With the advance of EHR, a great variety of data is gathered in clinical practices, using which treatment and treatment policies can be evaluated for patients. The study focuses on learning individualized treatment rules(ITRs) to come up with a treatment policy that is expected to give better results for an individual patient. For real world dataset the model is applied to learn the optimal ITRs for intravenous (IV) and vasopressors (VP) administration. The method applied performs better as compared to both physicians and other baselines. The data was used from freely accessible database called The Medical Information Mart for Intensive Care(MIMIC-III) database. In total 20,944 admissions are included in dataset consisting of adult patients having a Sequential Organ Failure Assessment (SOFA) Score of 2 and above. Static information like age, gender and sequential information like time-varying heart rate and respiratory rate are used for study. In this approach there are two parts: a predictive model for propensity score estimation and an ITRs model trained with an objective function based on estimated propensity score.
Input variables : Age, Gender, Time varying heart rate, Respiratory rate.
Output Variables : The policy of the optimal ITRs
| 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 |
| 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 : github.com
Additional links : arxiv.org
| Model Category | : | Public |
| Date Published | : | July, 2020 |
| Healthcare Domain | : | Provider |
| Code | : | github.com |
| Patient Centric Care |
| Treatment Policy |