Decision Intelligence for Healthcare

Decision Intelligence in Healthcare

Improving operational efficiency in healthcare through better decison making

Maintaining high quality of care under pressure

null

Healthcare organizations, operating in a dynamic environment are constantly challenged by sometimes conflcting objectives. Providing the highest level of care with limited resources, shortage of staff and an increased focus on patient experience. With an ageing population and increasing demand for care, pressure on hospital resources will only continue to increase.

In order to continue providing the same or higher quality of care and 'patient experience', hospitals will be obliged to be more efficient with staff and resources, as well as more focused on evidence-based decision support. 

Decision Intelligence is bringing physicians and AI together

A new augmented team to address hospital critical KPI's

While AI might increasingly outperform physicians in specific tasks, AI and clinical knowledge need to go hand in hand, as  both have their unique strengths and limitations. Getting this complementary relationship right is what will ultimately determine the adoption and impact of AI in healthcare. Therefore we need to combine data science and technical expertise with a strong understanding of the clinical context in which physicians operate as well as the cognitive and emotional demands of their daily reality.  


Decision Intelligence is engineering how decisions are made and how they impact the outcome to drive performance of healthcare service systems like Length of Stay, patient flow, capacity, (re-)admission, costs,...                             

                       

Decision Intelligence optimizes towards better and less “biased” decisions, capturing  expertise through interaction of decision support with clinicians at every stage of the development.


Process & 

Decision  modeling

Designing process decision maps, creating and validating decision logic.


Optimizing  

decision process

Detecting bias, decision noise,..

From "prediction" to "decision" models that can be implemented effectively

Feedback &

“Continuous Learning”

Data, models, results and feedback are  monitored for adjustments

Making changes and quick roll-out 

 Human-centered AI, providing actionable recommendations, driving operational excellence through the entire patient flow.

Length of Stay (LoS)

With “less-urgent visits”, taking up a large proportion of all visits, the risk of "overcrowding" increases in the emergency department. LoS is an important indicator for the quality of patient care. Understanding underlaying mechanism and providing recommendations to reduce LoS, will drive improved patient experience and care. Accurate prediction of LoS will help to plan and organize bed allocation with proper lead time.

Demand for Emergency Department (ED)

EDs have become increasingly congested due to increased demand, increased complexity of care and blocked access to ongoing care. Hospital services can be managed more efficiently if an accurate short term demand forecast for ED is available. Due to the stochastic nature of ED arrivals, measures like rolling average, are not sufficient and cannot be adapted to reflect the case mix of people in the ED at a given point. Informing operational management real-time of the emergency admissions will help for more efficient use of resources..

Hospital re-admission

Healthcare costs due to unplanned readmissions are high and negatively affect health and wellness of patients. Discovering patterns in patient data to understand which patients are most likely to be readmitted in the short term and identifying which decisions are related,  would improve both the quality of life for patients and the financial wellbeing of the hospital. For elderly patients,  a geriatric assessment and interdisciplinary care plan can be set up in case of a high risk score.

Transitioning patients

 The ineffectiveness of transitioning patients into inpatient units within the hospital in a timely manner, is another source for congestion affecting operations. "Boarding patients" e.g. , waiting to be transferred from ED to hospital, and the extra  workload they introduce are a major concern in emergency departments. An early detection of patients with increased risk of a longer stay (disposition decision) would help planning to be adjusted or "boarding time" reduced.

Reducing the Length of Stay (LoS) in the emergency department through Decision Intelligence