Actionable Analytics and Data
In an earlier column I wrote about the reality of hospitals losing money by the bushel basketful. This is in spite of the fact that many US hospitals are armed to the teeth with internal IT systems producing an endless stream of operational data ranging from how much bread was used in the cafeteria in any given week to how many CT images were generated and how many slices remained on a given tube. Early in my career I was attending an echocardiography conference in Indianapolis where the course Director was none other than Dr. Harvey Feigenbaum, the father of echocardiography. During his introductory comments he made the following statement about all the data pouring out of the “new” 2D echocardiography systems and the reliance that a lot of the “new to the field” cardiologists were placing on that data – “I would rather have no data than bad data.” It was up to the cardiologist to determine if the echo data was actionable or not. For data to become actionable it must have two core attributes; (1) it must accurately represent what is going on in a system being evaluated, and (2) it must provide actionable analytics.
At Acertara we developed a software system called Acertara Equipment Management, or AEM. This system was designed to capture data in the probe testing and repair process that would accurately determine failure processes that were occurring in hospital use. AEM was used to highlight potential design or manufacturing issues that led to a probe failure. We shared this data with the OEMs and their feedback allowed our engineers and repair team to develop prophylactic repair processes that would mitigate such design or manufacturing related weak spots. AEM produced actionable analytics. Hospitals cannot get their costs under control using data like “four probes died this month”. That is not actionable data, that is an obituary! Actionable analytics would be “we have 78 diagnostic ultrasound probes in our active inventory, their use pattern and repair history indicate that the following four probes should be rotated through testing on a priority basis to prevent a failure during use: 1) S/N 308PY…etc.” That is true asset management: get the probe tested prior to failure (based on analytics), schedule its time out of the hospital, and save money that otherwise would have been unwisely spent on either repairing a catastrophic probe failure, or being forced to buy a replacement.
I look forward to seeing many of you at the upcoming RSNA meeting in Chicago.
Until next month,