Prediction using patient comparison vs. modeling: A case study for mortality prediction
Information in Electronic Medical Records (EMRs) can be used to generate accurate predictions for the occurrence of a variety of health states, which can contribute to more pro-active interventions. The very nature of EMRs does make the application of off-the-shelf machine learning techniques diffic...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2017
|
| Online Access: | http://hdl.handle.net/1721.1/112991 https://orcid.org/0000-0001-6349-7251 https://orcid.org/0000-0001-8411-6403 |
| _version_ | 1826210994161975296 |
|---|---|
| author | Hoogendoorn, Mark el Hassouni, Ali Mok, Kwongyen Ghassemi, Marzyeh Szolovits, Peter |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Hoogendoorn, Mark el Hassouni, Ali Mok, Kwongyen Ghassemi, Marzyeh Szolovits, Peter |
| author_sort | Hoogendoorn, Mark |
| collection | MIT |
| description | Information in Electronic Medical Records (EMRs) can be used to generate accurate predictions for the occurrence of a variety of health states, which can contribute to more pro-active interventions. The very nature of EMRs does make the application of off-the-shelf machine learning techniques difficult. In this paper, we study two approaches to making predictions that have hardly been compared in the past: (1) extracting high-level (temporal) features from EMRs and building a predictive model, and (2) defining a patient similarity metric and predicting based on the outcome observed for similar patients. We analyze and compare both approaches on the MIMIC-II ICU dataset to predict patient mortality and find that the patient similarity approach does not scale well and results in a less accurate model (AUC of 0.68) compared to the modeling approach (0.84). We also show that mortality can be predicted within a median of 72 hours. |
| first_indexed | 2024-09-23T14:59:02Z |
| format | Article |
| id | mit-1721.1/112991 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T14:59:02Z |
| publishDate | 2017 |
| publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| record_format | dspace |
| spelling | mit-1721.1/1129912022-10-01T23:46:09Z Prediction using patient comparison vs. modeling: A case study for mortality prediction Hoogendoorn, Mark el Hassouni, Ali Mok, Kwongyen Ghassemi, Marzyeh Szolovits, Peter Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Ghassemi, Marzyeh Szolovits, Peter Information in Electronic Medical Records (EMRs) can be used to generate accurate predictions for the occurrence of a variety of health states, which can contribute to more pro-active interventions. The very nature of EMRs does make the application of off-the-shelf machine learning techniques difficult. In this paper, we study two approaches to making predictions that have hardly been compared in the past: (1) extracting high-level (temporal) features from EMRs and building a predictive model, and (2) defining a patient similarity metric and predicting based on the outcome observed for similar patients. We analyze and compare both approaches on the MIMIC-II ICU dataset to predict patient mortality and find that the patient similarity approach does not scale well and results in a less accurate model (AUC of 0.68) compared to the modeling approach (0.84). We also show that mortality can be predicted within a median of 72 hours. 2017-12-29T19:45:54Z 2017-12-29T19:45:54Z 2016-10 2016-08 Article http://purl.org/eprint/type/ConferencePaper 978-1-4577-0220-4 http://hdl.handle.net/1721.1/112991 Hoogendoorn, Mark, et al. "Prediction Using Patient Comparison vs. Modeling: A Case Study for Mortality Prediction." 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 16-20 August, 2016, Orlando, FL, IEEE, 2016, pp. 2464–67. https://orcid.org/0000-0001-6349-7251 https://orcid.org/0000-0001-8411-6403 en_US http://dx.doi.org/10.1109/EMBC.2016.7591229 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) MIT Web Domain |
| spellingShingle | Hoogendoorn, Mark el Hassouni, Ali Mok, Kwongyen Ghassemi, Marzyeh Szolovits, Peter Prediction using patient comparison vs. modeling: A case study for mortality prediction |
| title | Prediction using patient comparison vs. modeling: A case study for mortality prediction |
| title_full | Prediction using patient comparison vs. modeling: A case study for mortality prediction |
| title_fullStr | Prediction using patient comparison vs. modeling: A case study for mortality prediction |
| title_full_unstemmed | Prediction using patient comparison vs. modeling: A case study for mortality prediction |
| title_short | Prediction using patient comparison vs. modeling: A case study for mortality prediction |
| title_sort | prediction using patient comparison vs modeling a case study for mortality prediction |
| url | http://hdl.handle.net/1721.1/112991 https://orcid.org/0000-0001-6349-7251 https://orcid.org/0000-0001-8411-6403 |
| work_keys_str_mv | AT hoogendoornmark predictionusingpatientcomparisonvsmodelingacasestudyformortalityprediction AT elhassouniali predictionusingpatientcomparisonvsmodelingacasestudyformortalityprediction AT mokkwongyen predictionusingpatientcomparisonvsmodelingacasestudyformortalityprediction AT ghassemimarzyeh predictionusingpatientcomparisonvsmodelingacasestudyformortalityprediction AT szolovitspeter predictionusingpatientcomparisonvsmodelingacasestudyformortalityprediction |