Expanding Cloud Computing usage at Health Systems
By Anil Saldanha, Chief Cloud Officer, Rush University System for Health
Health systems comprise a group of hospitals providing health care services to people. Information Technology (IT) plays a critical role in care delivery since it is typically responsible for the deployment, maintenance, and enhancement of the EHR system.
Cloud computing usage at health systems is on the rise. We can review the usage using cloud service models – SaaS (Software as a Service), PaaS (Platform as a Service), and IaaS (Infrastructure as a Service). In addition, using the cloud paradigms namely Private, Public and Hybrid Clouds.
The public cloud adoption for EHR systems at health systems has been slow but does show promise for the future. The EHR vendors have made progress in releasing software versions in the cloud that may get deployed at major health systems in the future. While clinical systems at health systems have been slow to embrace the scale, promise, and efficiencies of the multi-tenant public cloud, there is a growing desire to use the public cloud in the non-clinical areas of the health system, such as research, innovation, philanthropy, and education. Since the non-clinical areas of a health system seldom require EHR integration or real-time patient information, health systems have great opportunities to experiment and build expertise using the latest offerings from the public cloud vendors. This follows the general trend in the IT world to embrace cloud computing for new projects or initiatives. It is quicker to buy a subscription to a cloud software SaaS product or deploy/build a cloud infrastructure for bespoke purposes in a non-operational health setup compared to the rigorous procurement process involved in the operational side of a health system.
Additionally, talent can be augmented using professional services teams (from cloud vendors) and consulting organizations at health systems in the non-clinical areas. Even though the available tech talent experienced in cloud computing is expanding, there is significant competition in the fast-growing tech industry to hire talent. To maintain competitiveness, health systems are embracing the rapidly changing IT landscape (including cloud computing), making the working environment attractive to competent IT talent and retaining talent long term. To summarize, the non-clinical projects are good candidates to embrace the cloud service models (SaaS, PaaS, and IaaS) depending on the availability of ready-made software solutions in the market (SaaS) and technical talent expertise in-house (PaaS/IaaS).
The increasing focus on HL7® FHIR® interoperability standard at health systems accelerated by new federal mandates such as the 21st Century Cures Act and the proliferation of mobile computing and telehealth needs, has provided an opportunity for the health systems to explore API Management for vendor integration, AI/ML tools for analytics as well as internet-scale cloud-native services from the public cloud ecosystems. The pandemic introduced a critical need for health systems to integrate with city, county, state, and federal health systems for Covid testing, vaccination, and reporting needs. Cloud computing usage has grown during the public health crisis to enable API integration and secure file transfers between organizations. Cloud computing service and deployment models are a natural choice for the health systems for any external integration with public or private organizations. Support for the FHIR® Bulk Export in EHRs such as Epic® will accelerate the adoption of cloud-based systems and integrations.
Private Clouds do exist at many health systems. IT systems critical to patient clinical delivery are predominantly hosted in data centers or private clouds. As mentioned before, there is minimal usage of public cloud environments for clinical systems involving patients. This is in line with patient and clinical data’s regulatory and privacy requirements. Hybrid models with a combination of private and public clouds may grow in the future. Unfortunately, widespread hybrid model usage will have slow progress.
