Sometimes the highways need an upgrade.
If you’ve lived in a neighborhood that experienced sudden growth, you know that increased traffic is an inevitable inconvenience that comes with it. The same is true of the data environment. This doesn’t just pertain to server speed. Organizations can outgrow the way their code is structured when data demands expand.
Hospitals and other healthcare organizations are at particular risk for this. In these settings, the data grew so fast that congestion of legacy systems was bound to occur. When this happens, the engineering work becomes to build new roads while keeping existing traffic flowing.
Overview
New York-Presbyterian hospital is a world-renowned health institution of 40,000+ employees seeing over one million unique patients annually. Like all hospitals, their need to process data reliably and efficiently exploded over the past decade in the wake of widespread electronic health record system adoption and Meaningful Use. Hospitals also need to access hundreds of quality metrics to monitor domains of quality improvement, patient safety, and patient experience. They are accountable for these metrics to a broad range of stakeholders.
The Challenge
In this context, the application that generated dozens of metrics for weekly executive-level reporting was no longer serving the needs of this organization. The existing solution had evolved in an era where there were fewer reporting demands. The process was buckling under the strain of rapid data growth. Code that had been inherited several times had grown difficult to disentangle when things went wrong. New metrics were constantly being requested and bottlenecks had also begun to occur frequently.
Solution
The existing code had been sufficient for a smaller number of metrics, but could no longer serve the current need. This digital transformation involved breaking large scripts apart into individual metrics. This system allowed problems to be pinpointed easily and to give ownership of metrics to individual teams.
The code was also restructured to calculate first, and format second. All metrics get calculated in the master table, guaranteeing consistent values. Branching reporting tables layer on any elements that are specific to different reports, such as report-specific hover tool items and targets.
Results
Six months post-launch, this system, which had for years housed a maximum of 30 metrics at a time, now supports the weekly production of over 150. Errors were reduced by more than 90% and error resolution time reduced from days to a matter of hours when problems did occur.
Want more solution inspiration? Return to the gallery to read more.