When Data Meet Nursing: A Nurse-Informed Patient Acuity Model

Predicting proper nurse staffing levels in infusion centers is critical for safe, timely, and quality care. However, identifying staffing needs accurately in this setting is fraught with challenges, including a wide variation in treatment complexity, high patient throughput, and the absence of a universally accepted nurse-to-patient ratio.

Additionally, while acuity-based staffing is a well-known best practice, traditional methods to obtain and collate information for acuity-based decisions are heavily reliant on nursing teams and can be time-consuming and inconsistent. With the growing complexity and volume of treatments moving to outpatient oncology, there is a need for more streamlined and consistent methods to enable timely acuity-based staffing insights.

In light of this need, The US Oncology Network (The Network), an organization of community-based providers supported by McKesson, led a collaborative initiative between nurses and data specialists to modernize acuity-based staffing. The Patient Acuity Model (PAM) leverages predictive analytics and machine learning to automate acuity-based staffing insights, empowering faster, more consistent staffing decisions that truly reflect patient needs, laying the groundwork for safer, more effective care in the modern oncology landscape.

A Data-Driven Solution for Resource Allocation

The PAM was developed through close collaboration between nurses and data specialists, ensuring practical nursing insights were embedded in the design. A group of nursing experts generated a list of potential features believed to influence acuity, drawing on existing models, published literature, and their professional experience in infusion. Nurses from 9 care sites across 4 practices then documented acuity ratings for all patients in free-text clinical notes.

Using natural language processing, data scientists extracted these entries and applied a range of statistical analyses to test each feature against the documented acuity scores to determine which truly impacted acuity levels.

The final model prioritized features with predictive power for acuity, with the following factors emerging as the most influential:

1. Number of drugs
2. Routes of administration
3. Drug hypersensitivity risk
4. Drug emetic risk
5. Appointment duration
6. Number of antineoplastic drugs
7. Cancer diagnosis
8. Cycle 1, day 1
9. IV access difficulty
10. Performance status (eg, ECOG)
11. Depression screening result
12. Primary language
13. Mobility status

The PAM was then trained to use these features to predict future patient acuity ratings.

PAM Dashboard and Reports: Efficient, Consistent Staffing at a Glance

The PAM dashboard is a user-centered, interactive tool that provides nurse leaders with daily updated acuity predictions and historical staffing data. Designed for the needs of nursing and operational leaders, the dashboard offers smart visualizations and filters that allow users to quickly assess staffing needs for current and upcoming days, compare trends across multiple sites, and make efficient, consistent staffing decisions. The dashboard and accompanying downloadable reports include key patient information to streamline assignment creation and infusion room preparation, supporting proactive planning for safer, more effective care.

Real-World Successes

The PAM is used for 50 infusion centers across 5 practices in The Network. Users report a wealth of improvements across multiple aspects of infusion center operations, such as:

Greater efficiency

The PAM automates acuity predictions and data compilation for staffing insights, infusion room assignments, and preparation, shifting the burden from nursing teams.

In one practice, the impact of replacing manual processes previously completed by nurses and nurse leaders with the PAM’s automated predictions was estimated to save 10.4 minutes per patient. With an average of 2400 visits per week across the practice’s infusion centers, this equates to more than 400 hours that are now being refocused on other activities—including an initiative to improve proactive infusion room preparation and patient safety checks—further underscoring the positive effect of the PAM.

Another practice leveraged the PAM’s acuity predictions and patient-level reports to allow a licensed practical nurse to create balanced infusion room assignments, a role previously assigned to the registered nurse (RN) lead, saving 16 hours a week in highly skilled RN lead time.

The PAM’s consistent methodology for predictions and a user-centered dashboard allows leaders to instantly assess staffing needs across multiple sites, eliminating manual and inconsistent processes. For practice leaders and float pool managers allocating staff across multiple locations, these real-time insights have been described as “game-changing,” as they are crucial for resource allocation and safe, effective care delivery.

Consistent data-driven staffing decisions

Practices are actively using the PAM’s 7-day prospective staffing recommendations to support anticipatory planning, including float staff allocation, facilitating acuity-based staffing decisions that were either prohibitively time-consuming or impractical to achieve. Users also report that the enhanced consistency, objectivity, and transparency in the way decisions are made contribute not only to better outcomes but also to increased collaboration between sites, promoting equitable allocation of resources.

Historical data trends enabled by the PAM also help leaders understand baseline staffing needs and justify hiring additional nurses when appropriate. One practice leader was able to use these insights to justify the addition of 2 positions to support the growing volume and treatment complexity of the practice’s infusion center patient population.

Enhanced safety and quality of care

With today’s complex treatments in oncology infusion, a shift from volume-based to acuity-driven staffing is needed to identify the appropriate number and skill mix of staff required to support precise patient preparation, vigilant double-check processes, strict administration and monitoring protocols, and early escalation when necessary. This approach is recommended by the Oncology Nursing Society. The PAM has facilitated this shift, enabling staffing decisions and nursing assignments to be tailored to patient needs.

Improved oncology nursing experience

Oncology nurses are committed to safe, effective care, and leaders must ensure they have the right resources. PAM users report smoother workflows, optimized staffing, and more personalized patient care, enhancing professional satisfaction for nursing teams.

The Power of Collaboration

The PAM exemplifies how teamwork and data can solve some of today’s complex challenges in health care. The collaborative approach enabled the PAM to be effective from the beginning, as the project brought together different perspectives and expertise to meet a real-world need. Dedicated nurses were engaged from the outset, contributing practical insights that shaped the model. Today, the PAM puts actionable data at their fingertips, empowering them to make informed decisions and deliver high-quality care with confidence.

By leveraging data-driven tools such as the PAM, practices in The Network are advancing the mission of delivering safe, timely, and high-quality cancer care. While the PAM addresses a critical challenge in oncology care, there is a need for additional innovative, collaborative solutions in the future. As care becomes more complex, nurses must continue to work with experts across fields to develop smarter, safer, and more efficient ways to meet the growing needs in the oncology world.