Data Tolerances and Intolerances

By | 2019-10-03T15:11:06+00:00 September 24th, 2019|

We all know getting accurate hospital data can be a challenge. These barriers have lowered expectations industry-wide for what counts as a usable data set. But it has lowered our tolerances in other ways as well.

I was recently reading a peer-reviewed article in a respectable journal on various methods the author used to improve surgical block time utilization. In this article was a footnote that caught my eye. This is how monthly block utilization was counted:

“… based on available work days, which is 22 days / month.”

This footnote is essentially saying that block utilization is calculated using 22 days’ worth of available block time, no matter what. This means months with 23 work days will report 4.5% higher utilization because of that extra work day. And months with 21 days will report 4.5% lower utilization due to the missing day.

If we add in US federal holidays, the disparities get even larger. Novembers that start on a Saturday have only 18 workdays because the month has both 30 days AND two holidays — Veteran’s Day and Thanksgiving (*). If we count the day after Thanksgiving as a holiday, November can have as few as 17 workdays. This is 5 days fewer than the assumption above; about 22% built-in error. This is not a rounding error.

This assumption of 22 work days / month is fine for a quick back-of-the-napkin calculation. But any other modern industry would flatly reject this formula in a peer-reviewed journal. Airlines don’t build revenue models on an “average of 100 seats per plane” — they know different planes have different seating capacities and will only take seriously models that account for this variance.

Another footnote in the article alludes to the difficulty the author had in extracting the data and producing the necessary comparative analysis. Having witnessed a wide range of perioperative software, we at Hospital IQ can sympathize with this sentiment and understand that accounting for variable days in months and holidays can be a lower priority given the other challenges in dealing with healthcare data.

Furthermore, if we are making simplifications to streamline data analysis, the average number of workdays in a month, excluding holidays, is 21.75, not 22. So not only is this “22 days / month” an oversimplification, it’s an imprecise oversimplification.

The goal of this blog is not to dismiss the good intentions of this article’s author. The content is clearly well thought out and should be appealing to perioperative managers seeking to improve utilization. The point we’re making is the healthcare industry has such low expectations of data management they’re willing to accept such an elementary and imprecise simplifications.

Healthcare leadership can and should demand more. They should feel entitled to modern software that is usable, and can deliver basic analytics without users seeking sympathy from their peers. In 2019, software should be able to calculate the number of days per month at a bare minimum.

(*) Months with 30 days can have between 20 – 22 weekdays depending on how the weekends align. Februarys can also have 20 – 21 weekdays. So February and November have the same minimum number of days. And November has two holidays, while February only has one. The next November that starts on a Saturday is 2025

Leave A Comment

Hospital IQ, Inc.
© Hospital IQ, Inc.
All rights reserved

Privacy Policy
Hospital IQ, Inc.
Privacy Policy
Blog
Blog
Twitter
@HospitalIQ
LinkedIn
LinkedIn