The most award winning
healthcare information source.
TRUSTED FOR FOUR DECADES.
Predicting admits, discharges vital
The numbers don't lie, and having a handle on the numbers is a critical part of developing effective strategies for improving patient flow, says Pamela Kiessling, RN, MSN, director of patient flow & clinical integration, clinical & business integration, and patient services at Cincinnati Children's Hospital Medical Center.
"We knew that we were not as efficient as we could be with the whole process around discharges," Kiessling recalls. "For those who just needed antibiotics to take at home, for example, we did not plan in advance sufficiently to discharge them as soon as they were ready to go."
To improve in this area, you have to be able to predict discharges, she says. "The adult world has been doing this for a long time because of their payer structure. Pediatric facilities are paid differently [i.e., in terms of DRGs], so we have not been driven to be as proactive," Kiessling says. "But now we're doing it for the right reasons: to have the patients leave on time and have no delays."
Discharge prediction is a two-level process, Kiessling explains. First, the patient has to have discharge criteria. Goals need to be specific and well communicated to the entire team, including the patient and family. The second level of readiness has to do with the team tasks that need to be completed, such as home care arranged, prescriptions written, and orders written. "The goal is to have the team tasks completed prior to the patient's readiness for discharge whenever possible so that there is no delay for the patient once she or he is ready to go home," she explains.
Communication regarding the predicted discharge date and time is critical so that the entire team can execute the plan in a timely manner. For the ED, this timely discharge means a greater likelihood of a bed on the appropriate unit when it is needed and that any delay would be intentional and predictable and only to allow the right bed to be available.
In the absence of the ability to build new beds, Kiessling summarizes, timely discharge is a legitimate way to increase capacity in a hospital that operates with very few open beds at any given time.
In developing the predictive process, says Kiessling, "you have to build in the factor that you'll be wrong some percent of the time — anywhere from 20%-30% — not because you've not planned well, but because the child may not progress as well as you've planned." Still, she insists, "for any given unit, we can be right seven times out of 10." When planning for beds, then, you should look at your predictions and build in processes to account for the "unpredicted" beds that will be needed.
Where appropriate, you can write conditional discharge orders, i.e., when the patients meet these criteria, they can go home, she says. These criteria must be patient-specific, Kiessling emphasizes. Discharge medication orders and discharge summaries are among the things that can be done ahead of time, she says.
At this point, says Kiessling, some units are 80% correct in their predictions, while others are closer to 50%. While the discharge predictions are mainly done on the inpatient side, she notes, it still benefits the ED. "Oftentimes there are delays in the ED because of bed availability," she observes. "In order to set the stage for improvement, we had to have beds."
It's difficult to track time saved by this process, she says. "Some patients may meet the criteria at 2:30 in the morning," she explains. "Should we tell them to get up and leave because our numbers need to be good?"
At the same, says Kiessling, she began to look at how to predict admissions. "We have three kinds of admissions: scheduled, ED, and direct," she notes. "The trick is to know what's going on in the population you are looking at."
In January 2009, she says, a math formula was developed that allowed the ED to predict its admissions. The formula takes into consideration admissions from the ED "yesterday," "same day last week," "two weeks ago," "three weeks ago," and "four weeks ago," she says. These data are averaged. "We then look at trends for the last month in terms of percentage of ED visits admitted to the hospital and adjust accordingly," she notes. "It isn't an exact science yet, but we're working on it."
"We are within 90%-95% accuracy most of the time," Kiessling says. "Folks in my department and the ED clinical manager figured [the formula] out, and it's pretty good."