Home
|
Inbox
|
Search
|
View Abstract
Derivation and Validation of the Automated Search Algorithms to Identify Patients with Ischemic Stroke and Transient Ischemic Attack in Electronic Health Records
Description
.abstract img { width:300px !important; height:auto; display:block; text-align:center; margin-top:10px } .abstract { overflow-x:scroll } .abstract table { width:100%; display:block; border:hidden; border-collapse: collapse; margin-top:10px } .abstract td, th { border-top: 1px solid #ddd; padding: 4px 8px; } .abstract tbody tr:nth-child(even) td { background-color: #efefef; } .abstract a { overflow-wrap: break-word; word-wrap: break-word; }
A1489 - Derivation and Validation of the Automated Search Algorithms to Identify Patients with Ischemic Stroke and Transient Ischemic Attack in Electronic Health Records
Author Block: M. Sabov1, M. A. Hawkes2, M. A. Passe3, T. J. Weister3, A. A. Rabinstein4, R. Kashyap5; 1Division of Pulmonary and Critical Care, Mayo Clinic, Rochester, MN, United States, 2Neurology Critical Care Medicine, Mayo Clinic, Rochester, MN, United States, 3Anesthesia Clinical Research Unit, Mayo Clinic, Rochester, MN, United States, 4Cerebrovascular Neurology, Mayo Clinic, Rochester, MN, United States, 5Anesthesia and Critical Care Medicine, Mayo Clinic, Rochester, MN, United States.
Introduction/Rationale: Stroke and transient ischemic attack (TIA) are common reasons and important problems in critically ill patients. We developed and validated a reliable automated electronic search algorithm because of increasing requests to identify acute stroke and TIA in electronic medical records (EMR).
Methods: The automated search algorithm was derived from subsequent analysis of cohort of 9 518 non-cardiac surgery patients at a mid-western tertiary care center. The initial derivation included 100 patients followed by a validation of another 100 patients. Both cohorts had independent (gold standard) manual review done and then compared with automated digital search algorithms. We used International Classification of Diseases Ninth and Tenth Revisions (ICD-9, ICD-10) codes for the first search algorithm. Another algorithm contained comprehensive medical stroke search terms list. Sensitivity and specificity of both search algorithms in derivation and validation cohort were analyzed and compared with the manual chart review.
Results: In the derivation cohort, ICD-9 and ICD-10 automated search algorithms showed sensitivity of 96.4% and specificity of 95.8%. Sensitivity and specificity of the validation cohort were 89.2% and 92.1%, respectively. On the other hand, automated algorithms with stroke search terms showed different results for both cohorts. Derivation cohort had a sensitivity of 89.3% and specificity 87.5% and the validation cohort had sensitivity 75.6%, and specificity 100%.
Conclusion: In this study, we have developed an electronic automated search algorithm that identifies stroke and TIA in electronic medical records which is feasible and reliable. Algorithm with ICD-9 and-10 codes outperformed the stroke search terms algorithm in every aspect except specificity validation.
Author Block: M. Sabov1, M. A. Hawkes2, M. A. Passe3, T. J. Weister3, A. A. Rabinstein4, R. Kashyap5; 1Division of Pulmonary and Critical Care, Mayo Clinic, Rochester, MN, United States, 2Neurology Critical Care Medicine, Mayo Clinic, Rochester, MN, United States, 3Anesthesia Clinical Research Unit, Mayo Clinic, Rochester, MN, United States, 4Cerebrovascular Neurology, Mayo Clinic, Rochester, MN, United States, 5Anesthesia and Critical Care Medicine, Mayo Clinic, Rochester, MN, United States.
Introduction/Rationale: Stroke and transient ischemic attack (TIA) are common reasons and important problems in critically ill patients. We developed and validated a reliable automated electronic search algorithm because of increasing requests to identify acute stroke and TIA in electronic medical records (EMR).
Methods: The automated search algorithm was derived from subsequent analysis of cohort of 9 518 non-cardiac surgery patients at a mid-western tertiary care center. The initial derivation included 100 patients followed by a validation of another 100 patients. Both cohorts had independent (gold standard) manual review done and then compared with automated digital search algorithms. We used International Classification of Diseases Ninth and Tenth Revisions (ICD-9, ICD-10) codes for the first search algorithm. Another algorithm contained comprehensive medical stroke search terms list. Sensitivity and specificity of both search algorithms in derivation and validation cohort were analyzed and compared with the manual chart review.
Results: In the derivation cohort, ICD-9 and ICD-10 automated search algorithms showed sensitivity of 96.4% and specificity of 95.8%. Sensitivity and specificity of the validation cohort were 89.2% and 92.1%, respectively. On the other hand, automated algorithms with stroke search terms showed different results for both cohorts. Derivation cohort had a sensitivity of 89.3% and specificity 87.5% and the validation cohort had sensitivity 75.6%, and specificity 100%.
Conclusion: In this study, we have developed an electronic automated search algorithm that identifies stroke and TIA in electronic medical records which is feasible and reliable. Algorithm with ICD-9 and-10 codes outperformed the stroke search terms algorithm in every aspect except specificity validation.
|
Home
|
Inbox
|
Search
|