The role of non-oncology specialties in the real world longitudinal journey of prostate cancer patients.

2020 ◽  
Vol 38 (15_suppl) ◽  
pp. e19255-e19255
Author(s):  
Daniel Carnegie ◽  
Elenee Argentinis ◽  
Dibyajyoti Mazumder ◽  
Indu Dhangar
Keyword(s):  
Prostate Cancer ◽  
Cancer Patients ◽  
Real World ◽  
Index Date ◽  
Diagnostic Testing ◽  
Significant Proportion ◽  
Data Repository ◽  
Electronic Health Record Data ◽  
Source Data ◽  
Electronic Health

e19255 Background: Electronic health records (EHRs) are increasingly being recognized by regulators and researchers as reliable sources of evidence. It is therefore critical that real world evidence databases accurately reflect all diagnostics and interventions that could affect patient outcomes. Some of the commercially available datasets source data solely from oncology clinics, which may not reflect patients’ full care journey. This analysis assessed the contribution of non-oncology specialties to prostate cancer care. Methods: Newly diagnosed prostate cancer patients with encounters between Jan 2014 to Dec 2017 were analyzed from the deidentified Optum Electronic Health Record Data Repository. Diagnostic procedures 6 months prior to the index date (first diagnosis date within the study period) and 1 year post index date were identified. Attending physician specialties were identified. All treatment related encounters up to 1 year post the index dates were mapped by specialty. Codes were verified by certified medical professional. Results: A total of 186,299 prostate cancer patients were identified between Jan 2014 to Dec 2017. In the 6 months prior to index date, biopsy was most commonly ordered by urologists (70%), followed by surgical specialists (14%). Biomarker tests were ordered mostly by general practitioners (40%) followed by urologists (18%). The trend was similar for 1 year post biopsy and other histology procedures. Interestingly, a large portion of treatment encounters was observed outside oncology: 48% of surgical management by urologists (48%), chemotherapy was prescribed by both oncologists (27%) and urologists (31%) in a similar ratio, while radiotherapy was performed predominantly by radiation oncologists (81%). Conclusions: In prostate cancer, a large proportion of care encounters occur outside oncology specialties, with urology conducting a significant proportion of diagnostic testing and early treatment. Restricting source data to oncology specialties may omit key factors affecting patients’ outcomes, therefore data for such studies should reflect the entire care continuum.

Assess how non-oncology care settings influence the longitudinal journey of cancer patients.

2020 ◽  
Vol 38 (15_suppl) ◽  
pp. e19254-e19254
Author(s):  
Daniel Carnegie ◽  
Elenee Argentinis ◽  
Dibyajyoti Mazumder
Keyword(s):  
Cancer Patients ◽  
Cancer Care ◽  
Healthcare Delivery ◽  
General Medicine ◽  
Surgical Excision ◽  
Data Repository ◽  
Primary Case ◽  
Electronic Health Record Data ◽  
Care Continuum ◽  
Electronic Health

e19254 Background: Electronic health records (EHRs) are increasingly being accepted by regulators and as real world evidence (RWE). In a recent literature review (published elsewhere in this conference), we identified that many of the commercial databases source data solely from oncology clinics. This analysis aimed to understand the extent to which melanoma skin cancer patients receive cancer care specialties other than oncology. Methods: Newly diagnosed melanoma patients with one or more melanoma encounters (ICD10 CM) from January-Dec 2017 were included from the Optum Electronic Health Record Data Repository. Optum EHR data repository has more than 90 million de-identified patient lives mostly sourced from integrated healthcare delivery network EHRs within the US. Optum data includes the entire care continuum across primary case, across specialties, inpatient and outpatient encounters. Index date was defined as the first date of melanoma diagnosis within the study period. EHR history was assessed for 6 months prior to index and 1 year post index. All pathology, radiology and treatment encounters were identified, and all specialties who participated in the patient’s oncology diagnosis and care were counted. All ICD codes were verified by certified medical professional. Results: A total 7351 patients had one or more melanoma related encounters in 2017. During the 6 months prior to diagnosis, histological tests were ordered mostly by dermatologists (38%), followed by pathologists (19%), while oncologists and surgeons ordered 6% and 5%, respectively. Analysis of surgical excision on or after index indicated that dermatologists performed 47% of procedures followed by plastic surgeons (10%) and oncologists (9%). Oncologists comprised 44% of chemotherapy prescribing, while general medicine accounted for 14%. Conclusions: A large proportion of dermatological cancer care encounters occur outside the oncology setting. Therefore, to understand the factors that affect treatment outcomes, the EHR data used for outcomes research should include the entire care continuum.

scholarly journals Real-World Data Analysis of Confirmatory Biopsy Utilization in Suspected Prostate Cancer Patients

2018 ◽  
Vol 21 ◽  
pp. S161
Author(s):  
J Scott ◽  
R Concepcion ◽  
D Garofalo ◽  
S Verma-Kurvari ◽  
B Xu ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Data Analysis ◽  
Cancer Patients ◽  
Real World ◽  
Real World Data ◽  
World Data

scholarly journals Explainable artificial intelligence models using real-world electronic health record data: a systematic scoping review

2020 ◽  
Vol 27 (7) ◽  
pp. 1173-1185 ◽  
Author(s):  
Seyedeh Neelufar Payrovnaziri ◽  
Zhaoyi Chen ◽  
Pablo Rengifo-Moreno ◽  
Tim Miller ◽  
Jiang Bian ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Electronic Health Record ◽  
Scoping Review ◽  
Real World ◽  
Research Trend ◽  
Health Record ◽  
Electronic Health Record Data ◽  
Explainable Artificial Intelligence ◽  
Record Data ◽  
Electronic Health

Abstract Objective To conduct a systematic scoping review of explainable artificial intelligence (XAI) models that use real-world electronic health record data, categorize these techniques according to different biomedical applications, identify gaps of current studies, and suggest future research directions. Materials and Methods We searched MEDLINE, IEEE Xplore, and the Association for Computing Machinery (ACM) Digital Library to identify relevant papers published between January 1, 2009 and May 1, 2019. We summarized these studies based on the year of publication, prediction tasks, machine learning algorithm, dataset(s) used to build the models, the scope, category, and evaluation of the XAI methods. We further assessed the reproducibility of the studies in terms of the availability of data and code and discussed open issues and challenges. Results Forty-two articles were included in this review. We reported the research trend and most-studied diseases. We grouped XAI methods into 5 categories: knowledge distillation and rule extraction (N = 13), intrinsically interpretable models (N = 9), data dimensionality reduction (N = 8), attention mechanism (N = 7), and feature interaction and importance (N = 5). Discussion XAI evaluation is an open issue that requires a deeper focus in the case of medical applications. We also discuss the importance of reproducibility of research work in this field, as well as the challenges and opportunities of XAI from 2 medical professionals’ point of view. Conclusion Based on our review, we found that XAI evaluation in medicine has not been adequately and formally practiced. Reproducibility remains a critical concern. Ample opportunities exist to advance XAI research in medicine.

Sign in / Sign up

Export Citation Format

Share Document