Inspiration
The healthcare environment—with its packed work days, demanding pace, time pressures, and emotional intensity—can put physicians and other clinicians at high risk for burnout. Burnout is a long-term stress reaction marked by emotional exhaustion, depersonalization, and a lack of sense of personal accomplishment. Burnout takes a toll on physicians, their patients, and their practices. Short visits, complicated patients, lack of control, electronic health record stress, and poor work-home balance can lead to physicians leaving practices they once loved, poor patient outcomes, and shortages in primary care physicians
What it does
The solution we realized is a form, that predicts the level of burnout of the doctors. They can fill out the form to get recommendations about their work life and how to prevent burnout on their level of job satisfaction. The goal of our project was to devise a solution that reduces the time-and-effort burden for providers or healthcare staff using data analysis techniques. In other words, we needed to identify the factors (socio-demographic or work environment-related) that positively influence a medical worker's burnout score and to propose a solution that would help to prevent burnout. To solve this problem we created an ML pipeline, starting from data preparation and processing: coding variables and creating new ones based on several judgments (job satisfaction and our dependent variable - burnout score). Then we applied a linear regression algorithm to find the weights or coefficients of our factors.
How we built it
We used the dataset of Jordan doctors, who took the survey during COVID about their work, demographic and geographical conditions. On this basis, we have built a linear regression model, which weighs the coefficient of every factor into burnout. Next, we have built the form to help doctors self-evaluate their condition.
Challenges we ran into
The main challenge is the creation of a convenient way to fill out the form for doctors. They are busy people, who do not have time for new formats and long questionnaires, so we developed a small form with 20 quick questions, that going to predict the burnout score.
What's next for Predictors of burnout among doctors
We going to build a website for doctors with our calculator and working materials. It may be the basis for further data analysis of burnout.
What we learned
As factors positively influencing a burnout score we discovered: single (marital status), female (gender) and a positive COVID status. The more mentally well-off workers work with universities, while the greatest burden is on those who work with the military, police, etc. (Jordanian Royal service). Oddly enough, the size of the salary does not allow to cope with burnout and does not affect its degree: for this reason, medical workers overwork and take night shifts, which are paid twice as much. This turns out to be a mistake: the highest weights are found in a 48+ hours/week work schedule, a large number of night shifts (3+/week), and full-time work. Among other things, the lack of COVID incentives and insufficient COVID PPE affect our dependent variable positively. All in all, we're going to build a website for medical workers with our calculator to help them understand which factors affect their mental state most and we want to give some recommendations to hospital administrations to create a better working environment for medical workers.
Log in or sign up for Devpost to join the conversation.