Improvements in the stroke care pathway have been catalysed by nationwide routinely collected data, standardised through the Sentinel Stroke National Audit Programme (SSNAP). However, Professor Dame Caroline Watkins, Director of UCLan’s Applied Health Research hub and leader of their Stroke Research Team, and Professor Liz Lightbody, Professor of Stroke Care and Improvement at the Faculty of Health and Care, School of Nursing and Midwifery and Lancashire Teaching Hospitals NHS Foundation Trust, suggest that a more granular digital approach to capturing data from admission to outcomes could further enhance patient care quality and make research results more robust.
We are increasingly recognising the need to understand if changes to treatment and care are improving outcomes for patients. This is vital, as making changes to services and systems may increase healthcare costs as well as the demands on the workforce. It appears to be very much the focus of projects such as SBRI Healthcare’s competition for technology developers and innovators to address health inequalities in stroke care, and to facilitate the collection of evidence in real-world settings from early diagnosis to rehabilitation and life after stroke.
There is no doubt that in order to make informed stroke care improvements, we would benefit from a better mechanism for capturing outcome data. But context and detail are vital if data is to be interpreted properly for the improvement of patient care quality.
At the moment, SSNAP data collection includes case mix, process and some hospital outcome data, but the collection of longer-term outcomes in the community is patchy: whether people have been seen by key professionals, for example, and if they have had the required tests and investigations along the stroke care pathway. It doesn’t necessarily give the detail of the results of the assessments, or whether the subsequent management reflects the results. Just because somebody has seen the speech language therapist and had a swallowing assessment, it doesn’t mean that the initial recommendations for management have been followed, or whether the patient has been reviewed and plans changed to reflect changes in their condition.
This lack of granularity can be seen at every stage of the care pathway. At the beginning, did the patient get enough food and fluids, did they have a catheter put in unnecessarily – and if so, was it removed? Did somebody check that the catheter was needed before the patient was sent home with it still in? Were there any continence promotion activities? Did anybody assess why the patient was having incontinence episodes, what caused them, and was an appropriate management plan put in place? Those questions are not asked later either, so we don’t know whether the plan has been any use or not. If it was available, data could tell us whether people were following guidelines or not more closely, and we could see if outcomes were linked to adherence or lack of adherence to those guidelines.
A mechanism for better data collection in stroke care
There is certainly a role for a digital stroke care pathway which provides a mechanism for the routine collection of data throughout, addresses some of the gaps in data availability, and helps with researching the effectiveness of patient care in a more contextual way. If we had good, routinely collected, anonymised data, we could also do research without having to get individualised patient consent, thereby negating having to approach patients at difficult or stressful times to gain their consent.
The current need to recruit patients who are able to give consent for their data to contribute to a study creates some challenges – for example, to the validity of the design of research studies: simply knowing they are in a research study can affect a patient’s outcomes. Furthermore, the inability to recruit all patients because of the need to gain consent can also introduce bias.
It is easier to recruit milder patients who have had a stroke, those who can understand what is being asked of them, and those who can communicate their approval. For example, the preponderance of mild stroke patients in a recent study of vocational rehabilitation meant that there did not seem to be a difference in outcome between the intervention and the control groups. But that doesn’t suggest that vocational rehabilitation doesn’t work – it just means it didn’t make a difference for the milder patients who were included in the study. A more comprehensive sample, with mild, moderate and severe stroke patients could have told us whether it was beneficial or not, and for whom. The opportunity to easily collect more detailed data from a more comprehensive sample of patients would help us to do better, more useful research, and to understand more about what we do and how it affects patient outcomes.
As things stand, SSNAP data is certainly useful in terms of overall indicators of quality of services, and it has no doubt been instrumental in getting some services to realise that they weren’t doing fundamental aspects of care. They will have had the ammunition to insist on additional resources to put processes in place to try and improve care. But essentially, it is a large, standardised dataset and the idea of adding to it – which might be more work – is a challenge for everyone. So, if we add more data capture requirements, we have to consider how to make that less work and not more. At the same time, the data itself has to be more informative about what’s actually happening with individual patients.
Data value in real time
However, clinicians won’t spend their time collecting additional data unless they think it could be useful for them in their day-to-day work, rather than just auditing practice. Capturing data has to be useful in real time as well as retrospectively: What happened with this patient today? What have we done to them? What are they looking like? What does the future look like for them, based on what we can see right now?
Clinical staff are very short of time. A digital system that saves time and gives them a more streamlined way of inputting information, rather than writing lengthy case notes or nursing notes, might persuade them that data capture is a value-added component of their job. In order to achieve this, staff need to feel comfortable with the format, the analysis, and the ease with which they can interrogate it. Ultimately, if you’re going to collect data for research or practice, interpreting it properly will always depend on the detail, and the ability to interrogate it.