Being bogged down in administrative tasks is tedious at best, but at worst, it can be a major contributing factor to employee burnout. We’re seeing this in the healthcare industry today. NPR shares that nearly half of all health workers want to quit their jobs due to the stressors piled on by administrative tasks, navigating health records and dealing with insurance companies, all when they could be caring for patients. This extends beyond clinical care to healthcare facility employees, administrators and others.
As massive operations employing hundreds of employees, hospitals can often see inefficient administrative tasks snowball across a facility and back up everything from HR to patient discharge procedures. This is where we’ve seen ERP with integrated generative AI solutions lift some of this major burden from healthcare workers’ shoulders and ultimately provide better care for patients.
What are LLMs and what can they do for healthcare?
We’re all familiar with GenAI these days, thanks to open-source tools like ChatGPT. AI-driven tools that generate human-like content and assist with tasks like writing and conversation, code generation, output validation and information retrieval are possible thanks to large language models (LLMs). Understanding and derive meaning from language, LLMs can equate to all of the above applications and more enhancing a user’s experience through a more natural interaction with technology – with no coding or IT training necessary.
When integrated into ERP systems, LLMs can identify patterns in even the largest volumes of patient data, whether structured or unstructured, and – for healthcare specifically – can streamline a number of processes like processing payments, transcribing patient visits, providing follow-up services communications to patients and sharing healthcare plan options that are tailored to one specific patient.
Clinical use cases
LLMs can bring new insight into the proprietary patient and organizational data without compromising the security and integrity of the highly secure private data, which means LLMs give providers the ability to quickly understand an individual patient’s healthcare needs, past procedures and more.
For example, discharging a patient from the hospital often includes a printed set of follow-up care instructions that says what patients should do following this procedure. However, it does not account for that individual patient’s experience, healthcare needs, occupation, etc. With LLMs, post-care instructions can be tailored specifically to a patient’s unique procedure, doctor and medication. This can be followed by personalized messages from the care provider to see how the patient is feeling and what questions they may have. Those answers then determine the next messages you receive for continued care.
Taking care of facilities
Hospitals have massive HVAC systems that are critical to patient comfort and care. They also have many moving parts all needing consistent maintenance. While many facilities teams likely use dashboards to track the regular maintenance of these systems as well as warranties, repairs, compliance checks and so on, LLMs allow facilities teams to cross-reference these with ease. This can include forecasting weather variations over the next ten years combined with past maintenance schedules to determine when the system may need replacement parts. Team members can also quickly ask a question such as, “When was the last time the south AC unit had maintenance?” and receive the answer back in moments without having to dig through records.
Maintaining best practices, mitigating concerns
Healthcare has strict data-sharing and privacy requirements. From concerns about data-sharing under the Health Insurance Portability and Accountability Act (HIPAA) to cross-network proprietary information, healthcare providers can be understandably hesitant to mingle data with outside sources. The great news is that LLMs can be securely integrated into an organization’s data, meaning the data the LLM is trained on is the organization’s alone.
To best mitigate security concerns, it is important to work with experts in GenAI and LLMs. They are well-trained in how GenAI works and can help mitigate compliance and security risks to ensure the system is leveraging proprietary data carefully. They can also help a system, if needed, pull appropriate outside sources without mingling confidential data.
To elicit the best insights from LLMs, it’s important to first remember that an internal AI solution is not the same as a Google search. Users can input short keyword searches, such as “Patient Smith date of birth, last check-up.” However, it is much more useful to ask a question in humanlike language because the LLMs will give you a humanlike response: “What is Patient Smith’s date of birth and when was their last checkup?” Now, instead of just getting a couple of dates, the LLM can extract those as well as additional information about Patient Smith: next checkup, recent lab results, potential issues to keep an eye out for, and more. Then you can go one step deeper and ask for more specifics: “What was their last checkup on and why?”
There are nearly infinite numbers of questions healthcare workers can ask LLMs, but the first ones that need to be considered are how, when and why LLMs should be incorporated into healthcare in the first place:
What value will this add to my practice? What training will need to be implemented to ensure proper use? What new insights can we gain?
But perhaps, when considering this technology, an important question may be “What is the risk if I don’t implement an LLM?”
Few technologies have such a broad application across an entire organization. In healthcare, the tools that hospital maintenance professionals use are unlikely to coincide with the systems doctors and nurses use. The cross-functional applications of LLMs mean they are not limited to a feature, tool or group of users. LLMs are a strategic initiative that, when implemented across an organization, will help provide insight into all areas of a hospital across all departments and across an entire healthcare system.
The risk of not implementing LLMs may not be as stark as losing patients to competitors, but the ongoing insights they can provide could be the difference between employee burnout and employee empowerment.
