Two years ago, King Abdulaziz Medical City, a Riyadh-based hospital of the Kingdom of Saudi Arabia's National Guard Ministry of Health, became the first in the world to reach Stage 7 with four different HIMSS models. (The company recently pioneered the impressive effort to reach Stage 6 with another model.) The company's advanced use of medical information and technology has benefited the health system's 1.3 million patients. is bringing about.
Since then, the 3,720-bed MNGHA has focused on digital health across a variety of specific use cases, including an ostensibly simple use case that has long plagued healthcare provider organizations around the world: no-shows in ambulatory care settings. We have continued our reform efforts.
These can be disruptive, add unnecessary costs to the care delivery process, and have a real impact on care management and patient outcomes.
However, according to Huda Al Ghamdi, director of data and business intelligence management at MNGHA, the National Guard Ministry of Health has achieved remarkable results in terms of reducing no-shows by applying artificial intelligence to analysis, and AI The company uses this to predict no-shows in advance. It may be in the ambulatory care setting that patients are most likely to miss an appointment.
Health systems are using machine learning to take data from electronic health records (patient summaries, clinical information, appointment history) and process and train it for AI models that can alert physicians within the EHR. This allows doctors to send necessary reminders to patients and even book appointments within their own workflows.
MNGHA is comprised of more than 30 hospitals, specialty hospitals and primary care centers across Saudi Arabia, all linked to an integrated EHR system called BESTCare.
This gives “the advantage of having a huge amount of data,” Al Ghamdi explained. “Advanced Analytics, Prediction, and Machine Learning”
Innovative approaches to analysis are helping health systems in many areas, but no-shows are an area of particular concern, she said.
“The reason we address this issue in particular is because the outpatient setting is considered the largest channel through which MNGHA provides health care services to patients,” she said. “Unlike inpatient or ER, outpatient is considered the largest, as it performs approximately 20,000 visits per day. [on] average. “
This adds up to 5-6 million visitors per year.
“So when issues like no-shows occur, it definitely impacts healthcare providers, resources and the patients themselves,” Al Ghamdi said.
She points out that the fact that MNGHA is a public hospital means it can be difficult to calculate the cost if a patient doesn't show up for their appointment, but it does cost money. “And we need to recognize that and start thinking about savings.” ”
Fortunately, MNGHA “has a huge amount of data that we can start analyzing and studying and try to figure out the factors that influence this,” Al Ghamdi said. “We have an integrated electronic medical record system with different modules for registration, admission and outpatient care.
“In terms of the datasets that we are utilizing in this project, in addition to information related to the clinic itself, it is mainly demographic information, mainly very simple information such as gender, age, etc. from one clinic to another. ” she explained. “The third part of the data set is the patient's own medical history. We find that some patients, like all patients, have a high rate of missed appointments, so this type of The medical history gives us insight.''For those types of patients. ”
Importantly, she added, “we did not address any kind of clinical data” in this project. This requires expert clinicians to determine what types of clinical factors may influence no-shows.
However, by using a basic dataset of patient information, we were able to create several initial models and validate which were the best and most accurate.
“This project started two years ago and we have to go through stages to make sure we are ready. [incorporate the model] “So in the first year, the model was created and now we are in the stage of validating the model. This validation stage will take about four to six months,” Al Ghamdi said.
“Some of the validation will be done within data science, and then we will start validating it with small groups of clinicians and nurses and patient services staff,” she added. “That phase took about six more months. At that point, it took her a year to verify that the model was reliable and to see if we could really trust the results of that model.” I did.”
Once data science experts were satisfied with the algorithm, MNGHA took steps to incorporate the model into its EHR system and integrate it into clinical workflows.
“Clinicians can know that a patient who was scheduled for that day may not be able to attend the appointment. And by setting this type of flag within the medical record system, clinicians can You can send a reminder or ask the patient, for example. It's a service that makes a kind of phone call to remind the patient,” Al Ghamdi said.
Eventually, we plan to implement this model at all MNGHA facilities in all regions.
For healthcare organizations looking to try something similar in their own organizations, Al Ghamdi has some advice.
“It's better to do this kind of implementation, even if you start with a small dataset, a small scoop of data or a small list of parameters, because the data can tell you a lot about the patient. Because we believe this is the type of hidden pattern that can be discovered using machine learning and artificial intelligence techniques.
“It's really important to take steps to process data and gain knowledge from it,” she says. “This is a very simple model to create, but it has a huge impact on your organization.”
A detailed case study on MNGHA's use of machine learning for predictive analytics can be found here.
Mike Miliard is the Editor-in-Chief of Healthcare IT News
Email the author: mike.miliard@himssmedia.com
Healthcare IT News is a HIMSS publication.
