Healthcare providers are changing their fundamental business processes in various ways with the advent of data, analytics, and AI. Some changes are because of regulatory pressures such as readmission penalties. Others stem from challenges such as staff shortages, resurging patient volumes, cost pressures and growing care-delivery models.
Artificial Intelligence (AI) solutions have been proven to help alleviate cost pressures, plan for different scenarios, and dynamically manage staffing requirements to achieve the appropriate staff-to-population ratios.
At OmniData, we leverage our extensive domain knowledge and capabilities in deploying AI solutions to aid healthcare providers to unlock business value.
The Problem
The average hospital readmission rate in the US is 14.5%, with rates ranging between 11.2% and 22.3%. In 2022 Medicare assessed hospitals and found that 42% were penalized for readmission rates exceeding 30-day risk-standardized readmission rates. Preventable rehospitalization costs Medicare hundreds of millions of dollars each year and can be avoided through better care.
Imbalance in the healthcare staff supply and demand will have negative effects on the quality of care provided. Oversupply of healthcare staff also leads to an increase in cost to serve, while an insufficient supply will lead to deteriorating care.
The business problem can therefore be summarized as follows:
- Can the demand for different healthcare staff members be forecasted accurately to address the dynamic needs of patients?
- Can patient readmission risk levels be assessed and managed accordingly?
- What is the optimal mix of healthcare staff needed at any point in time to address the demand at the optimal cost.
- Can a balance be struck between hospitalizations and longer-term care needs for special care institutions.
The Solution
OmniData’s approach to the business problem would entail combining data sources, healthcare domain experts and the use of AI and optimization algorithms to identify and forecasts the drivers of readmissions and staff-to-population dynamics.
Health Human Resource Planning
Efficient health staff-compliment planning is critical to guarantee adequate number of staff available. Choosing the most appropriate forecasting technique will involve selecting amongst various algorithms. Feature engineering will be used to create context from the data such as demographic, diagnosis, procedure types, indicators of recency and frequency of medical events per patient, amongst others. These features will be passed through a modelling framework to forecast the demand of healthcare staff members. An optimization approach will be applied to these forecasts by selection of the optimization criteria like cost and quality of service.
Readmission Risk Assessment
A similar approach to feature engineering will be followed to predict readmission risk including demographics, number of ER visits, number of procedures and diagnosis. An AI model will be applied to classify patients to be readmitted or not. A score will be provided as part of the solution to indicate the probability of a patient being readmitted, from which a treatment strategy will be determined. The treatment strategy will be based on different levels of risk of readmission for example High/Medium/Low risk with associated treatment strategies like a nurse having in in depth discussion with high-risk patients on the necessary steps to be taken post discharge vs an email explaining the steps for a low-risk patient.
Scenario Planning
Scenarios provide views on the possible forecasted impact on an organization given certain assumptions. OmniData follows an approach where the forecasting models provide the base-case scenario, but then also enables clients to adjust certain factors to see the possible impact on planning.
Conclusion
The effective management of healthcare staff supply is a key business activity for companies in the healthcare sector. Balancing the supply with the demand and finding the optimal staff-to-population mix provides opportunities for value unlock. The value of having a robust solution to classify patients at risk of readmission is visible in the amount of dollars paid in penalties. Some studies show that upwards of $560M have been paid in penalties under the HRRP. OmniData helps clients in the healthcare industry by applying AI methodologies on client data and providing solutions to aid in the decisioning process for healthcare providers to identify patients and provide relevant actions.
Next Steps
To explore how OmniData can optimize your generative AI capabilities, schedule a complimentary AI in Action workshop to gain insights into leveraging Azure AI services tailored to your needs.
Our process includes:
- Environment assessment.
- Reference architecture provision,
- Roadmap creation.
Book your workshop to initiate your AI transformation journey.
