Although cervical cancer is preventable, it still continues to be a leading cause of death. Following the evidence- based guidelines for cervical cancer prevention is challenging for healthcare providers, due to which many patients do not receive the optimal preventive care. Clinical decision support (CDS) systems can potentially improve the care delivery. However, the current CDS systems only identify patients overdue for screening, and do not suggest the optimal screening interval. Moreover they do not help with surveillance of patients with abnormal screening results. This is because the existing systems lack the capability to process free- text clinical reports that contain information needed for applying the guidelines. Hence there is a critical need for natural language processing (NLP)-enabled CDS systems that can utilize discrete as well as free-text patient information for enhancing the decision support. Our long-term goal is to improve healthcare delivery of cervical cancer prevention with guideline based reminders. The central hypothesis is that NLP- enabled CDS system will significantly improve the quality of care delivery for cervical cancer prevention. The rationale is that use of NLP will improve granularity of the guideline implementation, which will in-turn enhance care delivery. As preliminary work we have developed an NLP-enabled CDS system that automatically interprets the patient information from the electronic health record and applies the national guidelines to compute the optimal recommendation for screening and surveillance. We have performed validation of the system in a non-clinical setting.1 In this application we will proceed towards deployment of the system in the clinical setting, and will carry out studies for measuring the impact on the quality of care delivery. In ai one, we will validate the system in the clinical setting and will optimize its usability and workflw integration.
In aim two, we will test the hypothesis that reminders from the NLP-enabled CDS system to primary care providers will improve the quality of care delivery, by performing a one year intervention control study across four sites of a primary care practice.
In aim three, we will test the hypothesis that reminders to non-adherent high-risk patients will improve their surveillance rates, by performing a randomized intervention study for three months. In this study, care coordinators will utilize the CDS system for sending reminders to patients that are non-adherent and at high risk due to abnormal screenings. The main contribution of this project will be knowledge about the effectiveness of NLP in enhancing the impact of CDS systems for cervical cancer prevention, and for clinical practice in general. This project is innovative because the CDS system will utilize NLP to generate screening reminders for normal patients and surveillance reminders for patients with abnormal findings. This is a major advancement over existing systems that can only identify patients for screening.
The proposed research is relevant to public health because it will yield knowledge about the effectiveness of natural language processing (NLP) to enhance impact of CDS systems for cervical cancer prevention, and for clinical practice in general. This research will foster implementation of similar CDS systems across the nation for cervical cancer prevention and for other decision problems, which will improve the quality of healthcare delivery. Thus, the proposed research is relevant to AHRQ's mission to improve the quality, safety, efficiency, and effectiveness of health care for all Americans.
| MacLaughlin, Kathy L; Kessler, Maya E; Komandur Elayavilli, Ravikumar et al. (2018) Impact of Patient Reminders on Papanicolaou Test Completion for High-Risk Patients Identified by a Clinical Decision Support System. J Womens Health (Larchmt) 27:569-574 |
| Ravikumar, K E; MacLaughlin, Kathy L; Scheitel, Marianne R et al. (2018) Improving the Accuracy of a Clinical Decision Support System for Cervical Cancer Screening and Surveillance. Appl Clin Inform 9:62-71 |
