Artificial intelligence (AI)
As predicted, AI and machine learning were seen everywhere, but a big change was that the industry was more focused on the actual value in terms of advanced analytics this year, rather than on the AI technology itself. This shift was well received by the attendees and gave greater substance to AI as more practical examples of its use were presented.
A commonly addressed challenge when it comes to applying AI in healthcare is the need for powerful data management and processing power, capabilities that healthcare providers for practical and economic reasons do not currently possess. Cloud services are seen as one solution, where processing and data storage take place in a remote hosting provider’s environment, such as Microsoft Azure, Amazon or Google. However, there are still concerns regarding patient privacy and an inherent reluctance to relinquish control of the data, which is making the shift towards cloud services very slow. The cloud providers, on the other hand, spoke about AI as a trigger that will speed up the adoption of cloud services.
Creating value from data by using decision support tools
A red thread throughout the conference was to create actual value of the increasing amounts of data. Healthcare was mentioned as the biggest contributor to data growth among all industries, which makes advanced data analytics an even more important component when it comes to extract information that that can guide physicians’ decisions.
Decision support tools was frequently discussed, especially among the EMR vendors. However, a major challenge is to provide guidance in the right situations, and not to overwhelm physicians with information and warnings.
One example that was mentioned of a situation where decision support would be useful was when prescribing a type of medicine that the patient previously had shown an allergic reaction towards and alerting the physician about this.
Physician burnout is a major global issue and many different areas showed potential solutions to ease physicians’ workload, from better usability in health IT systems that create inefficiencies and frustration among users, to AI-based tools and voice assistants. Many blamed poor EMR systems for contributing to physician burnout, but a study that was presented showed no such correlation.
A common theme was applications that could help to simplify the daily lives of physicians. Some of the most interesting were prototypes that could automatically generate patient journal entries simultaneously as the physician met the patient. Other tools included digital doctors, such as Babylon Health and services to help chronic patients, such as Livongo.
As at earlier HIMSS conferences, interoperability remained a hot topic, although the focus had faded a bit due to the slow progression in the development of standards over the last 10 to 15 years.
HL7’s FHIR (Fast Healthcare Interoperability Resources) is largely seen as an enabler of health data exchange, which of course it is, but executives at IBM, Google and Microsoft said it will also lay the foundation for AI and machine learning. A substandard that was mentioned was FHIRcast, which enables instantaneous desktop synchronization between IT systems, which will minimize the need for time-consuming switching of applications.
In addition to FHIR and the importance of open APIs, the need for consensus among healthcare providers with respect to their internal language was brought up—“yellow at one hospital should mean the same as yellow at another hospital” was used as an example by the CEO of a large EMR company.
On this topic, HL7 and SNOMED announced an agreement on the formalization of a license agreement in which a relevant “free-for-use” set of SNOMED CT coded concepts will be used within the HL7 International Patient Summary (IPS). The main purpose is to facilitate the exchange, integration, sharing and retrieval of electronic health information.
The slow progress in interoperability have clearly prompted the US government to take action and one of the biggest news was a new law proposal that will make sure health providers provide health data to their patients digitally, but also to other care providers that require it for diagnostic or treatment purposes. This will hopefully accelerate the development.
The slow progression in interoperability was also addressed by some of the Nordic vendors. Sectra, Cambio and Tieto held an interoperability showcase, which received positive feedback, where they launched a project to enable equal care independent of location across the five Nordic countries. They showed how the project would enable patient data and medical images to be accessible from any hospital for one type-1 diabetic patient.
Overall, there was a common belief that an increased collaboration between vendors, government and healthcare is necessary and we will probably see a lot of initiatives to nurture such partnerships in the future.
Technology usable in a clinical context
Also discussed were the importance of developing technology that can create practical value in healthcare. This was a reaction to earlier frustrations over good technology that never succeeded in reaching clinical practice. The need for “useful” technology was mainly raised by healthcare professionals underlining the need for the industry to develop technology in a clinical context where it can be applied in a real environment.
On an adjacent topic, the organization KLAS Research had asked 150 healthcare providers around the world how satisfied their users were with their EMRs. Not surprisingly, the level of satisfaction varied considerably, from 10 to 20% up to 80 to 90%.
What was surprising was that it also varied between 20 and 80% among users with the same EMR. KLAS’s explanation was that only 20% of customer satisfaction could be derived from the availability of the system, reporting times and functionality, while the other 80% was due to HOW it was implemented, and how the account management and the dialogue with the users were conducted.
Hence, selecting an IT system that matches all the functionality requirements might not always result in the best solution. Other factors need to be taken into consideration for a successful implementation.
Another trend was “empowering patients to be informed customers”, which means giving patients access to their own data. The focus was on offering new, innovative solutions for patients to access their data and be involved in their own healthcare.
We will probably see a conflict in this area between the established big EMR vendors and new private equity-funded up-and-comers that challenge the EMR vendors—in a similar way as we already have seen with the digital healthcare providers that have disrupted the established healthcare system.
Wearables and patient-generated data
Wearables could be seen everywhere in the HIMSS19 exhibit hall. For example, one of the many vendors in attendance was VitalConnect, a vendor of wearable biosensor technology for wireless monitoring, which showcased a system that helps caregivers detect specific health outcomes by measuring weight scale, blood pressure and pulse oximetry, core temperature, single-lead EKG, heart rate, heart rate variability, respiratory rate, skin temperature, body posture, fall detection and activity.
The VitalPatch is the smallest and lightest FDA Cleared Class II medical device approved for use in both hospital and post-discharge home settings. A big challenge is for healthcare to make use of patient-generated data, something we haven’t seen yet.
As more healthcare providers are suffering from data breaches, malware and hackers, cybersecurity was a popular topic of discussion and even had its own area of the exhibition hall. A commonly held belief was that within three to five years will see the first major data breach in healthcare with catastrophic consequences. Increased use of AI, IoT and shared data will jointly raise the risk of data breaches and the common belief was that IT security needs greater attention.