Cardiff University student uses SAS® to identify potential 40% cost saving for district nursing service
Elizabeth Rowse of Cardiff University used SAS® to model patient demand and identify the ideal skills mix in district nursing teams to meet this
Elizabeth Rowse, a student at Cardiff University, used Base SAS® as part of her MSc in Operational Research and Applied Statistics to model the patient demand faced by a local district nursing service and to identify the optimum nursing team size and skills mix to meet this demand in a cost-effective way.
There are 27 district nursing teams in the local area. Within the teams, there are four different types of nurse: band three (a lower skill band) to band seven (team manager); each team is made up of several nurses with varying skills. The district nursing teams visit patients with differing treatment requirements, which are categorised into patient-need levels.
The local health board wanted to know if the nursing teams were spread across the city in the right way, and didn't know if they had the right skill level, or number, of nurses in each team to meet this varying patient demand.
"I was able to use SAS to model patient demand against resource, thanks to Cardiff University's SAS licence," says Rowse. "The health board were also keen to explore the potential of this analysis under Proof of Concept."
Rowse decided to model the skills mix of district nursing teams to understand how effectively they currently meet patient demand. Rowse also wanted to identify how a different team structure, in terms of skills mix and size, could help them work more effectively.
"I modelled patient demand in terms of time, travel and treatment using SAS," explains Rowse. "SAS was much easier to use compared to the alternatives, which would have needed to be run multiple times. The program allowed me to make a graph of travel time, and customisation of the output was really straightforward."
Transforming data to unlock insight
Rowse received a large data set from the local health board, covering the activity of every team in the district nursing service: this amounted to 1.5 million rows of data, each with 20 variables. The data had been collected by the district nurses themselves over the course of several months, and included data such as travel time, treatment time and patient-need level from every patient visit.
Rowse then used SAS® to split this data by team, and to analyse each individual team's data set to identify the distribution of travel time, treatment time, and how many patients were seen in each category. The results of this analysis were used as input in Rowse's simulation and linear programming model, which she built in SAS. Using the output from the model, she was then able to suggest an ideal team composition, detailing how many nurses of different abilities were needed to match patient demand.
"I found that the lower banded, less qualified nurses were only able to visit some patient categories. This meant that the higher banded, more qualified nurses were really stretched as they were the only ones who could attend to all the patients," says Rowse. "Using SAS, I found that the ideal structure included fewer low-banded nurses and more mid-band nurses who can visit and treat more patients. My results indicated that patient demand can still be better met with a restructuring of teams and a resulting reduction in cost – this represented a 40% cost saving for one team. By restructuring in this way, the whole programme could become much more cost-effective."
Despite the project being undertaken solely as part of Rowse's MSc project, one middle manager has expressed interest in applying the findings across the whole service.
Allocating district nurse time appropriately
Rowse also used her model to investigate if it would be possible to manage workloads and allocate time differently, so that patient demand could be met without the need for team restructuring. In addition to patient-facing time, each nurse also has time allocated for study leave and administration.
Rowse simulated supply time against patient demand, and found that her results suggested a different allocation of the nurses' time between these three activities from that recommended by the legacy scheduling tool that the nursing service currently uses. By adjusting the proportions of time spent on these different activities, it could be possible to meet patient demand without restructuring the nursing teams.
Providing optimum research capability
One of the modules available on Lizzie's MSc is operational research, including SAS programming, during which students are taught how to code in SAS, and learn how to manipulate data sets and carry out statistical procedures.
"SAS was much easier to use than many of the alternatives," says Lizzie. "It can deal with significantly larger data sets than competitive products, which was particularly important with this research as the original data from the district nurses included approximately 1.5 million records."
Taking SAS® to the hospital ward
Rowse is now studying for a PhD and is using SAS for her project, which also uses analytics in healthcare: this time, she is focused on hospital beds and how best to schedule operations in surgical theatres so that the availability of beds in the hospital's wards is optimised. Rowse aims to use SAS to understand how best to schedule operations, so that there isn't an influx in demand for beds one day, leading to operations being cancelled. Ultimately, she is seeking to level demand through the intelligent scheduling of operations and resources.
"SAS is an integral part of our MSc course," said Dr Janet Williams, Senior Lecturer at Cardiff University. "We chose SAS because it gives our students the skills employers want: many of the organisations they are likely to work for use SAS on a daily basis. Lizzie was an outstanding student, and we were very pleased to award her the SAS prize for Applied Statistics. I'd also like to thank the local Health Board that supports the MSc by providing interesting and real-life projects for students like Lizzie: they make all the difference."
Copyright © SAS Institute Inc. All Rights Reserved.
To model the skills mix of district nursing teams, to understand how well they meet the present demand. To identify the required skills mix and size of teams to better meet the demand.
Base SAS® is used to split the data set and model patient demand, in order to provide suggested team composition.
Results indicated that patient demand can still be better met with a restructuring of teams and reduction in cost – up to 40% cost savings for one team. Rowse has had a paper, based on her analysis, accepted for submission to an academic journal. One middle manager has expressed interest in applying the findings of the study across the whole service.
“SAS was invaluable throughout the data analysis in my MSc. I found it is possible to still meet patient demand and significantly reduce costs – up to 40% cost savings for one team. I'm pleased to see that the health board have found the outcomes of my MSc project useful, and that they are considering how they might best be applied.”
MSc and PhD student, University of Cardiff.