Personalized Medicine — A Fantasy without Patient Directed Data
Personalized medicine is a medical procedure that separates patients into different groups — with medical decisions, practices, interventions and/or products being tailored to the individual patient based on their predicted response or risk of disease. Personalized medicine, with its ability to tailor diagnostic and treatment decisions for individual patients is seen as the evolution of modern medicine, and the terms personalized medicine, precision medicine, stratified medicine and P4 medicine (personalized, predictive, preventive, and participatory medicine) are often used interchangeably.
Why Personalized Medicine
Most drugs don’t work. Statistics show that among the 10 highest-grossing drugs prescribed in the US, even the best work in only one in four patients. Some work in only one of twenty-five. Statins, commonly prescribed cholesterol drugs, work correctly in only one in fifty patients. Now consider that the cost to produce a new drug is $2.6 billion and rising 8.5% every year. So, we spend billions of dollars every year inventing new drugs that work for less than 25% of the target in the best of cases. If there is ever a prize for the best “What is wrong with this picture?”, this would be it. There are multiple reasons for this, but the basic reason is that our different genetic makeups (genome), proteins in our body (proteome), and body flora (the bacteria that grows inside of us) affect how drugs work.
Doomed without Data
The informatics resources that can support clinical interpretation for personalized medicine must come from genomic test results, mobile internet real-time data and electronic health record data. Personalized medicine must use all of these three Big Data.
However, to create and dispense personalized medicine, researchers and clinicians must navigate and integrate clinical information (e.g. medical diagnosis, medical images, patient histories) and biological data (e.g. gene, protein sequences, functions, biological process and pathways) that have diverse formats and are generated from different and heterogeneous sources. Add to that all the Patient Generated Health Data (PGHD) from Wearables and other devices, and you can see how this becomes an untenable situation. Today, no single knowledge base can currently support all aspects of personalized recommendations. It is up to the provider or researcher to consolidate several resources and findings across a variety of data sources and then analyze and interpret this data. For personalized medicine to become a reality, a more dynamic, consolidated and collaborative knowledge base is absolutely essential.
Data integration and making use of different data sources is at the core of personalized medicine. The only vehicle through which we may truly realize the personalization of medicine, beyond population-based genetic profiles, is a patient centric, cross-institutional and longitudinal information entity possibly spanning the individual’s lifetime. It provides physicians and researchers an integrated view across genotypes and phenotypes for establishing new patient-stratification principles and for revealing unknown disease correlations.
As personalized medicine is practiced more widely, a number of challenges arise. The current approaches to patient privacy and confidentiality, ownership and control of data, reimbursement policies, as well as regulatory oversight will have to be redefined and restructured to accommodate the changes personalized medicine will bring to healthcare.
Data Aggregation Challenge
Although all Electronic Health Record Systems (EHRs) are required to provide a Continuity of Care Document, not all EHRs are open and not all formats are easily aggregated. With over 300 different EHR systems in use today, when visiting a doctor, the average patient’s experience is recorded within at least 3 different information systems of record. That same patient also sees 18.7 different doctors in his or her lifetime. For patients over 65 years of age, the average increases to 28.4 individual doctors, including primary care, specialists, hospital and urgent care providers. Among all of these doctors and all of these systems, there is little-to-no communication. This leaves a patient’s medical record scattered across many different types of systems controlled by many different IT departments and behind many different firewalls.
Data Ownership Challenge
Medical records have traditionally been compiled and maintained by health care providers. A recent study showed that 62% of insured adults rely on their doctors to manage their health records, and nearly 29% of the respondents indicated that they keep them in a home-based physical storage location like a folder or even a shoebox. For personalized medicine to succeed, individuals need the ability to access their health information electronically, and actively direct its flow so as to take charge of their own health and make more informed decisions. The requirement for individuals to securely and electronically authorize the movement of their health data to destinations they choose cannot be understated for personalized medicine to succeed.
Data Interoperability Challenge
An additional challenge to building a longitudinal health record for a patient is matching unique records across multiple sites such as physician offices, hospitals, ambulatory care centers and home health. So, despite the general availability of datasets, interoperability is still lacking. Whereas much of the data is accessible in structured formats, there is also a lot of data that is only available in unstructured format, such as images or text. The heterogeneity and diversity of data thus limits their accessibility and re-usability.
The individual is the Solution
These challenges are not insurmountable, but relying on traditional regulations, institutions and incumbents has rarely produced innovation at any kind of reasonable speed. The key to rapid innovation in personalized medicine may actually lie with the individual. The reason is simple. It’s the individual who is most motivated and most incentivized to take advantage of personalized medicine. Health information, traditionally in the sole possession of healthcare providers, increasingly is also in the possession of individuals; and advances in mobile health (mHealth) applications, interoperability standards and online data storage have led to the development of personal health records (PHRs) that are maintained by patients themselves.
Recipe for Individuals to create Longitudinal Medical Record
The most logical starting point for any longitudinal record is the capture of lab and imaging orders and results, including one’s genome sequencing. Because 70 percent of clinical decisions involve lab and imaging data, it is critical to ensure the inclusion of this information to provide a good beginning to a comprehensive longitudinal record.
The next step is to incorporate all other clinical information from EHRs, including notes, symptoms, diagnoses, allergies and medications. Inclusion of this information provides the 360-degree patient view needed by providers throughout the care continuum.
The final component of the longitudinal record is financial or claim information. Claims are an important source of information to understand what services have been provided to the patient. This includes data beyond what would be in the EHR, such as outpatient therapy, out of town emergency room visits, etc.
User Control and Ownership
The promise of personalized medicine will incentivize individuals to have their genomes sequenced, and their health history, lifestyle habits, and environmental exposures aggregated, collected and organized on their electronic device or personal cloud storage. Individuals manage information from their own devices such as smartphones or tablets, and share that information seamlessly across multiple electronic platforms as appropriate. With mHealth applications available for iPhone and Android, users can effectively add cryptographically signed permission slips giving other entities rights to access their stored data in particular ways. Healthcare providers, Health Information Exchanges (HIE), insurance companies, and researchers can query health data, but the power to give permission to third parties to run queries and computations, and the power to revoke that at will lie with the individual.
Interoperability Standards
Today more than ever, the availability of FHIR (Fast Healthcare Interoperable Resources) and Substitutable Medical Apps, Reusable Technologies (SMART) offers health apps a unique opportunity to innovate in the healthcare space without encountering interoperability challenges, outdated APIs or lack of standards. FHIR offers a cheaper, easier and more modern platform that enables healthcare specific solutions to easily interact with patient healthcare data regardless of EHR vendors. FHIR harmonizes all patient data into standard code sets that can be easily accessed by all providers. SMART is an open, standards based technology platform that enables seamless and secure connectivity ensuring access to patient information across the continuum of care providers. So today, if the individual is motivated, there are tools available to aggregate data from various sources including one’s genomic data, which in turn can be accessed in a standardized and interoperable manner.
Data Analytics
Once the data is aggregated and consolidated, even though most of it might be under the complete ownership and control of the individual, advances in IT and analytics can do the rest to aid in personalized medicine. Using big data analytics, drug companies and care teams can query, interpret and analyze the data to come up with new targeted drugs and care treatment. On an ongoing basis, they can track the effectiveness of medicine based on genetic markers and identify certain biomarkers that signify that people might be at risk for a given disease, and dispense personalized medicine appropriately.
The Side Effects
As drug companies create more accurate drug trials, the cost and time of developing drugs will have to be controlled and managed. Tools to efficiently create patient pools with the right genetic makeup will have to be developed. Changing the way we design and administer treatment trials, using big data to bring “personalized” or “precision” medicine to drug trials and research, can potentially reduce costs, allow the right drugs to be prescribed faster, and improve outcomes at lower costs. Pharmaceuticals could even go back to a lot of dead end drugs that seemed to work on one set of patients and not another to see why that was so. We could recoup a lot of money, and more importantly, while saving a lot of lives.
