AI-Powered Healthcare: Balancing Expectations with Data and Privacy
By Tanya Thouw, Senior Global Operations Manager Healthcare, SAP
Imagine a doctor’s appointment without a wait time of 40+ minutes, redundant paperwork, and repetitive questions. Instead, the doctor hits a button and pulls up your entire medical history, including hereditary risks and findings from the latest relevant scientific studies.
With a data-informed evaluation specifically tailored to you, the doctor utilizes the system to make recommendations for a treatment plan, including antibiotics, holistic remedies and even lifestyle changes. In the end, thanks to artificial intelligence (AI), the time and money you spend on this visit are significantly lower than what you’ve spent in the past.
"The most realistic approach to sorting our siloed health data is to develop healthcare platforms that can gather and organize data and overcome different data formats"
Does this sound like a dream? It’s slowing becoming reality. AI holds promise for this kind of personalized medicine, but we have a long way to go. Today, for every hour physicians see patients they spend two additional hours on paperwork.
Today’s Siloed Data, Tomorrow’s AI Future
Technological advances taken by progressive companies will change the world of healthcare, but progress is persistently impeded by disorganized health data. Think about all required annual visits, shots and test results you’ve accrued over your lifetime. All that data is collected and stored in a range of formats, making it difficult for doctors to have all relevant data at all times.
The most realistic approach to sorting our siloed health data is to develop healthcare platforms that can gather and organize data and overcome different data formats. This may sound simple, but the data in question is very sensitive and personal information that must be treated with the utmost care.
Health Data and Patient Privacy
Regulations like the General Data Protection Regulation (GDPR) and the “right to be forgotten”are designed to protect the patient’s privacy, but these regulations also present significant hurdles when it comes to advancing medicine. Under the GDPR, patients are entitled to have the ability to access their own health data at any point, and to decide where and how their personal data is used. This means that any patient data can only be utilized for research under certain conditions, like having the patient’s consent, be completely de-identified, kept for only a limited time and also be subject to the right of erasure. This presents a number of complications for the medical research process. For example, completely de-identifying information for research will not result in useful takeaways and being unable to hold data for extended periods of time significantly impacts the efficacy of longitudinal studies.
The goal is to respect patient privacy while simultaneously advancing medicine to give these same patients better care. To make this work, we must find a compromise, or a “good enough” satisfying level standard that guards trust between the patient and the healthcare system but also supports important research initiatives. Utilizing Fast Healthcare Interoperability Resources (FHIR) standards will help in overcoming data silos and exchanging clinical and administrative data for the better patient care.
The definition of this “good-enough” standard will support medical professionals, technologists and governments so that measures created to protect patients don’t end up hurting them instead by hindering the development of potentially life-saving medical innovation. With AI becoming smarter and more valuable by the day, the time to start this conversation is now.