Rather than spending months getting a data warehouse up and running you can do a “commando-style raid” on data sources in a week and discover how it might be able to help you. In the data raid, an intrepid analyst can go on a voyage of discovery into what was previously “data incognito”.
Powerful insights are more accessible than ever before. This opens great opportunities to do things differently. It is a great way to see what the data can tell you.
But what exactly can you achieve from these data insights?
In my experience working with NHS customers, it’s clear organisations can make significant gains that can have a direct impact on patient care, staff productivity and organisational performance.
We use data discovery very specifically as an analysis of a care process. We are looking for patterns and outliers that tell us how well a process is working. We do that by looking at the data that describes the processes and pathways used to deliver care, how many people, how long, etcetera.
Often this data is far from the mainstream data in the warehouse. It might be in any of the hundreds of minor applications, in spreadsheets with rosters or other data. We have used all sorts.
Finding opportunities not problems? Looking to the future
Currently, much NHS data is used to report KPIs and their exceptions, and problems. With a data raid we are looking for areas for improvement. Starting with the end goal in mind we ask the question “Can we find ways to improve the quality and cost of what the organisation delivers?” Be it patient care, administration, sales or services.
Combining data, process and people
As well as using a range of data sources, we also like to add the thoughts, views and opinions of the people delivering the processes.
By systematically analysing the inputs and outcomes found in any typical NHS systems data, we can create a robust, evidence-based assessment of the efficiency and effectiveness of any department. For instance, we look at the volume of attendees, time of day, how long they waited, how often they were seen, what their outcome was. We combine that data with on-site observations and with interviews with staff, to gather insights and preferences.
It’s a conversation, not an interrogation
We often find that IT trained analysts have a list of answers that they need, and by golly, they intend on getting them or telling the world that the data is “not good enough”. Our experience is that a more flexible approach can leap these issues. Often you can get some sort of a proxy or replacement measure that the folk on the ground will agree is good enough. We’ve used time in a cubicle as a measure of treatment resources and time. Not perfect but we could then see the bigger picture.
Eliminating waste and driving self-sufficiency
Essentially, through data analytics, we aim to identify process “waste” that can be eliminated. Waste is any step or action in a process that is not required to complete a process. That might include patient waiting times, unneeded interactions, repeated interactions or re-work. In a clinical setting, anything that is not needed to deliver diagnosis, treatment or prevention is waste.
This is not the way you should do it! There is an entire blog to be written on the objections that we see from clinicians and informatics professionals.
People often tell me that their data quality is too poor. Data quality is fitness for purpose and generally, the quality they have in mind is that needed for auditable reporting. So far, every “data set” that is used to track activity is a useful and meaningful place to start. There are always great insights to be found. But like any “Discovery” it might not be the insights that you expected. In one case of high Did Not Attends (DNA) we quickly discovered that a typo in the appointment letter meant patients couldn’t contact the clinic easily to arrange a different date. We didn’t know what answer we were going to get so couldn’t judge the quality.
Do you need a single version of the truth?
For the audit and commissioner reporting, you need consistent data that doesn’t change depending upon who you ask. However, accountants will tell you about a “full and fair reflection” and is this “a material sum”. They recognise that you can only go so far with accuracy.
What is the single version of the truth for the number of people who attended A&E between 9 and 10 pm? I can give you a single version of the truth in 6 weeks or I can tell you it’s about 30 in real time. But I can also tell you in real time that that is about 25% more than you would expect and that it infers six beds needed. That version may get revised later but that “truth” could be really useful. As Oscar Wilde commented, “The truth is rarely pure and never simple”. Perhaps a single version of the truth is a mirage and we just need to have data that we understand.
But where is the data warehouse?
A data raid is essentially finding out how you can use the data you have to address issues that are worth addressing. But in doing this we also identify the data that matters. If we want to create new data in the “normal” data warehouse then we know exactly what is needed, where it is and any transformations.
Equally, for some solutions, the in-built data manipulation can simply be automated on a local or cloud server in minutes and the job is done.
What does it look like?
The clinical services are crying out for these types of insights and it is up to us in IT and informatics to meet the need. Yes, there may be difficulties and challenges, but this modern technology offers us a way to meet that need.
Analytics to improve processes in action
Here are three examples where we have used analytics to improve processes within Healthcare organisation:
We have recently completed this analysis for an A&E in an NHS England trust. It was designed to help them return to 95% treated or admitted in 4 hours and saw a significant increase in their performance.
Community Care. Fixing an unhappy and poorly performing service.
Combining clinical patient data with community nursing services data. The data showed that the old model of counting patients and sharing them between care staff had become inefficient and ineffective as patient care profiles had changed. We were able to use the data to develop a more accurate workload planning that ensured staff had adequate time to deliver the care required. This delivered a much happier service for both nurses and patients as well as improving staff retention.
The data used was a combination of a Microsoft Access application with the main PAS data and some reporting spreadsheets.
Ensuring Patients can be moved as needed
The tension was rising between the portering service and clinical staff unable to move patients in a timely manner.
We found we could identify when portering constraints affected the A&E and Outpatient clinics, and equally when resources were scheduled but unneeded. Using simple predictions to inform rostering led to reduced waits in A&E and better service to clinics as the client could adjust resources to meet the actual portering needs that had changed over the years.
Using data from a clunky, 15-year-old SQL tool to analyse portering workloads. We combined this data with A&E and Outpatient data and interviews with the users. A range of issues were identified that were mainly to do with the wrong staffing levels at the wrong times. Also, the opportunity was identified to upskill some porters to be able to deliver certain patient transfers without nursing assistance. Conceptually a simple solution but the availability of the data analysis took the evidence from emotional narrative informed by some experience to a much less contentious numerical exercise.
What else is possible? We’ve only just scratched the surface.
Unlock the power of your data analytics and improve processes
We have a highly skilled Healthcare team dedicated to supporting NHS customers with their Business Intelligence and Data Analytics.
To find out how Trustmarque can help you to unlock the power of your data email me at [email protected]
About Dick Wall
In 2011, Dick joined Trustmarque as a BI consultant for the public sector and from 2013 went on to specialise as a BI Solutions Architect for Healthcare. His role is to ensure active business intelligence delivers measurably better outcomes for NHS customers. Dick advises the Trustmarque Healthcare team on implementing BI solutions to improve operations, quality, speed and cost. He ensures that teams can use data in a methodical way to improve safety, quality and risk.