Predicting Claims Denial
An automated machine learning (ML) system is built that enables healthcare providers to accurately
predict which claims will get denied on the basis of "medical necessity" reason by the payer even before the claim has been submitted. This ML system helps revenue cycle staff to focuse attention on claims those that have a strong likelihood of being overturned, and allow them to correct claims before submission to increase first-pass payment rates. The ML model built made use of claims 835, and 837 data. The developed XGBoost model showed sensitivity of 95% and specificity of 65% for the given data.
Input variables : Historic data from EDP (Member 360)+ Claims (Rejection status, Rejection reason, Dx, Rx, CPT, ICD codes)
Output Variables : Approve/deny
Metrics to Monitor
|
Statistical
|
:
|
Somers D |
Accuracy |
Precision and Recall |
Confusion Matrix |
F1 Score |
Roc and Auc |
Prevalence |
Detection Rate |
Balanced Accuracy |
Cohen's Kappa |
Concordance |
Gini Coefficent |
KS Statistic |
Youden's J Index
|
|
Business
|
:
|
Claims Processed |
$ Saved |
Denial Rate |
Top Denial Reasons |
Denials by CPT |
Denial by DRG |
Denial by Payer
|
|
Infrastructure
|
:
|
Log Bytes |
Logging/User/IAMPolicy |
Logging/User/VPN |
CPU Utilization |
Memory Usage |
Error Count |
Prediction Count |
Prediction Latencies |
Private Endpoint Prediction Latencies |
Private Endpoint Response Count
|
Visit Model :
medictiv.citiustech.com