I mentioned last week that I would be attending our quarterly “All Provider” meeting and had been hoping that Accountable Care would continue to be the focus of physician anger rather than EHR. Although it continued to draw a large amount of complaints (along with ICD-10 and Pay for Performance) EHR was once again in the spotlight. The current issue revolved around system availability.
In a nutshell, providers don’t ever want the system (or any part of the system) to be down. They expect upgrades and patches to be magically applied with no disruption. I don’t blame them – no one wants to be without the information they need to safely (and efficiently) care for patients. We do have to remember, though, that we’re dealing with machines and networks and the people who install, program, and maintain them. Downtime can be minimized but it is not completely avoidable.
One of my providers is really fond of using statements like, “How come you can’t just patch this thing like Microsoft does?” Being a long-time user of Microsoft products, I think that shows a remarkable lack of insight. I don’t think Microsoft is particularly adept at making user-friendly patches.
The average end user typically has no idea what is in them and has to just accept them through the auto-update process without thought and frequently without concerns for timing. I love rebooting after batches – when the system is trying to shut down and it warns you not to unplug, touch, or look funny at the device because “Windows is configuring updates.” There is no estimate of how long it will take or what it’s really doing.
Our department painstakingly combs through vendor release notes to make sure we fully understand everything we’re installing and how it will impact the end user experience. We communicate, re-communicate, and over communicate using a variety of media and strategies and yet it seems to never be enough. Many of our maintenance tasks can be done with users on the system. However, there is one item that has to be done with all users logged off. We typically do this once a month after midnight and it takes about 10 minutes. You’d think that we were asking people to give up an organ they way they respond to this.
People cannot possibly be without the record! There might be an emergency! The sky might fall! I’m not talking about a hospital system here – I’m talking about an ambulatory EHR in a large group that’s about 60 percent primary care. In my experience with having colleagues contacted through the after-hours exchange in the wee hours of the morning, it takes more than 10 minutes for them to respond to texts or calls. One would think that if those 10 minutes were critical, they’d be answering instantaneously and not making the emergency department secretary chat with voice mail. Even better, perhaps they should consider taking in-house call.
Once upon a time in a land far, far away, we managed patients from home without the chart. We used the data we had (patient, nurse, ER physician, resident, intern, exchange, partners, pharmacy, etc.) to give the best advice we could. Certainly there are benefits of having home access to charts, but being without them occasionally is not the end of the world. Patients did well the vast majority of the time. We used things like clinical judgment to treat the patient in front of us, not numbers.
Of course, scheduled downtime and unplanned “events” are two different things. However, with a solid business continuity plan, they should have very little impact to clinicians. Some of my colleagues can access their disaster recovery servers directly. Others have “mini-charts” sent to a network drive every night. Some colleagues have no plan and can’t tell me the last time they actually tested a backup to see if it could be used to perform a system restore.
Despite the crucial nature of clinical data systems, we are at risk of outages and it’s time to be prepared. If you don’t know what your disaster plan is, find out. If you do know your plan, then good for you. If you’re like some of the colleagues I dealt with today who think that clinical systems are the only ones that ever go down, I offer the same challenge I gave at our meeting: take a week and see how many “issues” you have with non-clinical systems. By issue, I mean an instance where the system doesn’t perform perfectly. I think you’ll be surprised at how often they happen and how often we simply move past them and get on with our work.
How do your end users cope with downtime? Do you have processes in place to maximize availability? E-mail me.