Healthcare systems across Canada continue to face growing pressure from patient flow challenges, prolonged hospital stays, and increasing demand for acute care beds. A key contributor to these pressures is Alternate Level of Care (ALC), a designation used when a patient no longer requires the intensity of services provided in a hospital setting but continues to occupy a bed while waiting for care elsewhere. Across Ontario, more than 80 per cent of ALC designations involve adults aged 65 and older, highlighting the complex care and transition needs of an aging population. 

Provincially, there is a growing shift from simply managing ALC days to preventing ALC through evidence-based leading practices, outlined in Ontario’s ALC Leading Practices Guide. These strategies recognize that delayed transitions are not solely the result of limited community or long-term care capacity. Hospital care processes also play a critical role, as prolonged hospitalization can increase the risk of preventable complications and functional decline. Improving how and when discharge planning begins is therefore central to both patient outcomes and system performance. 

At Humber River Health (Humber), a new artificial intelligence (AI)-enabled initiative is taking a proactive approach to ALC by supporting earlier identification of patients who may experience complex transitions and enabling care teams to begin planning sooner. This work is about more than patient flow – it is about improving access to the right level of care, strengthening system capacity, and ensuring hospital resources are available for patients who need them most. 

Through investment from Scale AI, Humber is working with Deloitte, Niagara Health, Cambridge Memorial Hospital, Seafair and SE Health to enhance iPlan, Humber’s digital discharge planning platform, with advanced predictive capabilities. The goal is simple but innovative, identify patients at risk of becoming ALC earlier in their stay and support timelier and more focused coordinated discharge planning. 

The need for innovation in discharge planning is especially urgent in Northwestern Toronto; the community Humber serves. The region has the highest number of seniors over the age of 80 in Ontario, a figure expected to double by 2036. At the same time, 23 per cent of seniors live alone, many residents face complex socioeconomic challenges, and the area has long had one of the lowest rates of primary care attachment in the province. With 60 per cent of residents identifying as racialized, 54 per cent as immigrants, and higher-than-average rates of material deprivation, care transitions are often influenced by factors that extend well beyond a patient’s immediate medical needs. These realities highlight the importance of earlier, more coordinated planning that takes these aspects into consideration. 

The Pathway 

The enhanced iPlan system uses machine learning models to analyze clinical and demographic data in real time, beginning at inpatient admission. These models generate early risk signals that help care teams anticipate which patients may face complex transitions out of hospital, suggest most likely discharge destinations, and group patients with similar care needs to inform earlier planning discussions. Patients with higher risk profiles are flagged sooner, triggering earlier involvement from physicians, nurses, social workers, allied health professionals, and Ontario Health atHome partners. 

By shifting discharge planning further upstream, teams can begin parallel planning days earlier rather than waiting until the patient reaches medical stability. This improves alignment between patient needs and available services and helps prevent extended hospital stays that can impact both patients and the broader system. 

“Innovation in healthcare must deliver both efficiency and compassion,” says Beatrise Edelstein, Vice President, Post-Acute Care and Health Systems Partnerships at Humber River Health. “By embedding AI into iPlan, we are giving our teams the insight to anticipate patient needs and plan transitions with greater precision. This project will not only improve patient flow at Humber but also contribute to a more connected, efficient, and sustainable healthcare system across Canada.” 

At the heart of the initiative is a holistic approach to understanding patients beyond their immediate diagnosis. The predictive models consider a wide range of factors that can influence discharge readiness, including mobility status, previous admissions, social supports, and other social determinants of health. This broader lens enables care teams to identify potential barriers early and connect patients to the right services sooner, whether that means returning home with supports, transitioning to rehabilitation, or moving to another care setting. 

“This is about bridging the digital and clinical worlds in a responsible way,” says Gagan Grewal-Sagoo, Transformation Lead, Innovation Excellence Team at Humber River Health. “By bringing together clinical teams, operational leaders, and technical partners, we are building something that is clinically meaningful, operationally embedded, and structured for responsible, real-world use.”  

Development and Collaboration 

Collaboration has been essential to the project’s development. Representatives from across the patient journey, including allied health, social work, nursing, physicians, information systems, privacy and ethics teams, as well as executive leadership, have contributed to shaping how the tool works in real clinical environments. Health system partners beyond the hospital, including home and community care providers, have also been engaged to ensure the solution reflects realities across care settings. 

“Having multiple disciplines involved ensures the model reflects what actually happens on the unit and in the community,” says Nadine Sriskantharajah, Professional Practice Leader and Social Worker. “The goal is not just to predict risk, but to make sure the information leads to practical, timely actions that improve the experience and overall wellbeing of patients.” 

The project follows a structured three-phase approach: define the essential features and requirements, conduct clinical testing, refinement and integration within iPlan, and deploy the tool through a pilot at Humber by the end of 2026. Clinical feedback is being incorporated throughout this process to refine the models and workflows, ensuring they support, rather than disrupt, frontline care. 

Beyond Humber, the long-term vision is scalability. While the initial integration is within iPlan, the underlying approach is being designed so it can be adapted by other hospitals and health systems, even those using different digital platforms. By demonstrating how AI can be thoughtfully implemented into existing clinical workflows, the project offers a blueprint for responsible, practical AI adoption in healthcare. 

Looking Ahead 

By combining earlier risk identification with smarter, data-informed decision-making, care teams can plan discharges sooner and more effectively, reducing mismatches and delays and driving system-wide outcomes such as shorter patient stays, improved patient flow, and more efficient use of staff time. The impact, however, reaches far beyond hospital operations. This proactive approach contributes to a more patient-centred experience by looking beyond just patient diagnosis, while also freeing up beds for those who need acute care most. 

As Canadian hospitals continue to navigate complex system pressures, Humber River Health’s Scale AI – ALC project demonstrates how innovation, when grounded in collaboration and clinical insight, can help create a more responsive and sustainable healthcare system for the future.