Professional Documents
Culture Documents
outpatient visits
Introduction:
Healthcare systems consumes a significant part of the U.S. Gross Domestic Product (GDP).
To confront rising costs, limited resource capacities, and burgeoning demands facing U.S.
healthcare providers, efficiently using clinical resources is critical. A major obstacle in cost-
effective health care delivery and patient safety is no-show. Patient no-show is defined as a
patient who does not appear for the scheduled appointment.
This is a recurring problem in healthcare that a high percentage of patients who miss their
appointment, be it a consultation or a hospital test. The hospital outpatient non-attendance
imposes huge financial burden on hospitals every year. In USA, it is reported that an average
no-show rate of 62 appointments per day and an estimated annual cost of $3 million in a
community hospital setting.
Consequences of No Show:
1. Non attendance at clinic appointments (no-show) significantly affects delivery, cost of
care and resource planning
2. Huge financial burden on hospitals every year. Almost $156,340.80 per provider, per
year
3. Manually following up with every single patient who no-shows, you are looking at
about 31.9 minutes per provider, per day
4. Longer wait times at clinics
5. Dissatisfaction among the patients
Model Selection:
Since we want probability score i.e. a score associated with each member to being a No Show
or Late Cancellation, we can use Logistic Regression model for this analysis and more
importantly LASSO Logistic Regression to avoid over fitting while training.
Model Evaluation:
We can use different techniques to evaluate our model like concordance, KS test and Confusion
matrix. We should also look at the ROC AUC curve after training the model.
Conclusion:
Predicting No show or late cancellation can help us re-schedule the appointments thus better
resource management and patient satisfaction.
Patients with a flag of no show can be given alternate appointment schedule or an option for
virtual appointment.