Interrupting Healthcare’s Imminent Meltdown and Why One Digital Ontology is the Only Option
Received Date: Dec 12, 2017 / Accepted Date: Dec 18, 2017 / Published Date: Dec 20, 2017
Healthcare in the United States is imploding and it’s only a matter of time before we lose our tenuous grip on high-quality care. It looks ominous enough from the outside with yearly costs ballooning 50% in the last 10 years to nearly $3.8 trillion in 2016 and insurance premiums steadily rising, but it’s not until you peek inside that you see how dangerous the situation has become. Routine over-diagnosis, frequent over-treatment, expropriated incentives, unnecessary busywork and provider burnout with little money, less time and no clear path to a better system makes a dangerous combination. When spending tightens, and you can bet it will, what happens to quality? Healthcare today in this country or any other can’t perform at half the price.
So, what’s keeping an industry steeped in technology from exploiting the blazing computer speed and widespread connectivity that’s transforming other industries? Amazon wants to know. Google, Apple, and IBM want to know. Certainly, the U.S. government wants to know. While there’s a lot of money to be made by the company that simplifies healthcare, the amount of money we’d save by transforming our broken system is staggering. Recouping revenues from the past decade’s failed policies alone amounts to $1.4 trillion per year and we were far from cost-effective in 2008. Before I reveal the answer everyone’s looking for, I need to introduce a term. An essential term that you may not have heard before but one that will change how you view Healthcare’s problems, and solutions, from now on. This term is Ontology.
Keywords: Healthcare; Routine over-diagnosis; Healthcare’s problems
Borrowed from metaphysics, Ontology is the formal naming of everything in a sphere of knowledge (Figure 1) including their properties and inter-relationships. More than a glossary, inventory or dictionary, Ontology describes what exists and how everything fits together. For example, if we use your home closet as the sphere of knowledge, its Ontology would be an exhaustive description of your shoes, slacks and tops, their colours, shapes and materials, including every property of each material, and how one or more items go together by style, season and function. No one bothers with a comprehensive description of their closets, but if you did and provided enough detail, your Ontology would not only describe your closet it would describe many items you’ve never seen in thousands of other closets and stores. Taking it one step further, if you expanded your Ontology to include every color, every shape, every material, every property, every outfit for every style, season and function, and added every size, your single Ontology could describe the entire clothing industry. So, what does Ontology have to do with the cost or quality of healthcare? In a word, everything .
In computer science, the Ontology dictates how data is organized and labelled. Also referred to as the data structure, every database has one. Unfortunately, healthcare has hundreds if not thousands of databases built over the past 50 years by universities, hospitals, insurance companies, regulatory bodies and vendors to store and manage data specific to their particular business, and they’re all locked tight. But even if they were wide open, one thing will always prevent the fusion of these electronic file cabinets into one meaningful dataset-their different Ontologies.
Computers aren’t smart. Computers process and software tells them what and how to process. Thirty years ago, when someone searched a health database to find disorders with the symptom fatigue, the search term had to match precisely. The software would ignore the words Fatigue, fattigue and fateeg. Fortunately, with today’s spellcheckers software always finds the right word but back then what would it find? Would it find every disorder? Would it separate total body fatigue from leg fatigue from hand fatigue from eye fatigue? Would it include disorders for tiredness, weariness, sleepiness, drowsiness, exhaustion, lethargy or lack of energy? Would it distinguish between these similar symptoms and use prompts, information from a prior search or this someone’s purpose or profile to shorten the list? Of course not. It would return a long list of records with the word fatigue in the title. You’d be surprised at how little has changed in the last 30 years (Figure 2).
Welcome to Health Care’s data management dilemma. Isolated databases contaminated with meaningless data including misspelled or missing words, which ignore vast segments of healthcare, that can’t be combined and software that will never be better than the data it can find, regardless of how fast a computer finds it. And it’s all because of a legacy data management strategy using flat paper-based Ontologies. Now, I don’t mean to be overly critical. These Ontologies were adequate and in some cases excellent for a long time but what tech has a shelf life of 30, 40, or 50 years? They were a product of their time. A time when computers were slow and used primarily as electronic file cabinets, a time when no one bothered to include definitions, and everyone treated simple and complex terms the same, a time when it was unthinkable to describe the entire healthcare industry including every person, place and thing. Today is no longer that time. We desperately need an upgrade. We need to make health data easier to manage.
Why It matters
Before I detail the Ontology we could be using, let’s step into the future for a moment and imagine it’s already the industry standard. Silos of isolated, contaminated information have been replaced by shallow trays of clean granular health data. Security has evolved from medieval walls to personal armour protecting identities better than ever before and software developers are busy working to integrate health data into the lives of everyone who needs it precisely when they need it. But is it enough? Can Ontology interrupt the imminent meltdown we’re facing today? I’ll let you be the judge but here are a few examples of what health software based on Ontology could do for healthcare.
Better health software must start with the patient, not merely because it’s our data and our lives but because the health of our bodies and minds is the sun around which healthcare orbits. It’s also because most of us can’t and don’t want to be doctors so letting us peek at our medical files or putting the file cabinet on our phone isn’t the answer. Without knowing anything about health or care, our data could be working for us with software that generates the best short list of foods, activities and relaxation to maintain our health and energy on any given day and seamlessly presents viable options to our shopping lists, calendars and activity planners .
When we’re not feeling well, it offers short lists of possible disorders, remedies and additional diagnostics we can do on our own, best people or facilities to engage if we’re not improving, and when we should seek immediate attention.
We don’t need another isolated health app or portal. We need a data platform that allows software to become our concierge, automatically pulling health data pertinent to us, assessing its value to us and gently guiding us through health and care. Who knows, when we’re honest with our private concierge about what we’re doing and how we feel, we might pay more attention to what it says.
But it doesn’t stop there. When new software for workers in the healthcare industry comes on-line, the benefits multiply - for them and us. Envision a work environment for providers where every computing device became their professional concierge when they logged on. Software pulling health data pertinent to their specialty from anywhere including saved organic resources and their patient’s health profiles refined by their schedules and displayed just how they like it for the task at hand. Not only could this time-saving software translate into higher quality face-time with our providers when we’re not feeling well, want explanations or need reassurance, focusing the full power of the entire healthcare dataset on issues specific to us makes every provider better. When we’re not face-to-face, software sharing the same data platform could also keep our providers apprised of how we’re doing and one-click-ready to help when we need them. Providers who are more connected with patients, better equipped to help each one with more time to do so would go a long way towards slowing the slide.
The dilemma: Valuable health data is locked away in hundreds of silos with no way to tie it all together or deliver it efficiently. This file cabinet paradigm is preventing Healthcare from exploiting todays computing power to raise quality, reduce cost and remedy access to care.
The approach: Use one Digital Ontology to describe all health data making it easier for every stakeholder to manage what they need. Like a giant CAD engine, this Ontology can model anything in Healthcare and sharing source materials means they’re all inherently connected.
The path forward: Inspire people capable of orchestrating the construction and adoption of one Digital Ontology. When we make health data easier to manage all software gets smarter and the benefits of true Digital Health can start reversing Healthcare’s expensive and dangerous slide .
While these examples of better software highlight the business end of healthcare, its vast support network has a significant influence on outcomes and a much more substantial role in how much it costs. Therefore, to find the most cost-effective solutions and improve overall quality the data platform has to connect everyone. Descriptive terminology for academic disciplines, training programs and research must share common ground with the terminology used in clinical medicine. The same is true for pharmaceuticals, equipment, food and supplements as well as data used by service and professional organizations, regulatory agencies and payers. Connecting apparently unconnectable entities in an industry is the role of Ontology, but it’s important to recognize that like the closet analogy when your Ontology described its contents in exhaustive detail, it didn’t necessarily reveal who owned the closet or where it was.
You might be asking yourself, how could anonymous, clean, contextually-rich health data about every man, woman and child in the country contain costs or improve quality? Well, health systems might want to know what products and services patients are using online and in person, so they could anticipate future needs, allocate resources and stay in business, especially when new consumer-facing software goes viral. Vendors of these products and services from big Pharma to therapists might like to know if and how well their offerings work on different patient subsets under real-world circumstances and would probably enjoy saving money on research, development and marketing. Payers may like upgraded billing codes that accurately reflect health, disorders and best treatments, decreased risk, absent fraud and a healthier population to manage. Policy-makers from Congress to the Department of HHS and the FDA should like all of this and finally be able to align best practices with best policies for each of us. After all the rhetoric, time and money spent in the last decade trying to improve access to healthcare, wouldn’t it be ironic if universal access was as easy as giving everyone a new phone? When one Ontology becomes the industry standard, it just might be that simple.
This Ontology Is Different
Unlike legacy databases, this Ontology was designed to describe the entire Healthcare industry and every different thing, person and place in it, or more accurately the data that represents these entities. Like most industries, some of these things are simple and others quite complicated so the Ontology had to be detailed enough for the simplest thing and provide a method to combine these details to describe the most complex entities. Healthcare is also one of the most dynamic industries, so it had to evolve naturally without disrupting its core functionality. And finally, the Ontology could not constrain the data it described. I repeat, the labelling mechanism can’t restrict the free flow of health data, especially the insensitive data.
So how much health data is insensitive? Nearly all of it. The results of an imaging study, the name of a disorder, the medicines we take or how we felt this morning are all insensitive, as long as sensitive data like our name and location are invisible. Contrary to the all-or-none architecture of current Ontologies that ink identifiers on everything, this Ontology focuses on the abundant and infinitely more valuable insensitive health information and then carefully manages identities (Figure 3).
Like super-selective magnets, detailed health information pulsing in our personal profiles and primed to pull matching data could bring our concierge to life, except for one thing. Ontologies in use today also lack the depth and precision to describe one person accurately let alone the whole industry. But what if we  assembled everything we know about health and care into one multi-level Ontology and designed it to generate digital labels instead of hold data? Could these labels provide enough detail to adequately describe each one of us and everything else in healthcare? Absolutely.
The levels and the details
Similar to your closet Ontology which probably started with the shape and surfaces of the body, the foundation for this multi-level Ontology is the entire human body. From skin to spine, all the regions, systems and structures that constitute the body’s physical form for each sex at every age. Something we already know a lot about and won’t change very much. This starting point is already a significant departure from (Figure 3). Biological systems don’t store products described by DNA in the DNA so why is our health data stuck in the structure that describes it especially since every silo describes it differently and when it’s moved all the descriptions disappear? To build the health biosphere we need we must first replace our old data management model with one that provides inherent interconnectivity, exploits today’s computer power and accommodates tomorrow’s discoveries. Existing Ontologies which frequently build on complicated entities like diagnoses or outcomes but it’s about to get radically different. The name for each structure is included only once on this Form level and there is five other levels.
Our body is a machine, not a mannequin, so the second level is dedicated to Function, or the elements necessary for activity like genomics, physiology, microbiology, biochemistry, fuel and the substructures where these elements operate. Again, each name or term is included only once. The reason for using separate levels of unique terms is that a term on one level can be associated with many terms on another level. For example, a specific biochemical reaction or microbe on the Function level is associated only with the structures on the Form level where it operates or occurs. Non-redundant, perfectly granular and when we discover new relationships, infinitely upgradeable or better yet, organic. These coupled terms also become part of the de facto definition of that particular term (Figure 4).
While powerful, associations between pertinent terms can’t always tell the whole story so there’s a parallel relational catalog on each level using these unique terms to describe more complex processes like pathways, mechanisms and procedures. Every level in the Digital Ontology above Form is composed of these two independent but connected catalogs. The next level is where things start to get interesting.
The Examination level is devoted to the different methods we use to measure form and function - not just see but measure. This is an important distinction because even though many of these methods involve a visual component, the Examination level is where qualitative data is quantified.
Accurately describing a thing means measuring all of its properties
On a blood test what is the concentration of a substance? On a brain MRI how much does a structure enhance after contrast? On an EKG how regular is the heart’s rhythm? On physical exam how red and raised is the rash or during an interview how agitated is the patient and how severe is their pain? Each of these methods to examine the body, from samples to sophisticated machines to verbal cues, is just a glimpse of how well it’s working so we must make the most of them. We must also be ready for new ones when they’re discovered. You may have noticed that the Ontology has yet to address a single disorder or even what constitutes normal. All of this happens on the next level.
On the Condition level, the Ontology finally applies names to collections of lower-level terms or more accurately it associates findings from our measuring techniques with specific disorders. Diabetes isn’t just high blood sugar, it’s everything from the causes to the effects including where they occur and how we measure them. This level is also where environmental factors, exposures and family history are integrated with this operational information to create a contextuallyrich representation of health for each one of us - in other words our normal range. And because constitutional findings like fatigue, pain and malaise are tied into all this descriptive detail, when we’re not feeling well the best short list of what’s out of balance isn’t far away. Now that the Ontology has exhaustively described human health and disease it finally has enough detail to address the Health Care familiar to most people. This process starts on the fifth level with Interventions .
The Intervention level organizes all of the products and services from around the globe rumored or proven to keep us healthy and return us to health when things go wrong. Past efficacy is not a prerequisite for inclusion here but just like other levels, every property of every intervention is fully exposed. The contextual framework on this level also includes purposes like prevention, surveillance and treatment, training if any, compliance measures, time and cost (Figure 5).
Segregating interventional options from the human elements in lower levels not only creates a more accurate description of what the intervention does and how well it works, but the inter-relationships between terms on every level could also uncover hidden benefits, predict specific side effects and illuminate clear paths back to health. As we’re being given more responsibility for managing our health, having a complete inventory of what’s available will only become more critical in the years to come, especially when our health concierge starts composing our personal menus.
Sandwiched between the five lower levels and the data to be labelled, the Application level assumes several roles. As the Ontology’s top level, it’s responsible for describing the most complex entities in healthcare. As the surface level, it has to sync the Ontology with realworld terminology and applications. It also must manage the sensitive information. To satisfy the first two requirements, the Ontology divides the relational catalogue on the Application level into Sets and Profiles, where a Set is a list of terms sharing a common theme and a Profile is the digital representation of a complex entity using terms to describe every attribute. It then provides a method to add any term and its definition to any Profile. A combination resume-builder and digital dictionary except for one thing, there’s no contact information on the resume.
As unique terms, descriptions for sensitive information like the subject’s name, ID numbers and location are managed on the associative side of the Application level, but they’re protected differently and are hidden by default. While some restrictions apply, users including patients, providers, professors, researchers, vendors and organizations can adjust their privacy settings and expose one or more identifiers to specific people, under certain circumstances or in real-time like when our phones ask if we’ll allow an app to access our location. Everything in Health Care should have a Profile and if a person does business in Healthcare, they should have two - one Patient and one Professional. There’s no reason a health professional should be completely anonymous .
Profiles for patients are a no-brainer. Profiles for Providers are too. But what about Profiles for disorders, drugs and departments? Why could these improve quality and reduce cost? Simply put, Sets and 3D Definitions (Figure 6).
A Set or list of terms sharing a common theme is critical to syncing the Ontology with healthcare as it is now. Our whole health system was built on broad terms like specialty, work, therapy, cancer, infection, device, department, Anatomy, Physiology and hundreds more. Our educational, training and research institutions use them every day as do professionals who manage licenses, quality assurance, billing codes and compliance. Of course, we still want to and need to use broad terms but when every term in the list shares thousands of data points with one another via their 3D Definitions, they become interconnected and far more meaningful. In our quest to prevent dementia, cure cancer and prolong optimal health we have to connect everything in Healthcare but for some discoveries, how two entities differ may be even more important.
It’s worth noting that because Sets and Profiles reside on the Application level, both can be customized by software developers. New Sets can be added to address issues specific to organizations, states or countries, and the formless Profiles can be molded to satisfy the functionality and aesthetic of the end user. One more thing, the entire Digital Ontology is numerically coded to save space and improve performance (Figure 2).
The definition of any term in the Ontology, in this example Term Y in a Set on the Application level, is the string of pertinent terms in both the primary (gray) and secondary (black) catalogs including the relationships between these terms. This is the 1st order definition (a). But each term in the 1st order definition also has it’s own definition, in this example Term 95 (b). The 3D Definition of Term Y is therefore it’s 1st order definition plus all it’s 2nd order definitions (a+c). Every term in the Digital Ontology has a 3D Definition and since all terms reside in the Ontology, Profiles using them are not only the deepest, most consistent description of an entity and it’s attributes, they’re tailormade for precise data management, whether matching a few terms or thousands .
Over the past few years, there’s been a lot of chatter about Digital Health transforming healthcare with Wellness, Patient-centric, Exponential and Functional medicine, Artificial Intelligence, Big Data and Big Tech all trumpeting a better future. Billions of dollars and 250,000 health apps later why is healthcare in worse shape today than it was ten years ago? The simple answer is they don’t work. The harsh answer, the one that has everyone scratching their heads and apologizing to their investors, is they can’t work and never will work until we solve our data management problem. Fortunately, once we do, these remedies will have a legitimate chance to work and more likely than not work together.
By now, you should appreciate the role Ontology plays in labelling data and how adopting one Digital Ontology for the Healthcare industry could solve its data management problem and set the stage for health software that works for each one of us. What remains to be seen is how quickly industry leaders will use this tool to interrupt healthcare’s slide or who might rise to the challenge since it’s no longer a mystery. A few scenarios we might see in the next few years deserve mention here [8,9].
As the biggest payer on the planet, the Federal government has both the incentives and means to mandate digital data management for the Healthcare industry. They already did it with electronic health records and for some time, regulatory agencies have been exploring methods to upgrade their businesses for a digital future. The FDA is already working to streamline processes for proven companies so it’s not a stretch to think they wouldn’t do the same for companies using one Digital Ontology, especially since it could help them plug into their industry.
While the government must be involved, it’s possible that as Digital Health policies take shape in Washington a company selling or delivering things in Healthcare incorporates the Digital Ontology into their data platform. The company may be large and already established or just a start up but when they roll out their digital products or services everything clicks, and they go viral. Within months people fall in love with them because they’re simple and they work. It’s happened before. Or maybe a billionaire who’s already pledged their fortune to charity recognizes that one Digital Ontology would have an even greater impact on global healthcare and builds a business to manage and disseminate the Ontology. Becoming the steward of accurate and meaningful health data for humankind just might appeal to someone.
Regardless of how it happens, the key is that it does happen. Make no mistake, keeping our old data management model on life-support is the major reason Healthcare is so expensive. It’s also why Digital Health is stalled at the starting line. It’s not like we have to wait for a supercomputer to show us how to connect everything and take control of our health data. Whether it’s this Digital Ontology or something similar, the information and expertise exist right now. So, I ask you, what are we waiting for.
Citation: Sewall B (2017) Interrupting Healthcare’s Imminent Meltdown and Why One Digital Ontology is the Only Option. J Health Med Informat 8: 296. Doi: 10.4172/2157-7420.1000296
Copyright: © 2017 Sewall B. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Select your language of interest to view the total content in your interested language
Share This Article
4th World Congress On Protein and Bio-Medical Engineering
March 15-16,2018 Singapore City, Singapore
10th International Conference on Proteomics and Bioinformatics
May 16-17, 2018 Singapore City, Singapore
13th International Conference on Metabolomics and Systems Biology
June 11-12, 2018 London, UK
International Conference on Computational Biology and Bioinformatics
Sep 05-06 2018 Tokyo, Japan
- Total views: 279
- [From(publication date): 0-2017 - Feb 23, 2018]
- Breakdown by view type
- HTML page views: 252
- PDF downloads: 27