Company: Mediclinic International plc
Company background: We are a diversified international private healthcare services group, established in South Africa in 1983, with divisions in Switzerland, Southern Africa (South Africa and Namibia) and the Middle East. We place science at the heart of our care process by striving to provide evidence-based care of the highest standard. The group operates over 80 facilities with over 11 000 beds and employees over 37 000 people. Mediclinic is focused on providing specialist-orientated, multi-disciplinary services across the continuum of care in such a way that the Group will be regarded as the most respected and trusted provider of healthcare services by patients, medical practitioners, funders and regulators of healthcare in each of its markets. Mediclinic takes a sustainable, long-term approach to business, putting patients at the heart of its operations and consistently delivering high-quality healthcare services. In order to deliver on these priorities, the Group upholds the highest standards of clinical governance and ethical behaviour across its divisions; invests significant time and resources in recruiting and retaining skilled employees; makes considerable investment into its facilities and equipment; and respects the communities and environment in the areas in which it operates. The Data Science and Information Management department (based in Stellenbosch, South Africa) is responsible for delivering data and analytics solutions across all of the company's operating divisions. We have been SAS customers for almost 10 years
Contact: Shravan Lalbahadur
Title: Group Manager: Data Science and Visualization
Country: South Africa
Award Category: Innovative Problem Solver
Tell us about the business problem you were trying to solve?
- Our centralized architecture brought all the data to South Africa and for analysis and reporting. This approach was slowing us down as the team was struggling to produce reports on SAS 9.4 in a timely manner because data was only being refreshed 1-2 times per month.
- New data privacy laws in Switzerland and the UAE were promulgated that prohibited the transfer of healthcare data outside their borders
- We want to deliver predictive and prescriptive analytics to doctors and nurses but this would be difficult to implement with our existing SAS 9.4 deployment and architecture
How did you use SAS to solve that business problem? What products did you use and how did you use them?
We expanded and restructured out team to include analytics presence in the UAE and Switzerland. We moved our data warehouse to the cloud so that it would be able to support the new analytics environment. We then decided to deploy Viya 4 on our Virtual Private Cloud in the UAE and Switzerland as it would be the most cost effective option (TCO over 5 years). This option also required the fewest amount of work hours needed to modify existing code/programs.
We migrated our previous jobs from SAS Enterprise Guide on SAS 9.4 to SAS Viya Enterprise which was deployed on our own Virtual Private Cloud in Azure. SAS Studio was used to migrate the old process flows from SAS Enterprise Guide. The deployment was accelerated by the fact that Mediclinic is using several Microsoft/Azure products which are supported by SAS.
What were the results or outcomes?
We immediately saw 40% to 60% reduction workload times after migrating to SAS Viya on Azure. Workloads that took 24-30 hours to run now complete in 10-13 hours. The flexible compute configurations & scalability available as part of Kubernetes and Viya will allow us to save 30%-50% on compute costs on an annual basis.
Our data scientists are able to move seamlessly between Python and SAS as a result of the Viya's native support of Python and Azure Databricks. They are also able to deploy their open source ML models on SAS.
Why is this approach innovative?
We are able to match compute demand and supply in a cost-effective manner according to the needs of the team, organization and our customers. For example, we are using Azure Logic Apps to define rules for scaling our Viya environment on Kubernetes.
We were able to leverage on the partnership between Microsoft and SAS , and the existing resources for SAS on Azure to deploy SAS in 6 months end-to-end (infrastructure+SAS deployment time).
