Post Top Ad

Post Top Ad

news

As emergency department admissions rise again, how can AI tackle crowding once and for all?

[ad_1]

If you visited an emergency department (ED) prior to the pandemic, you most likely don’t need to be told that ED crowding is a problem. It is visibly embodied by queues of ambulances outside hospitals and bed trolleys in corridors.

During the pandemic, efforts to keep patients out of hospital wherever possible lowered ED admission rates, but as the world returns to normal, and admission rates rise again, so too does this internationally recognized issue which runs deep into the heart of hospitals, putting patient lives at elevated risk.

According to the CDC, approximately 50% of EDs experience overcrowding, and 90% of ED directors report overcrowding as a recurrent problem. In England, it is clear that supply is also struggling to meet demand for emergency medicine as in 2018-19 only 88% of ED patients were treated within 4 hours, compared to 98% in 2011-12 and a predicted 79% in 2025-26.

And it has indeed been proven to impact levels of care.

In terms of quality of care delivery, ED crowding is proven to be associated with higher workload, higher cost of treatment, delayed patient assessment, more frequent discharging of patients with high-risk clinical features, poor infection prevention and control measures, and lower patient satisfaction.

And this of course translates into lower patient outcomes, specifically in the form of high patient readmission rates, increased walkouts, prolonged hospitalization, a higher frequency of medication errors and adverse events, along with both increased morbidity and mortality.

Economically, the effects of extended lengths of stay and inefficient operations are also severe. One study found that in one 627-bed New York hospital, extended ED length of stay equaled almost $9 million in increased excess charges. In the UK, emergency admissions cost the NHS £17 billion in 2016-17 alone.

So it’s obvious that ED crowding is an issue that requires urgent attention if we are to secure the sustainability of our healthcare systems and empower healthcare professionals worldwide with what they need to deliver top quality care.

Enter: AI

Over the years, there have been attempts to mitigate and resolve this issue.

Tools which provide superficial measurements of ED crowding to aid decision making such as the NEDOCS and ICMED (International Crowding Measure in Emergency Department) scores, for instance, are available to emergency medicine leaders, though their limitations and shortcomings are widely agreed upon.

Initiatives to increase access to primary care and general practitioners have been launched with varying success. Alternative models of care, such as the Discharge to Medical Home model, have also aimed to reduce the number of low-acuity patients entering the ED. However, the Royal College of Emergency Medicine’s stance remains that, in the U.K. at least, the proportion of low-acuity patients who could be treated in alternative healthcare settings is no more than 15%, which suggests that the effect of these solutions could be limited.

So why, then, is such a long-standing, well-recognized problem still plaguing our EDs? Well, because crowding is an extremely complex, multivariable phenomenon with causes unique to each and every ED, interconnected with wider hospital operations.

And it is this complexity which means that AI and machine learning is perfectly positioned to make a global impact. Algorithms can nowadays ingest inconceivable volumes of historical and real-time data from EDs to not only make predictions personalized based on the unique features of each individual ED, but also visualize the results of theoretical interventions on current and future crowding pinch points.

What is most exciting is the advent of truly advanced digital health technologies which promise to reveal extremely rich, previously inaccessible information on the real-time physiological health (vital signs etc.) of patients. There is thus the opportunity to inform ML-driven crowding models with this kind of data also, which could transform our understanding of what constitutes quality hospital care management for good.

ED crowding is therefore an issue in the healthcare landscape where desperate need for change is meeting huge potential for innovation, and an international consortium should be formed to capitalize on this convergence.

Photo: pablohart, Getty Images

 


Source link
[ad_2]

source https://earn8online.com/index.php/310771/as-emergency-department-admissions-rise-again-how-can-ai-tackle-crowding-once-and-for-all/

Related Posts

No comments:

Post a Comment

Post Bottom Ad