Tuesday, July 8, 2008

Responding to Medical Error

Yesterday, Paul Levy posted an entry on his blog about a recent surgical error at BIDMC.

IS and the clinical departments of BIDMC have a very strong collaborative relationship. Working together, we first enhance processes, then automate them, since even the best technology is generally not the solution to workflow and communications problems.

Here's the application enhancement we're making as part of a process change in the Operating Room to prevent future patient harm.

Standard Operating Process in the OR includes a "time out" by all OR personnel in the moments before surgery to double check all aspects of safety - equipment, right surgical site, team readiness etc. Currently the "time out" is documented on the paper intra-operative record, which means that the scrub nurse needs to look at both the paper record and the electronic peri-operative information system during the "time out." We will add a "Time Out" button to the electronic OR journal screen containing the case times. When this button is clicked we will pop up a window with the "time out" fields. The nurse will fill in the time out information and enter her/his password. We will not allow the nurse to enter an incision time for the case unless the "time out" has been completed, with one exception - we will provide a check box on the time out screen to indicate the time out could not be completed prior to incision due to a life-threatening situation.

The standard process we've put in place to respond to sentinel events such as this one is that the root cause is reviewed with the Board (PCAC committee) and the Quality Improvement Directors. IS staff work with Quality Improvement Directors to determine which process improvements need to be made, then what additional automation should be added. Using this approach, we've created a balanced way to add new technology at the appropriate time.

5 comments:

PJ Geraghty said...

I'd read about this incident in Paul Levy's blog. Commendations to all of you for the transparency involved and sharing with the rest of us BIDMC's response.

I'd also recommend that any time the checkbox is used, the case should be reviewed by senior surgical staff in whatever QI committee is established already. Checkboxes such as this one are easy to click when someone is simply in a hurry. There should also be an audit to find out when the time-out information is entered as opposed to the case times. For example, if the system logs a time-out screen completion at 0814, but incision time is entered at 0819 as 0742, that also should be a flag for review by the committee.

Congratulations again on such a quick response.

Ian Furst said...

agreed about the check-box. It's a fact of life that people will forget to do the time-out or do it after the first incision. If you force the check-box before incision time it forces them to falsify the record when they forget. If you allow the time out to be filled in seperately you'll be able to track how often and who systematically forgets/overlooks it. My suggestion would be to allow the incision time to be completed seperately but have the time-out box automatically pop-up after the pt is admitted to the surgical theatre. ???that way you both prevent and track the error rate.

The Critical patient said...

In case folks wonder why this is not a simple fix, I defer to Drs Charles Friedman and Jeremy Wyatt's excellent text, (Evaluation Methods in Biomedical Informatics 2nd edition - Springer), where they use a venn diagram to describe the complexitiy of the convergence of medical science, evaluation, and computer science. There is much in this text about the prospective problems in foreseeing this kind of issue, whilst the retrospective solution seems so intuitive. Readers interested in the topic may find this useful. Thanks for posting the solution.

kmccrensky said...

Just a couple of things to consider...

MULTIPLE PROCEDURES/TEAMS
Cases involving multiple procedures and/or surgical teams may require more than one time out. You may want to consider having the system provide the ability to complete multiple Time Outs under certain conditions. ex. Mastectomy with Reconstruction involves Surgical Oncology and Plastics.

Otherwise you will have one process (computer) for single team/procedure cases and a different process (computer+paper) for some of your most complicated cases.

WRITE ONCE
Some systems have Time Out screens that cannot be viewed or edited once the screen has been signed. This eliminates peer pressure to retrospectively modify the Time Out record. The nurse can accurately record what happened and "blame the software" if instructed to modify the record.

TRIGGER
You may want to consider automatically triggering the Time Out screen after a certain timestamp has been signed rather than forcing users to click for it.

TIMING
At first glance one might think the logical place to require/trigger the Time Out screen is after the Patient Ready timestamp and before the First Procedure Begins timestamp. However, my research (300+ OR nurses at my last position) indicates this is not ideal. The perioperative nurses said the actual timeout (1) occurs before they document the timeout...NOT at the same time, and (2) is not initiated due to the inability to progress in the electronic record. Therefore forcing the timeout to be entered before the documeting the First Procedure Begins timestamp may not have the desired effect.

Ian Furst said...

it would be interesting to survey nurses about what trigger the time-out in their own mind. Some do it when they see me prepping the pt, others when I drape, others when I inject local and some just before I start cutting. At the beginning of a case almost no-one is at the computer screen -- everyone is busy with clinical tasks.