How AI Can Lighten the Documentation Load for Nurses

As the nursing profession grapples with heightened burnout, researchers are investigating how large language models might help simplify clinical documentation and care planning.

Nurses are vital in providing quality care and driving positive patient outcomes. However, shortages and burnout have placed immense pressure on this workforce.

According to the American Nurses Association (ANA), healthcare organizations must prioritize maintaining adequate staffing, fostering supportive work environments, and enacting policies that back the nursing profession to fully unlock nurses' potential.

The ANA further explains that the COVID-19 pandemic has intensified existing challenges for nurses, such as rising healthcare demands, lack of sufficient workforce support, and a widening gap between retirements and the entry of new nurses.

As demands on nurses continue to escalate, some experts argue that technologies, particularly AI, may ease certain burdens, including clinical documentation and administrative tasks.

In a recent study published in the *Journal of the American Medical Informatics Association*, Fabiana Dos Santos, Ph.D., MSN, RN, a postdoctoral research scientist at Columbia University School of Nursing, spearheaded an exploration into how a ChatGPT-driven prompting framework could assist in generating care plan suggestions for lung cancer patients.

In an interview with Healthtech Analytics, Santos elaborated on the promise — and potential risks — of AI chatbots in the nursing field.

Addressing Documentation Challenges in Nursing Care Plans

Crafting care plans is critical to ensuring that patients receive timely and appropriate care tailored to their specific needs. Nurses play an indispensable role in this process, yet they face numerous barriers when documenting care plans.

“Nurses are at the forefront of patient care, contributing vital clinical insights to electronic health records (EHRs),” said Santos. “However, many existing documentation systems are burdensome, consuming a significant portion of nurses' workdays. This can lead to cognitive strain, stress, frustration, and disruption of direct patient care.”

The American Association of Critical-Care Nurses (AACN) points out that documentation tasks occupy an average of 40% of a nurse’s shift. As documentation time increases, direct patient care time decreases, leading to potential declines in job satisfaction, heightened cognitive load, and greater burnout.

These issues also impact patient outcomes, increasing the risk of medical errors and hospital-acquired infections.

“The demands of patient care, combined with administrative responsibilities, make it difficult for nurses to dedicate sufficient time to individualized care plans,” Santos continued. “The complexity of many EHR systems only exacerbates this challenge, making it harder to fully capture all aspects of a patient’s condition, including their physical, psychological, and cultural needs.”

To address these challenges, Santos and her team evaluated whether ChatGPT could be leveraged to improve clinical documentation.

“These negative effects highlight the urgency of enhancing EHR systems,” she noted. “Properly designed AI tools could streamline the development of individualized care plans and reduce the documentation burden.”

The Opportunities and Challenges of AI in Nursing

Nurses rely on their expertise to design care plans that address complex patient needs, such as symptom management and comfort care. Santos emphasizes that generative AI (GenAI) could enhance this process by improving documentation workflows and alleviating time constraints, documentation inconsistencies, and redundancies.

“AI can rapidly process large data sets, generating care plans more efficiently than traditional methods,” Santos explained. “This could allow nurses to focus more on patient care rather than administrative duties.”

However, Santos stresses that rigorous validation of AI models is essential. Nurses’ clinical judgment and expertise remain crucial when evaluating AI-generated care plans.

“While AI technologies can improve documentation, the integration of nurses' critical thinking is vital to ensure that patient care remains accurate and comprehensive,” she said. “AI has the potential to alleviate workloads, but its outputs must be thoroughly reviewed.”

AI tools are also limited in their ability to grasp the nuances of patient care that nurses, through personal interactions, can provide. This includes understanding patients' cultural and spiritual needs.

Despite these limitations, large language models (LLMs) are gaining traction in healthcare, particularly in EHR workflows and nursing efficiency, and Santos’ research is a notable step in this direction.

The research team designed a method for structuring ChatGPT prompts to generate consistent, high-quality nursing care plans. By refining the Patient's Needs Framework through multiple iterations and diverse hypothetical cases, the team ensured that AI-generated plans aligned closely with standard nursing practices.

“Our study shows that ChatGPT can prioritize crucial aspects of care, such as oxygenation and infection prevention, while providing detailed explanations for interventions,” Santos said.

Looking Ahead: AI's Role in Nursing

Though the initial study focused on lung cancer, Santos emphasized that it’s just the start of exploring how LLMs can assist nurses.

“This research represents an important step in integrating AI into nursing documentation,” she stated. “Our next steps involve demonstrating the framework’s versatility by applying it to various clinical contexts.”

The team intends to expand their studies to other conditions, such as gestational diabetes and renal failure, testing the framework’s robustness across a wider range of patient populations.

While promising, Santos cautioned that AI tools will not replace the critical work nurses do. “AI can enhance care plans, but nurses must continue to apply their judgment to ensure the patient’s unique needs are met,” she said.

Santos and her team continue to refine their framework, working towards a future where AI complements — rather than replaces — the essential human elements of nursing.