The challenges presented by pursuing novel and heterogenous methods of prioritising elective waiting lists.
Throughout the country, integrated care systems are trying to redefine what we have previously understood by the concept of treating in line with clinical need. The delivery plan for tackling the elective backlog called on systems to analyse their waiting lists through the lens of distinctive characteristics including deprivation, age and ethnicity, to eventually develop operational plans which delivered clinical services in line with this expanded definition. We are beginning to see the response to this challenge, with some NHS providers showcasing work which reprioritises the elective waiting list by taking account of factors influencing inequality. The intention behind these projects is both important and impressive; seeking to revise an individual’s position on the waiting list by accounting for a series of clinical and non-clinical factors. Nevertheless, there are a number of hazards in both the methodologies being employed and the organic nature in which the work is being developed that should be discussed promptly and openly in order to maximise the effectiveness of the overall project.
Sponsored by
Algorithms are data hungry and of course, “You are what you eat”, so a good diet of data items is essential when it comes to a complex prioritisation system. Where a patient tracking list uses new data sources such as those we know are now being incorporated, even though we are familiar with them, sadly the quality of these data is by no means guaranteed across the NHS. For example, taking account of the number of accident and emergency attendances and their impact on an elective pathway is no mean feat as these are often managed entirely separately in provider’s electronic systems. You would have to assume that these new data items will need to be collected and interpreted consistently for each person on the waiting list. How the algorithm equitably handles empty fields or those populated incorrectly will also be key.
Of course, some areas are much newer to us. The way that we interpret and apply indicators of deprivation is a very new science to the NHS and remains relatively unrefined, relying on using demographic factors such as postcodes. Moreso, any social value judgments, for example which individuals are at greater socio-economic risk of waiting longer is simply not a pitch on which NHS provider organisations have ever played. How the weighting of these factors is applied will have big impacts on the results of any relative reordering of patient waiting times and so it is vital that the results are modelled and tested for sensitivity. Naturally, any team developing such an approach will look to do this but this would be enhanced enormously with support from experts in this field to help with the burden of managing these new risks and spread a consistent methodology across the NHS. Unfortunately, something that no one will have the benefit of is any longer-term analysis of any unintended consequences on elective waiting lists because any such effects will simply not be visible yet.
Many provider organisations are looking to adopt this kind of software to support their waiting list reprioritisation. Certainly, organisations starting from a shared foundation is a good thing but a natural tendency for adaption as opposed to adoption brings some risks. It is inescapable that different NHS providers will need to modify variables and weightings to suit their own system – and rightly so – but this customisation leads to a hugely expanded range of possible outcomes. It is vital that material health inequalities are not replaced by a set of systemic inequalities hiding inside a multiverse of different waiting list algorithms.
Finally, there is the issue of modelling which is overtly agile in its design that is to say subject to regular change and revision. When the principle is applied to something like demand and capacity modelling or other types of service planning, flexibility can be of great benefit. When dealing with a list of individuals waiting for surgery, stability and predictability have been the objective for teams managing these services. It follows, with an agile model, that people may move around the waiting list as their own variables and the variables of others around them change. Beyond presenting some difficulties of how this should be managed from a booking and scheduling perspective, there are several other consequences that should be visible from the outset. It may well make communications with patients and primary care clinicians more complex, specifically in relation to waiting time expectations and prehabilitation needs. It also presents us with some issues around the management of waiting times standards and providing assurance over the successful treatment of long-waiting patients across the short-term timeframes often required.
Set the NHS a challenge and it will rise to it. The pioneers in this work should be praised and supported but it is time to look at the questions that this vanguard work presents and review whether it is necessarily the right project to be encouraging local innovation for much longer. NHS England’s transformation directorate have this field of work within their scope and so there is a good prospect of national leadership, direction and support in this area. At this point in the development of these tools, it is crucial that this regularisation takes place.
Ultimately, a waiting list is inherently an amalgam of many individuals. The burden of effectively managing these waiting lists in new and complex ways increases exponentially when you accept that, for it to work at all, it needs to work for everyone and in the same way.
