Latest / The Operator's Manual with Fexingo: Daily Decisions, Tactical Wins, and Business Execution / How One Hospital Cut MRI Wait Times 45 Percent With a Scheduling Tweak
Transcript
- Lucas: So there's a hospital in the Pacific Northwest, about three hundred beds, and their MRI wait times were stuck at around fourteen days. That's not the worst you'll see, but it's enough to push urgent cases into the danger zone. Luna: And they didn't buy a second machine or hire more radiologists, right? Lucas: Exactly. They changed one thing in their scheduling software, and wait times dropped to around eight days in about six weeks. A forty-five percent reduction with zero capital expenditure. Luna: That's the kind of story I want to hear. What was the tweak? Lucas: So before we get into the mechanics, I want to mention something that keeps this show going. We're able to keep these conversations ad-free because of listeners who chip in a few bucks every now and then. If today's episode helps you see a bottleneck in your own work, and you feel like the coffee's on you, the link is buy me a coffee dot com slash fexingo. That's it. No perks, no pressure. Luna: Yeah, and it genuinely helps us keep digging into stories like this one. Now, what was the tweak? Lucas: The scheduler was booking every MRI slot as a fixed thirty-minute block. But the actual scan times varied wildly depending on the body part and the protocol. A head scan might take fifteen minutes, a liver scan with all the sequences could push forty-five. Luna: So they were padding every slot to the worst case, and then the machine sat idle when a scan finished early. Lucas: Right. And they couldn't book extra scans in that freed-up time because the system treated each slot as rigid. So they switched to a variable-length scheduling model. The software now estimates the scan duration based on the triage notes and the specific protocol, and it leaves small buffer blocks for overflow. Luna: And that alone cut fourteen days to eight? Seems almost too simple. Lucas: It's not just the duration change. The buffer blocks matter just as much. They used to have no slack at all, so any delay rippled through the whole day. Now, if a scan runs long, the next patient waits less because there's a built-in cushion. Luna: So it's a queue design problem, not a capacity problem. Lucas: Exactly. And that's the lens I want you to walk away with. Most operations assume you need more machines or more people. Sometimes you just need to model the variability honestly. Luna: There's a classic stat that in healthcare, utilization and wait time have a nonlinear relationship. Once you hit about eighty-five percent utilization, wait times explode. Lucas: That's the queueing theory curve. The hospital was probably cruising around ninety percent utilization, so even a tiny improvement in how they packed the schedule freed up enough slack to break the spiral. Luna: And they didn't even need to see more patients to justify the change. They just cut the backlog by making the schedule more realistic. Lucas: The key was that they stopped treating a fifteen-minute scan and a forty-five-minute scan as the same unit of work. Once the software knew the difference, it could pack the day more intelligently. Luna: So what did they do with the extra capacity? Did they book more scans or just reduce the backlog? Lucas: Both, actually. They started by clearing the urgent cases that had been waiting, which took about two weeks. Then they opened up same-day slots for primary care referrals, which cut down on no-shows because patients don't forget an appointment they booked this morning. Luna: That's the kind of cascading win that makes operations people giddy. Lucas: And the finance folks too. The hospital estimated they avoided roughly two hundred thousand dollars in overtime and outsourced imaging costs over the next quarter. Luna: Now, you mentioned they used triage notes to estimate scan time. That's a data input problem as much as a scheduling one. Lucas: Right. The radiologists had to actually type the protocol into the order, which they sometimes skipped. So the hospital added a dropdown menu that made it two clicks instead of free text. That's a small nudge, but it got compliance up to about ninety-five percent. Luna: It's funny how the biggest wins often come from making the easy thing the right thing. Lucas: The same pattern shows up in other industries. Airlines overestimate taxi times at some airports, and they pad their schedules so much that the actual block time is longer than necessary, which hurts on-time performance. Luna: And in manufacturing, if you set cycle times based on the fastest worker, you build in constant firefighting. Lucas: The principle is universal: match your scheduled duration to the real distribution, not the worst case. And then build in a small buffer that absorbs the tail. Luna: So for a listener running a clinic or a repair shop, what's the first step they should take tomorrow? Lucas: Pull your actual completion times for the last ninety days. Just look at the median and the ninety-fifth percentile. If the median is thirty minutes and the ninety-fifth is fifty, you're probably booking everyone for fifty. That's your signal. Luna: And then you adjust to the median plus a small buffer, and you monitor the spillover. Lucas: Exactly. You don't have to solve for the worst case every time. You solve for the typical case and let a few overflow slots catch the long tail. Luna: I think the part I love most is that this hospital didn't need a big IT project. They tweaked two screens in an existing system. Lucas: That's the thing. The software vendor had the feature built in all along. Nobody had ever turned it on because the operations director assumed scheduling had to be fixed blocks. Luna: So the real bottleneck was the assumption, not the machine. Lucas: Right. And that's a lesson that travels well beyond healthcare. Before you buy more capacity, question your unit of work. Is every appointment, every task, every customer interaction really the same length? Because if it's not, you're leaving efficiency on the table. Luna: So if I'm a service manager, I should be asking: what's my equivalent of the thirty-minute MRI slot? Lucas: That's the question. And the answer usually comes from the data you already have but aren't looking at. Luna: A quick follow-up: did they see any downside? Longer waits for the buffer slots, or staff feeling rushed? Lucas: The techs actually liked it more because they weren't constantly behind schedule. The buffer blocks gave them a breather, and they could use that time to prep the next patient. Patient satisfaction scores went up too, because they weren't sitting in the waiting room as long. Luna: So it was a win win win. Rare in operations. Lucas: Rare, but not magic. It's just what happens when you align your schedule with reality instead of a guess. Luna: So what's the one thing you want a listener to take away from this episode? Lucas: Look at your longest-running bottleneck and ask if the scheduled time actually reflects the work. If you're always booking worst-case, you're always underutilizing your capacity in the best case. Luna: And a lot of times, the fix is just a few clicks away. Lucas: Yeah. Sometimes you don't need a bigger hammer. You need a sharper pencil.