By Eric Fischer, Senior Product Manager, Data Solutions
In part 3 of our series on how Accruent Data Insights works and how it helps the healthcare technology management (HTM) profession, we focus on end-of-support dates for medical equipment.
For most healthcare technology management (HTM) professionals, managing end-of-support dates for your medical devices starts when you receive a letter from the medical device manufacturer that a device is going to be end of support at some future date. The letter serves as a warning that service or parts will no longer be available, and it lets you know about the latest models that could be used as replacements.
Your ideal next step would be to somehow record this letter in your CMMS system, typically by scanning and uploading it to the CMMS asset or model record. In addition, you would record the date so it could be retrieved easily when the yearly capital planning cycle comes up, allowing you to justify the device’s replacement.
The problem with this process is that it is outdated, manual and time-consuming. As a result, documenting the end-of-support (EOS) date is often forgotten or missed. But with a solution that automates this process, you gain sufficient date information to act upon during the annual capital planning cycle.
Using Artificial Intelligence to Determine EOS Dates
Accruent leveraged its data warehouse of 12 million asset records to determine if it could derive a good EOS date for each medical device model. This was a challenge because, as I mentioned, the source of the data is in end-of-support letters that are sent to various individuals, and it is often not recorded.
First, Accruent started with the low hanging fruit. We checked to see if the end-of-support dates were adequately populated from the process described above. This resulted in some success, but it also led to problems when different dates were recorded by different hospitals or users. Also, we suspected that the manual nature of the process described above would lead to missing dates.
Accruent then called upon our customer base to see if they would be willing to donate their letters to us in exchange for electronically formatted dates in return. Unfortunately, many organizations did not retain these letters or were unable to dig them up (illustrating how disorganized keeping letters can be). In addition, we found that many organizations were uploading the letters to the CMMS but not recording the date.
However, some organizations did have these letters available and were willing to share. We also searched through the file systems in Accruent-hosted CMMS solutions to retrieve as many end-of-support letters as possible.
After we gathered these letters, we partnered with Butler to help us extract data from the documents. Butler is an innovative technology startup in the Bay Area that specializes in using artificial intelligence to solve this exact problem.
Validating the Artificial Intelligence
The technology behind Butler worked amazingly well! We used Butler to accurately find the fields we were searching for across these documents, even though each letter looks dramatically different from the others. However, we found that the letters themselves are often lacking in standardization: Many letters did not even explicitly state an end-of-support date and some letters had multiple dates for various reasons, such as dates based on serial numbers.
Below is how we solved each issue we encountered:
No explicitly stated end-of-support date
Butler extracted the date of the letter and used that as a proxy for EOS date.
Multiple end-of-support dates
Accruent indicated when various dates existed and provided a link to the letter so the user can determine for themselves which date is appropriate.
Dates written in different formats (e.g., March 21, 2011 vs. 3/21/2011)
Accruent used pre-existing date-parsing programs to standardize date formats.
Mapping EOS Dates to Standard Model Names
Once we had good end-of-support data, we then had to map the data to the Accruent Data Insights standard model names. This was a quick process because, as covered in the first blog post in this series, we had already developed the logic to match dirty medical device names to our standard.
Now that the process has been completed, you can view the Data Insights model lookup screen to see when a device is at end of support:
Additionally, you can use the Data Enrichment report in Accruent Data Insights to determine where EOS dates are missing in your CMMS data.
Why does this benefit your capital planning? It helps because you can intelligently apply the end-of-support date data to the capital planning process by modifying the configurable “Replacement Priority Score” – so that the score becomes higher if the device model has an associated end-of-support letter.
How can Accruent Data Insights Help You?
We often hear customers talk about the problems identifying devices that are end of support. With the good EOS data Accruent Data Insights provides, you have the information you need during the capital planning cycle.
In addition, Accruent Data Insights is automated, so getting useful results is quick and painless. Even simply mapping your cost data is a simple and easy. If you’d like to learn more, please reach out! Ask your Account Executive for a free data cleaning report card so we can show your how this technology can help you. As always, we can be reached at firstname.lastname@example.org.
Looking for the full Accruent Data Insights Deep Dive weekly series? See:
- Accruent Data Insights Deep Dive Part 1 – Standardizing MDM and Model Names
- Accruent Data Insights Deep Dive Part 2 – Standardizing Medical Device Categories
- Accruent Data Insights Deep Dive Part 3 – Mining for End-of-Support Dates
- Accruent Data Insights Deep Dive Part 4 – Determining Medical Device Costs
- Accruent Data Insights Deep Dive Part 5 – Measuring Life Expectancy