2. Pharmacy OneSource by the Numbers Healthcare's #1 Software-as-a-Service (SaaS) provider More than 44,000 pharmacy professionals are active members of the Pharmacy OneSource community. Nine quality healthcare application offerings More than 1,300 organizations in the U.S. utilize one or more of our applications
3. Healthcare’s Big Goals Efficiency: Human Resource Scarcity Declining Reimbursement Increasing demand and severity Effectiveness: Mortality and costs are too high Reimbursement penalties and rewards Growing readmission rates Preservation: Worsening anti-infectives efficacy
4. Pharmacy OneSource is Adapting to Market We are focused on delivering real-time data-driven analytics optimizing care in quality, cost and compliance Every day we process over 4 million messages supporting over 200,000 active patients We run 9000 rules against these messages in real-time We needed to find a more cost-effective way to support our market need
5. By Ensuring We Adopt Market-leading Solutions, Rather than Building Drive Innovation Not bleeding edge, but market-driven and competitive Technology innovation driven by ease of management and ubiquity Balance Cost, Quality and Speed Increase operational efficiencies in the delivery and hosting of our services and data sources through-out our product offerings Leverage Agile development methodology to increase quality and consistency in delivery of our product offering
6. Focusing on Ambitious Goals Aligned with the Business By being cost-effective: Our justification was to leverage existing investments in infrastructure by doubling capacity Drive down our cost per hospital to better manage our margins By meeting the scalability challenge to support real-time: Support an architecture that could scale to 10 million messages and beyond on a daily basis Enable our rules to process the messages and generate results back to the client in less than 10-secs, regardless of complexity Provide a bridge to the cloud to enable us to move away from the hardware business
7. Resulting in our Choice for a Spaces-Based Architectural Solution In-Memory Data Grids are inherently faster than databases because memory is faster than disk. A recent Stanford University study confirms that In-Memory Data Grids "can provide durable and available storage with 100-1000x the throughput of disk-based systems." An In-Memory Data Grid relieves the database bottleneck, which makes it difficult to scale most enterprise applications. It does this by scaling the data layer and sharing the load among numerous machines. Databases and file systems have inherently limited latency because they run on magnetic disks. According to a recent Stanford University study, an In-Memory Data Grid has "the potential for extraordinarily low latency ... 5-10 ms." For an in-depth overview of the In-Memory Data Grid concept and its benefits, see the Stanford article cited above: The Case for RAMClouds: Scalable High-Performance Storage Entirely in DRAM (Ousterhout et. al, 2009)
8. This will Address our Need for Speed and Power Today we are challenged by physics (or more precisely disk speed) and Scalability Disk speed is a massive bottleneck and leads to performance issues that result in: Messaging traffic backups List processing delay and timeouts Resource contention As we move away from traditional web-based architectures we need to Scale effectively Scaling out vs. scaling up Linear scalability Cost-effective scaling
9. Secure Early Wins, While Minimizing Risks in our Actual Approach Selection Strategies Evaluated market-leaders to see which could best serve our needs: GigaSpaces and Terracotta RFP sent to finalist vendors Piloted finalist before making decision We split our project into four phases: Replace our HL7 ADT messaging Replace all HL7 messaging (formulary, lab/microbiology, pharmacy, surgery and radiology feeds) Transition our customers to new platform Replace our rules processing with Drools rules processing Implementation Approach We use Agile/XP – co-located teams, TDD, Paired-programming, etc We established a team dedicated to messaging and another team for rules-processing Tackled implementation as an evolutionary replacement defined around four phases shown above Separated out the development from actual customer conversion Demonstrated success with early wins
10. Our Journey Has it’s Challenges Integration - Code complete is only half-way there, we still need to transition our customers to the new platform Resourcing - Finding the right skills to undertake the development and transition knowledge Culture - Building a culture around innovation and market-focus Learning Curve - With any new technology expect to adjust timelines to support learning with team – ensure the mission is well understood Technical - Be prepared to give up old approaches for new capabilities. At the end of the day it is only software and prone to change
11. Our Initial Results Support Our Projections Our initial pilot with GigaSpaces proved out capability and demonstrated support for our own technical considerations We completed first phase with ADT HL7 feed replacement We saw 6x increase in message processing from 25 messages per second to over 130 messages per second We have converted over 150 customers and on track to complete this by year end This has reduced our message load on our legacy architecture by 15%
13. Our Investment has Tremendous Upside for the Future Freeing $20 million from the bottom line! Rapid time to market: leveraging market-leader to deliver solution in 18 months Deliver processed lists faster than 10-seconds Scale to process over 10 million HL7 messages daily Bridge to the Cloud – taking us out of the hardware business by scaling via On Demand cloud vendors
15. Pharmacy OneSource and GigaSpaces is the Delivery Platform for our Customers Overcome performance challenges by leveraging the In-Memory Data Grid Get to market rapidly by leveraging market-leading solutions getting us to market rapidly Scale without passing the expense burden onto our customers A bridge to the Cloud – we will no longer need under-utilized hardware, but provide On Demand solutions as we add customers We save people’s lives by delivering critical information in real-time!
Notas do Editor
Set by JCAHO, IHI, CMS, Insurers, State Boards…Pursued by health system strategies and profession based visionsExecuted by healthcare professionals: Pharmacists, Nurses, Physicians, Infection Preventionists, Case Managers, Quality Directors…The human and economic burden of healthcare-associated infections (HAIs) is significant, and growing. According to the Department of Health and Human Services, nearly 1.7 million HAIs occurred in 2002, and they are among the top-10 causes of death in the United States. HAIs also generate up to $33 billion in excess healthcare costs each year.
GigaSpaces XAP is an In-Memory Data Grid that delivers an in-memory cache for fast data access, and an advanced distributed cache for extreme performance and scalabilityGigaSpaces’s XAP eliminates database bottlenecks and guarantee consistency, transactional security, reliability and high availability for the data