Total 587 words used.
Background
Emirates Health Services (EHS) through its advanced governmental health care system consisting of 17 hospitals and 100+ primary healthcare center caters to healthcare needs of patients with over 7 million Outpatient department(OPD) visits each year. EHS .is committed to instilling the principles and concepts of innovation practices in accordance with the best international standards at all strategic, organizational, and operational levels, to achieve the real and qualitative advancement of services provided to the patient community taking advantage of advanced technology.
As one of the largest providers of OPD services in UAE, every patient failing to attend the clinic appointments(no-show) significantly affect delivery, cost of care and resource planning. Globally the patient No Shows is estimated to cost the healthcare industry $150 billion annually. Patients who fail to show up in the OPD often require more expensive emergency and tertiary care later thus increasing the cost burden of the healthcare system.
It is therefore becoming not only a sustainability challenge but could also affect the health outcomes of the patient. EHS aimed to evaluate prevalence, predictors and economical consequences of patient no-shows and designed an innovative approach leveraging Artificial Intelligence and Machine Learning(AI& ML) to reduce no-shows to provide sustainable, integrated, accessible, efficient, innovative and high-quality healthcare services
Leveraging Artificial Intelligence for Predicting No-Shows
EHS consistently strives to use Artificial Intelligence for critical decision making using its analytics platform, ‘Manara’. The analytics platform of Emirates Health Services was designed to be the catalyst for the data driven healthcare transformation. It is a centralized platform that helps end users to access different available dashboards & insights created by Emirates Health Services.
This technical project was conceived as a practical and applied Artificial Intelligence Driven efficiency improvement Program using its analytics platform – Manara that would provide augmented intelligence to help OPD administrators identify the patients at risk of missing their scheduled clinic appointments and improve physician productivity and operational efficiency.
The tool was developed through iterative model development methods using both traditional machine learning techniques and as well as more advanced classification and regression trees for ML(CART) models to predict the patient no-shows.
Methodology & Design
A complete clinic appointment dataset was built using over 28 features which included patient demographics, appointment details, patient behavior and clinic details. Through the application of data science concepts, a retrospective cohort study was conducted to understand patterns, distribution and correlation.Exploratory Data Analysis was conducted to help understand data topology and certain features of importance.We used this understanding to support the design of a bespoke machine learning model that will accurately predict the future likelihood of no-shows on real time data
Outcomes
The AI model built had an Accuracy of around 80% when validated. The results from the model were then converted to insights that could guide the administrators and leadership make informed decisions and device strategy to follow up patients by calling/sending text reminders. The insights have several slicers and dicers to determine those patients who are most likely to miss their appointments.
The model identified the top 3 factors for patient missing their appointments as service category(specialty), E-clinic and the cumulative no-shows of a patient in the last three years. This was a crucial information that gave insights that some of the specialties like orthopedics, ophthalmology had higher likelihoods of no-shows and it helped the leadership analyze the reasons such as longer waiting days to get an appointment, long wait times etc. EHS were able to put forth a individualized strategy to tackle the no-shows in these departments
