‘Big data’ helps KU Medical Center researchers figure out some of nursing’s toughest challenges
September 18, 2020
By Anne Christiansen-Bullers
Faculty at the University of Kansas School of Nursing are taking a unique look at the value of nursing by combining electronic health records, human resource records and other data to measure the effects of nursing on pain management.
In fact, their work, "The Value of the Nurse in Pain Management and Acute Care Length of Stay" was selected as a Distinguished Abstract by the Midwest Nursing Research Society (MNRS), a 1,200-member organization encompassing nursing researchers from 13 midwestern states.
Amy L. Garcia, DNP, RN, CENP, director of the Office of Practice and clinical associate professor, and Ellen Harper, DNP, RN-BC, MBA, FAAN, clinical associate professor, presented at the annual conference. It was held, like so many other conferences in the COVID-19 era, via videoconferencing software.
The study data showed that expert nurses help patients manage their pain following surgery differently than novice nurses. They studied 1,728 nurses who cared for 43,269 separate patients at a children's hospital. By linking the nurse's personnel records, which showed their experience level, with electronic health records (EHR), the nursing interventions and the pain levels the children reported, they could take a new look at the value of on-the-job experience.
"We need to help nurses make the best use of their limited time with each patient."
Ellen Harper, DNP, RN-BC, MBA, FAAN
"We were able to show that expert nurses were better able to control pain during the first 24 hours after surgery," Garcia said. "We were also able to show that the interventions the expert nurses chose to use were different than the interventions that novice nurses chose to use."
Novice, experienced and expert nurses consistently used some of the same techniques, such as comfort measures, hot or cold packs, and distraction. But expert nurses, the researchers found, were more likely than novice nurses to use the following techniques:
- Reposition the patient.
- Teach patients and families about strategies they could use to control pain.
- Administer the medication prescribed on an "as needed" basis more frequently.
This information becomes important when healthcare providers create work schedules for nurses, Garcia said. "We don't want all of our expert and experienced nurses on one shift, and new or novice nurses on the night or weekend shift. We need to seek balance."
Impact on patient care
Harper said the results of this study could help nurses decide which interventions work best for patients, leading to optimal care for patients and the best use of time for nurses.
"We could start to look at work that doesn't add value and do a better job of knowing which activities provide better outcomes," she said.
Harper said nurses have traditionally added new techniques as the result of ongoing research, but rarely do they stop anything they've done before. The result: nurses have an ever-longer list of the things they need to do without knowing which items on the list can be eliminated without harming the patient.
"Nursing time is valuable," she said. "We need to help nurses make the best use of their limited time with each patient."
Tool for staffing
The research also could help with staffing, which has traditionally divided the workload by counting the number of patients and dividing by the number of nurses on the shift. That's incredibly equal -- until it isn't.
"We wanted to get deep down into the data. And we decided that a key way to do that is to look at the individual provider, the individual nurse, with the individual patient and then scale that up."
Amy L. Garcia, DNP, RN, CENP
"We know that patients are not average. Nurses are not average. And care settings are not average," Garcia said. "Some nurses have a stronger skill set. Some patients have more acute needs. Some settings have a better layout or better access to the resources needed to provide care."
"So, we wanted to get deep down into the data. And we decided that a key way to do that is to look at the individual provider, the individual nurse, with the individual patient and then scale that up," she said.
Getting the right ‘dose' of nursing
Harper said most research on nursing practice isn't focused on that level of detail. Instead, much of the data is focused on nursing units or even hospital floors. By examining data about nurses and patients at a more precise level, researchers can help determine the optimal number of nursing hours and skillsets to match patients' needs.
Harper explained, "This research is possible because hospitals are using EHRs that capture metadata, best described as ‘who' and ‘when' the data was recorded. Using metadata allows us to match one nurse to one patient at a time."
Since statistics are the fundamental building blocks of the research, the team also included Lakmal Walpitage, business intelligence analyst in medical informatics at KU Medical Center. Walpitage was responsible for data management, which he said was challenging because data came from "multiple source systems in high volume." He also completed statistical analysis. "Thankfully, the data analysis infrastructure available at KU Medical Center provided us with the bandwidth to process such a huge volume of data and complete this ‘big-data' analysis," Walpitage said.
That analysis was important so data could be broken down into one nurse-one patient interactions, which the researchers called "doses" of nursing.
Why use the word "dose," a term usually associated with medication instead of nurse interventions and interactions? Because like medication, one nurse's "dose" is more effective than another, one more expensive than the other, and all need to be measured to best understand what best suits the patient, the researchers contend.
Implications for future research
Studying these "doses" could be helpful in all kinds of research. "It's opening up new possibilities, especially as we build out the gaps in the current data standards," Garcia said. "Common data models help us research very big sets of data, including years of data from many different hospitals and health systems. But the common data models rarely have information about the experience or education of the nurses, therapists, and physicians who are providing the care."
"So, that is the next step: to use this new knowledge to scale up with more variables and very large data sets that will help us improve care," Harper said.
A slide from the researcher's presentation shows how the patient's pain score is managed the first 24 hours after surgery. Slide courtesy Amy L. Garcia.