from years to months, tools may be field tested sooner and improved as treatment philosophies evolve and new therapies emerge.
Ease Regulatory Pressures
There is little consistency in how various groups within DOD define CDS requirements, and no consensus on how tools should be vetted prior to use. This paralyzes implementation. Furthermore, to deliver the best and most efficient care anywhere in the world, independent cost savings and operational impact analysis must be performed to ensure the tools make sense. For instance, tools are being designed to expand the capabilities of organic healthcare assets. For burns, blast, or other shipboard trauma, objective data will guide treatment afloat, prioritize evacuations, and safely extend evacuation times for others.
Unfortunately, DOD’s own CDS tools in development have no clear path for implementation. Most end up in peer-reviewed journals to establish necessary academic credibility, but create issues for clinicians overwhelmed by paperwork. To circumvent this, most groups now place their tools online, but the process of manual transcription is frail, prone to error, and useless for operational providers who lack continuous Internet connectivity. In addition, the U.S. Food and Drug Administration approval process for CDS tools attempts to evaluate each CDS tool based on risk—a process already applied to medical devices. However, the amount of evidence supporting an FDA application varies, based on the significance of information being generated by the tool, and the context in which it is used. Regulatory approval can be long and arduous for those who choose to go it alone, even for well-funded DOD programs.
The alternative is to apply a judicious dose of standardization and oversight. Just as we have an obligation to develop CDS tools that work for all our beneficiaries, we must standardize the way each is evaluated and implemented—either in the EHR or in stand-alone versions on board operational platforms. Insisting on rigorous standards and a conduit for implementation will streamline regulatory approval and encourage research and participation within DOD, academia, and industry.
The Navy is in a position to drive the development of precision medicine solutions because it must deliver care across a diverse network of platforms and environments. Implementing it properly will standardize care, improve outcomes, contain costs, and extend the capability of operational providers. If we are to develop CDS as an operational capability, we must first develop the infrastructure to evaluate, regulate, and implement the tools that will make combat medicine more effective and efficient.
Author’s note: The author thanks Commanders Eddy Ha and Andy Hunt, U.S. Navy, for their insights.
1. The White House. Precision Medicine Initiative: Privacy and Trust Principles. 1–4 (2015).
2. Forsberg, J. A., Potter, B. K., Wagner, M. B. & Vickers, A., “Lessons of War: Turning Data Into Decisions,” EBioMedicine 2, 1235–1242 (2015).
3. National Research Council Committee on a Framework for Development a New Taxonomy of Disease. Toward Precision Medicine: Building a Knowledge Network for Biomedical Research and a New Taxonomy of Disease. 1–143 (2015).
Commander Forsberg, MD, PhD, is an orthopaedic oncologist and former Marine who leads a team of clinician scientists focused on clinical decision support and improving the lives of wounded servicemembers.