The right AI tools: One home care organization is paving the way

Artificial intelligence (AI) is rapidly becoming a foundational capability across healthcare. As workforce shortages, increasing service demands, and financial pressures continue to challenge providers, organizations must move beyond experimentation and identify practical opportunities to deploy AI in ways that improve care delivery, enhance workforce experience, and strengthen operational performance.

The rapidly evolving AI technology space offers healthcare organizations myriad options to address their AI needs, but organizations often struggle to select the right tool for the job.

As with other digital technologies, organizations are confronted with the “build” or “buy” conundrum. General-purpose AI models are capable of performing a wide range of tasks, providing flexibility across many business use cases.

However, for highly specialized workflows or domain-specific data, these models may not consistently deliver the level of accuracy, context or performance required. In these situations, organizations may need to customize, fine-tune, or develop AI solutions that are better aligned to their unique requirements.

To choose the most appropriate AI tool, there are three different approaches organizations can take: purchase an “out of the box” existing program, partner with vendors to trial software as part of a selection process, or build a solution in-house.

At VHA Home HealthCare (VHA), a trusted not-for-profit provider of high-quality, compassionate home care in Ontario for more than a century, we view artificial intelligence as a strategic opportunity to improve healthcare delivery and organizational effectiveness.

Rather than taking a wait-and-see approach, VHA has established a focused program to identify, evaluate and rapidly implement AI solutions that can deliver meaningful value for our clients, families, care providers and employees.

While we are moving with urgency to realize the benefits of these emerging technologies, we are doing so within a robust governance framework that emphasizes privacy, security, transparency, fairness and accountability.

This balanced approach enables VHA to innovate confidently, scale successful solutions, and ensure AI is adopted responsibly, in a manner consistent with our purpose and values.

Here are three of VHA’s AI initiatives, each of which has been selected using a different approach.

• Adopting an “out-of-the-box” AI model to better review and evaluate employment candidates: Like many recruitment departments, VHA’s team often receives a high volume of applications for certain positions. In some cases, more than 1,000 applications are submitted for a single job posting, many from candidates who do not meet the required qualifications. Reviewing applications, assessing whether candidates meet the requirements of the role and scheduling interviews, can be a time-consuming process, which may delay hiring for critical positions.

To improve the efficiency of this work, VHA’s HR team implemented a conversational AI platform to help match applicant qualifications with job requirements and automate parts of the screening process.

The selected solution leverages Paradox, a Workday-owned platform that integrates directly with the human resources and finance systems VHA already uses. This integration allows VHA to streamline recruitment activities while maintaining a consistent and secure user experience.

One of the most significant risks associated with AI tools used in recruitment is the potential for bias in algorithmic outputs. As a result, a key consideration in the product selection process was ensuring that the platform complied with VHA’s Artificial Intelligence Policy, including requirements related to bias management, monitoring and reporting.

“What excites us most about our Paradox implementation is its ability to connect qualified candidates with opportunities faster and more efficiently,” said Ernesto Sequera, VHA’s director, workforce strategy, safety & labour relations.

• Piloting and evaluating three AI scribe options: There is growing interest in AI scribes in healthcare as clinicians look for ways to reduce time spent on administrative tasks – like documentation – so they can spend more time providing direct client care.

However, limited data is available on the effectiveness of AI scribe technology within a homecare practice environment, so VHA is conducting a structured evaluation of three leading solutions to identify the platform most capable of supporting safe and effective implementation at scale.

This carefully monitored, real-world pilot provides an opportunity both to understand how well these tools work for homecare providers and to check for concerns that have been raised in other practice contexts, including monitoring and auditing for potential inaccuracies in AI-created notes.

Built into VHA’s pilot are multiple proactive controls, including a required review and authorization of the draft note by the clinician prior to submission into the electronic medical record. This process also addresses a new clinical competency for team members, building confidence in diligently reviewing AI.

“By piloting a few different solutions, we’re able to compare how they perform in our unique home care environment, get feedback from our point-of-care providers and make a confident decision before moving forward with one platform,” commented Jordan D’Souza, head of innovation at VHA.

• Building a unique solution for better charting and documentation: The review of nursing charts by clinical leaders is standard practice in healthcare organizations to ensure quality of care and patient safety. It can be a labour-intensive process, but it is crucial for driving improvements to documentation quality. Traditionally, this repetitive process has been performed manually, allowing only a small sample of available charts to be reviewed.

In the absence of existing solutions in the market, VHA partnered with a developer to build a customized AI-based nursing documentation review program.

The development process to create a chart review system that would reflect VHA’s standards was driven by internal guidance and oversight from nurses and clinical leaders.

The resulting software, ChartCoach, enables VHA to move from auditing a sampling to reviewing all nursing charts in a fraction of the time it took previously.

This reduces costs significantly and allows prompt feedback to be provided to our nurses, supporting better documentation and client care.

Adopting AI responsibly through effective oversight: The three initiatives highlighted in this article demonstrate that successful AI adoption is about selecting the right solution for the right problem. In each case, VHA identified a specific operational challenge and chose the most appropriate path forward, either implementing an existing solution, collaborating with a vendor to tailor a product, or customizing a tool to address unique organizational needs.

These experiences reinforced an important lesson, that there is no single blueprint for AI adoption.

As organizations move to incorporate AI into their operations, strong governance is essential. At VHA, oversight is provided through three complementary bodies: an AI Governance Committee, Ethics Committee, and Data Governance Committee.

Together, these groups help ensure that AI solutions and the data behind them are evaluated and implemented in a manner that is responsible, secure, transparent, and aligned with our organizational values.

Alistair Forsyth is VP, digital health and chief information officer at VHA Home HealthCare. Sandra McKay, PhD is VP, research & innovation and chief scientific officer at VHA Home HealthCare.