With the COVID-19 pandemic rearing its ugly head in 2020, the health care industry was in dire need of adaptability to continue serving patients. Telehealth existed before the pandemic, but it saw a more significant purpose when medical offices began to shutter their doors and providers needed to seek innovative ways to reach their patients. Ideally, almost overnight, a new data source emerged, leading to further complexity in the existing world of health care data. While this new source has the potential to facilitate care, it’s important to understand the barriers of integrating it into an already fragmented data center effectively.
To paint just a brief overview of the situation, let's consider where my health care my data resides. Since birth, I have been cared for by 13 health systems. While I'm sure it's somewhere, I have no idea where the claims data exists for the surgeries I had at ages 16 (broken leg — skiing), 22 (broken leg — skiing), 29 (broken leg — falling down the stairs), and 32 (vein stripping as a result of the previous three injuries/surgeries). I had different insurance providers for each claim, and they were handled by different health systems in different states. I have had at least two colonoscopies in different systems. I have a family history of cancer, but the relative is married and has a different last name. I received my flu shot and COVID-19 vaccinations at a supermarket (I think they asked for my insurance card, but I’m not sure). Since moving to Washington, D.C., I have received a tetanus booster, but I guessed when my last booster was on the form. For fun, I count my steps on my phone to keep healthy, but there is no integration with my medical record. Plus, my charts are populated from the forms I fill out, which may or may not be accurate, depending on what I remembered at the time.
When new technologies are introduced into health systems, it's important to consider where health data is coming from if it is integrating into the other data to help facilitate care and improve outcomes. For clinical data, the electronic health record (EHR) is a good data source. But, as demonstrated in my example, there may be multiple EHRs to deal with, some of which are populated by what the patient remembers. Insurance claims could be a good source of data, and, under value-based care, will need to be integrated with clinical EHR data to measure the outcomes and cost of care. Pharmacy data would be a good source, given the frequency of prescription refills, but rarely interface with EHRs. Telehealth data will need to integrate into the EHR seamlessly, thus demonstrating that intervention's impact in telling a complete story. We must also consider direct-to-consumer tools, such as patient monitoring, wellness apps, step counters, weight devices, and their data capture and integration challenges.
For practitioners to practice evidence-based care, there must be a benchmark to measure deviance from medical evidence in the literature. This should constantly be bumping up against the clinical data, so variation is readily available at the point of care and over time. Patient surveys or perceptions need to be incorporated into the clinical data. And finally, consideration needs to be given to social media sources.
Characteristics of Health Care Data
The characteristics of health care data have been referred to as the five “Vs” — voluminous, velocity, variability, variety, and veracity.
Health care data is complex, heterogeneous, longitudinal, and voluminous. As demonstrated in my single case, a lot of the data story covering my 66 years needs to be integrated. Let’s then consider the U.S. (or the world) and the amount of data that needs to be supported and incorporated; it's enormous. It would seem no single system could handle the need.
The velocity refers to the rate at which data is created, stored, visualized, exchanged — analyzed amongst care providers. Because multiple providers may see a patient, this data needs to be readily available and bumped up against the evidence in the literature to aid in decision-making. An example could be the unavailable chest X-ray showing a pulmonary nodule that should be followed up on. The radiologist dictated the result, but the family practice physician missed the report, and the patient forgot to ask. One year later, the patient presented with shortness of breath and a chest X-ray demonstrating metastatic disease.
The variability of data refers to how it is presented, such as flat files, relationship tables, images, or texts. It also refers to using words or abbreviations meaning different things depending on context. Examples would be the word "nodules" referring to a breast mass or a lung mass. Or the abbreviation APC refers to activated protein C, advanced pancreatic cancer, or antibody producing cells.
The variety of data has been addressed in the examples so far. There is data everywhere, generated by information regarding the state of my body, the medications I'm on, what I'm feeling, etc. Therefore, the therapies instituted may not be based on a complete picture.
And finally, the veracity of data refers to a lack of the standard. If we could agree on how the data is named and presented, systems could share data, and a single repository could be the source of truth. A good example is the titles and codes of case reports, drugs, diseases, and examinations vary in different hospitals and health care settings. This lack of standards hinders the ability to share. There needs to be a standard lexicon of medical terms.
The Health Insurance Portability and Accountability Act (HIPPA) of 1996 is a law enacted by the U.S. government to protect sensitive patient health information (PHI) and monitor the use, storage, and distribution. HIPAA applies to “covered entities,” including health plans, health care providers, and health care clearinghouses. To maintain faith in the benefits of emerging technologies, there needs to be safeguards to ensure information is secure and protected. And while covered entities are learning how to make their experiences HIPAA-compliant, other data sources that tell the health care story need to be analyzed.
If I share data from my fitness app with my physician, can that be stored in a data center? What if, ultimately, it is captured immediately and sent to the data center — is the device sending the data HIPAA-compliant? HIPAA compliance is a must, but new technologies, like audio and video recordings, have created challenges for telehealth.
The clinical benefits of telehealth became crystal clear amid the pandemic. Patients in remote areas experiencing trouble with travel or high-risk patients who would rather not sit in a crowded waiting room can utilize telehealth to keep open lines of communication with their providers. Others found the convenience of telehealth too compelling and questioned any trip to manage their health. Emerging technologies, such as wearable devices also empowered consumers to manage their health. With the benefits evident, more patients and physicians will be willing to adapt to changes in technology and improved systems as innovations arise.
However, with this new technology, the onus is on the data center experts to ensure organizations have a powerful enough system to address the five "Vs" and the HIPAA concerns. Data center systems for covered entities need to support the smooth delivery of services and articulate the data limitation that still exists.