Improvement Insights Blog
I ♥ u charts…
I got an email from the hospital where my friend Annie worked saying that they wanted to arrange a video call. It seemed unusual only because Annie and I have known each other for a long time, and if she needed anything she’d usually just call, send a text or corner me in a coffee shop. I haven’t been in a coffee shop very often lately, and I hadn’t heard from Annie, so I assumed things were going okay.
I connected to the video call and was greeted with not one attendee, but two. Annie was clearly joining from home, and an older woman with grey hair joined the call from the an office at the hospital.
“Hey, Annie!” I greeted her.
“Hey there!” she responded. “There’s a couple of people I’d like you to meet. This is Rebecca, and she’ll be filling in for me for a while at the hospital.”
Rebecca smiled. “Hello!”
“Well, hello! It’s very nice to meet you.” I said.
“The other person I wanted you to meet is right here.” She tilted the camera down and I could see a tiny newborn sleeping in her lap. “This is Isabella.”
“What?” I exclaimed. “I mean… uh… congratulations and all, but… didn’t I just see you at the coffee shop?”
“You realize that was about 7 months ago. It was still snowing, remember?”
“Oh yeah… I suppose it has been a while,” I admitted
“I’ll be at home with little Izzy for a bit, so I knew if I connected you with Rebecca you could explain things to her the way you do for me,” Annie offered.
“I’d be glad to.”
“I’m going to sign off now and put this gal down for a nap. Rebecca, you’re in good hands!”
With a wave, Annie exited the call. Rebecca and I began discussing the Quality Improvement initiative the hospital had in place.
“I just need to know what to do to look at our data, and Annie said you’d be able to show me how to do that,” she explained.
“Of course,” I replied. “Do you happen to have an example of the data?”
“I do! Let me open it up.” She shared her screen and in a moment I was looking at this:
“Okay, so you’ve got a column of labels by month, a column of Total Patient Falls, and a column of Total Patient Days, is that right?” I asked.
“Yes. Only I don’t know what to do with it,” she replied, frustrated.
“No worries. I see the hospital bought you a QI Macros license and it’s installed on your computer. I’d like you to highlight the data in all three of those columns.”
“Okay. Now what?” she asked.
“Well, if you knew which chart you needed, you’d go ahead and select that chart… but you’re not sure which chart you need, right?” I asked.
“I consider it a win to get my computer to turn on in the morning. I really don’t understand this statistics stuff,” she offered.
“Don’t worry about it, QI Macros can handle a lot of it from here. If you’ve got all the data in those columns highlighted, click the Control Chart Wizard in the QI Macros menu at the top.”
She did so, and I saw a dialog box appear on the screen:
“So what is this, now?” she asked.
“It’s letting you know it needs a little more information in order to choose the right chart. It’s asking you if the number of patient falls is a count of the patients who fell, or if it’s a count of the total number of times there were falls.”
“I’m not sure I follow,” she said.
“Does a patient ever fall more than once?” I asked.
“Oh, heavens yes. Usually men in their 20s and 30s… they just don’t believe they are too weak to walk on their own after surgery,” she replied.
“All right, so if one of those fellas falls four times, is it counted four times in your Total Patient Falls column?” I inquired.
“Yes, it is… so I would click Yes?”
She did so, and I saw a u chart appear on her screen:
“Well, that was fast… but why are some of these red? That looks bad!” she exclaimed.
“Not necessarily… in fact, in this case, it’s very good. The software is marking those points red not because they’re bad, but just to alert you to them. They’re statistically unlikely to have happened, based on the pattern beforehand.” I explained.
“How can it know that?” she asked.
“Hmmm… okay, think of it this way: If you were flipping a coin, you’d expect it to land heads just as often as it lands tails, right?” I began.
“But it’s not unusual if it lands heads twice in a row, right?”
“But it would be unusual if it landed heads 50 times in a row, right?”
“On this chart, the average of all the points is that blue line in the center. If you look at the points to the left colored blue, you see that they act like that coin we were talking about: just about half the points are above the line and just about half are below the line. See how they wander above the line and below the line?” I explained.
“But those points at the end are all on the same side of the line. Statistics tells us that if you see 8 points in a row falling on one side of a Center Line, that’s pretty unlikely. It’s kind of like that coin turning up heads many times in a row. Statistics tells us that these results are unlikely to be caused by the same factors that caused all the other points. There’s something else likely to be a factor here… usually, this means something got changed. When did you say you began the new procedures concerning patient falls?”
“Well we did all the planning last November, but we didn’t roll it out until December.”
“That’s what this chart tells me as well. You changed the way you do things, right?”
“In my business, that’s called a Process Change, and QI Macros can identify things like that and change your chart automatically. Do this for me: First, click anywhere on the chart itself.”
“Okay, I did.”
“Now at the top, you’ve got a new menu item: the QI Macros Chart Menu. Click on the QI Macros Chart Menu, then click on the Process Change Wizard.”
She did so, and I could see the software make changes to her chart and re-calculate it:
“Well isn’t that just amazing?” she remarked.
I smiled. “The software is pretty incredible, but it looks like the work you’ve done at decreasing patient falls is what’s amazing. This shows that you used to have about a .6% fall incidence and your efforts decreased it to about a .42% fall incidence. You decreased it by a third. Good job!”
“Why are the red lines all… wiggly?” Rebecca asked.
“Those are called Control Limits; the Upper Control Limit and Lower Control Limit. They mark the levels where about 97% of all your points should fall. The reason they’re not flat lines is because you don’t have the same amount of patients each month, so it’s taking that into consideration.” I responded.
“How so?” she asked.
“Well, let’s say you looked at two months of patient falls, and the first month was exceptionally quiet and you only saw 4 patients, but one of them stumbled and fell, that’s 25%, right?” I asked.
“Right…” she answered, hesitantly.
“And then the next month, you saw 4,000 patients, but 1,000 of those patients fell, that’s still 25%, right?”
“Ohhhh…” she said, nodding slowly. “One patient falling out of four could just be a fluke from such a small sample, while 1,000 patients falling out of 4,000 means a much larger problem, right?”
“Exactly,” I agreed.
“Well, you’re just as helpful as Annie said you’d be. Thank you!” Rebecca continued. “There’s only one other thing that Annie wanted me to go over with you.”
“All right,” I said. “What’s that?”
“She wanted me to make sure you aren’t still eating lunch at the hot dog vendor’s cart once a week.”
We both laughed. “No, you can tell Annie that I’ve cut way down on that.”
“Good,” Rebecca said jovially, but with a hint of sternness. “I’d prefer if we kept meeting via video call, and not with you looking up from a gurney in my E.R.”