RwHealth founder and CEO Orlando Agrippa explains how AI innovation in emergency admissions from asthma and COPD will be essential this winter.
As the nation holds its breath in the face of rising hospitalisations and fresh talk of a potential last-resort firebreak lockdown before the end of the year, protecting the wellbeing of some of our most high-risk patients is something we will need to prioritise with extreme urgency. According to the latest ONS figures, hospitalisations have soared from 1.12 to 7.40 per 100,000 since the start of June 2021, amid growing concerns around how the NHS will cope under “unsustainable pressure” and the potential force of a tougher flu season.
With the devastating impact COVID-19 has upon the respiratory system, it is of little wonder that the delivery of care and services has changed so drastically over the past 18-months within related specialties. As the RwHealth team works to support our colleagues on the frontline of healthcare, the impact the pandemic has had upon respiratory care pathways is something that has been raised again and again by NHS leaders across the country.
Growing evidence suggests that respiratory conditions traditionally known to see high susceptibility from respiratory infections, such as Chronic Obstructive Pulmonary Disease (COPD), could see similarly heightened risks of hospitalisation (~3.8x), ICU stay (~1.3x), and mortality (~2.2x) in the context of COVID-19 – patients defined as clinically extremely vulnerable and added to the Shielded Patient List early on in the pandemic.
Likely at least in part due to the resulting need for patients to avoid potential points of infection, the post-pandemic period has seen what some may otherwise consider a counterintuitive drop in COPD admissions. With potential compounding by changes to clinical coding – patients admitted for COVID-19 with underlying COPD and related symptoms not seeing these recorded as the primary diagnosis – it will nonetheless be incredibly important to track potential changes over the winter months amid a predicted severe flu season and the potential downstream impacts of delayed care for COPD exacerbations on patient wellbeing.
Combatting the associated risks for patients with COPD and those with related conditions requires significant proactivity in ensuring those strategising and providing care within respiratory services are well armed with the powerful data insights, advanced predictive technology, and sharp solutions needed to drive up-to-date decision-making under intense stress.
Applying the powerful capabilities of data science to interrogating curated datasets is a crucial component of undertaking this challenge – monitoring healthcare trends, exploring the underlying patient and provider-level complexity, and modelling what the coming months in respiratory care may look like. Combining this with insights around successful restructuring of such pathways in partner trusts that have already made important steps to tackling the pandemic’s impact, this utilises an approach that applies important turnaround work to other providers and services, to ensure delivery of care that protects patient outcomes and ongoing organisational efficiency in the face of extreme strain.
COVID-19’s impact on patients | Growing complexity, and the associated risks
Despite falling numbers of admissions with a primary diagnosis of COPD, we have seen a growing degree of complexity in those admitted that stress the potential for associated risks. Across both male and female patients, the post-COVID environment has seen an average 28% increase in the depth of diagnosis, with the incidence of those linked to viral diseases rising to within the top 10 most common secondary diagnoses.
Indeed, analysis driven by the DSP has found that this growth in complexity appears to accompany a rise in mortality rates in comparison with pre-pandemic levels, with rates peaking in April 2020 and January 2021 at 5.6% – in-line with the peaks of COVID-19 waves. Whether this reflects a higher rate due to less severe patients avoiding contact with healthcare environments and the overall drop in admissions or are a true reflection of heightened risk post-COVID, will therefore be an important avenue of exploration in the fight to protect patients from poor outcomes.
Trust responses in a post-COVID landscape
Provider benchmarking is a longstanding element not only at RwHealth, but within healthcare itself – creating integral comparisons of performance, outcomes, and clinical variation across trusts in order to drive the subsequent targeting and reduction of the disparities that in turn translate to poorer outcomes for patients. Driven by the DSP’s advanced Artificial Intelligence/Machine Learning technology, we have applied the fundamentals of benchmarking to exploration of the data and identification of key trends and outliers in regards to COPD care across the NHS.
Despite increasing patient complexity, national averages have seen an overall drop in length of stay from five days pre-COVID to 3.9 days in the months since the first lockdown. With many providers fighting to make space for surging influxes of patients and in light of the risk of COVID-19 infection in healthcare settings, it is likely that this may reflect a prioritisation for expedited discharge processes for patients for whom it is clinically appropriate.
While this has promisingly overall not correlated with a subsequent increase in the average number of COPD-related readmissions in comparison to pre-pandemic levels – instead falling by ~40% from March 2020 onwards – this has certainly not been true of all trusts across England. Indeed, of the 164 trusts analysed, only four recorded an overall percentage increase in the number of COPD readmissions since the pandemic’s onset, with a number of providers at the higher end of the length of stay spectrum seeing higher mortality rates in turn.
For such providers, it will be vital to better understand the underlying patient profiles, patient flow pathways, and current procedures at a local level, in order to pinpoint potential strategies to tackle these associated risks to patients – deriving best practices from other organisations in order to improve outcomes for patients with COPD.
Predictive insights for evolving COPD demand
A critical part of the DSP’s modelling capabilities in supporting NHS organisations to manage dynamic trends in demand and meet this gap with available bed, resource, and workforce capacity, is powered by in-depth predictions built from historical insights and current global, national, and local developments in underlying factors.
With predictions generated at both a national and provider level across multiple scenarios, our predictions highlight a potential surge in monthly admissions of up to 239% by 2023 at some of the largest COPD treatment providers in the country. With ever growing concern around the long-term impacts of COVID-19 upon chronic conditions such as COPD, ensuring that trusts are equipped to provide capacity and standards of care for these increasing demands, set only to arise over the coming months, is essential.
The DSP’s ability to generate these deep clinical variation, provider level, and patient level insights is one that has been truly instrumental in our ongoing work to support healthcare partners in light of the challenges brought by COVID-19. Supporting planning with advanced modelling capabilities and innovative predictive technology is something that we truly believe will be a vital part of strengthening best practices for key at-risk specialties and patients – both as the pandemic continues and as we one day return to a world closer to normal.
You can download the complete report here for a deeper dive into key findings in COPD care.
