Automatic Process Comparison for Subpopulations: Application in Cancer Care

Processes in organisations, such as hospitals, may deviate from the intended standard processes, due to unforeseeable events and the complexity of the organisation. For hospitals, the knowledge of actual patient streams for patient populations (e.g., severe or non-severe cases) is important for quality control and improvement. Process discovery from event data in electronic health records can shed light on the patient flows, but their comparison for different populations is cumbersome and time-consuming. In this paper, we present an approach for the automatic comparison of process models that were extracted from events in electronic health records. Concretely, we propose comparing processes for different patient populations by cross-log conformance checking, and standard graph similarity measures obtained from the directed graph underlying the process model. We perform a user study with 20 participants in order to obtain a ground truth for similarity of process models. We evaluate our approach on two data sets, the publicly available MIMIC database with the focus on different cancer patients in intensive care, and a database on breast cancer patients from a Dutch hospital. In our experiments, we found average fitness to be a good indicator for visual similarity in the ZGT use case, while the average precision and graph edit distance are strongly correlated with visual impression for cancer process models on MIMIC. These results are a call for further research and evaluation for determining which similarity or combination of similarities is needed in which type of process model comparison.

Download Full-text

Abstract P148: Heart Disease Among Breast Cancer Patients

Circulation ◽

10.1161/circ.137.suppl_1.p148 ◽

2018 ◽

Vol 137 (suppl_1) ◽

Author(s):

Tekeda F Ferguson ◽

Sunayana Kumar ◽

Denise Danos

Keyword(s):

Breast Cancer ◽

Heart Disease ◽

Electronic Health Records ◽

Cancer Patients ◽

Cancer Diagnosis ◽

Breast Cancer Diagnosis ◽

Health Concern ◽

Breast Cancer Patients ◽

Health Records ◽

Electronic Health

Purpose: In conjunction with women being diagnosed earlier with breast cancer and a rapidly aging population, advances in cancer therapies have swiftly propelled cardiotoxicity as a major health concern for breast cancer patients. Frequent cardiotoxicity outcomes include: reduced left ventricular ejection fraction (LVEF), myocardial infarction, asymptomatic or hospitalized heart failure, arrhythmias, hypertension, and thromboembolism. The purpose of this study was to use an electronic health records system determine if an increased odds of heart disease was present among women with breast cancer. Methods: Data from the Research Action for Health Network (REACHnet) was used for the analysis. REACHnet is a clinical data research network that uses the common data model to extract electronic health records (EHR) from health networks in Louisiana (n=100,000).Women over the age of 30 with data (n=35,455) were included in the analysis. ICD-9 diagnosis codes were used to classify heart disease (HD) (Hypertensive HD, Ischemic HD, Pulmonary HD, and Other HD) and identify breast cancer patients. Additional EHR variables considered were smoking status, and patient vitals. Chi-square tests, crude, and adjusted logistic regression models were computed utilizing SAS 9.4. Results: Utilizing diagnoses codes our study team has estimated 28.6% of women over the age of 30 with a breast cancer diagnosis (n=816) also had a heart disease diagnosis, contrasted with 15.6% of women without a breast cancer diagnosis. Among patients with heart disease, there was no significant difference in the distribution of the type of heart disease diagnoses by breast cancer status (p=0.87). There was a 2.21 (1.89, 2.58) crude odds ratio of having a CVD diagnoses among breast cancer cases when referenced to cancer free women. After adjusting for age (30-49, 50-64, 65+), race (black/white), and comorbidities (obesity/overweight, diabetes, current smoker) there was an increased risk of heart disease (OR: 1.24 (1.05, 1.47)). Conclusion: The short-term and long-term consequences of cardiotoxicity on cancer treatment risk-to-benefit ratio, survivorship issues, and competing causes of mortality are increasingly being acknowledged. Our next efforts will include making advances in predictive risk modeling. Maximizing benefits while reducing cardiac risks needs to become a priority in oncologic management and monitoring for late-term toxic effects.

Download Full-text

Abstract LB-278: Metformin exposure on survival in localized breast cancer patients with type II diabetes mellitus using electronic health records

10.1158/1538-7445.am2014-lb-278 ◽

2014 ◽

Author(s):

Matthew K. Breitenstein ◽

Leiwei Wang ◽

Richard M. Weinshilboum ◽

Jyotishman Pathak

Keyword(s):

Breast Cancer ◽

Diabetes Mellitus ◽

Electronic Health Records ◽

Cancer Patients ◽

Type Ii Diabetes ◽

Type Ii Diabetes Mellitus ◽

Type Ii ◽

Breast Cancer Patients ◽

Health Records ◽

Electronic Health

Download Full-text

Leveraging Genetic Reports and Electronic Health Records for Predicting Primary Cancers Based on FHIR and RDF (Preprint)

10.2196/preprints.23586 ◽

2020 ◽

Author(s):

Nansu Zong ◽

Victoria Ngo ◽

Daniel J. Stone ◽

Andrew Wen ◽

Yiqing Zhao ◽

...

Keyword(s):

Machine Learning ◽

Electronic Health Records ◽

Cancer Patients ◽

Genetic Data ◽

Precision Oncology ◽

Primary Cancer ◽

Health Records ◽

Web Resource ◽

Cancer Types ◽

Electronic Health

BACKGROUND Precision oncology has the potential to leverage clinical and genomic data in advancing disease prevention, diagnose, and treatments. A key research area focuses on early detection of primary cancers and the potential prediction of cancers of unknown primary in order to facilitate optimal treatment decisions. OBJECTIVE This study presents a methodology to harmonize phenotypic and genetic data features to classify primary cancer types and predict unknown primaries. METHODS We extracted the genetic data elements from a collection of oncology genetic reports of 1,011 cancer patients, and corresponding phenotypical data from the Mayo Clinic electronic health records (EHRs). We modeled both genetic and EHR data with HL7 Fast Healthcare Interoperability Resources (FHIR). The semantic web Resource Description Framework (RDF) was employed to generate the network-based data representation (i.e., patient-phenotypic-genetic network). Based on RDF data graph, graph embedding algorithm Node2vec was applied to generate features, and then multiple machine learning and deep learning backbone models were adopted for cancer prediction. RESULTS With six machine-learning tasks designed in the experiment, we demonstrated the proposed method achieved favorable results in classifying primary cancer types and predicting unknown primaries. To demonstrate the interpretability, phenotypic and genetic features that contributed the most to the prediction of each cancer were identified and validated based on a literature review. CONCLUSIONS Accurate prediction of cancer types can be achieved with existing EHR data with satisfactory precision. The integration of genetic reports improves prediction, illustrating the translational values of incorporating genetic tests early at the diagnose stage for cancer patients.

Download Full-text

PCN294 IDENTIFYING NON-SMALL CELL LUNG CANCER PATIENTS FROM A COHORT OF HETEROGENEOUS LUNG CANCER PATIENTS USING BOOSTED TREES ON ELECTRONIC HEALTH RECORDS DATA

Value in Health ◽

10.1016/j.jval.2020.04.1758 ◽

2020 ◽

Vol 23 ◽

pp. S75

Author(s):

P. Chandrashekaraiah ◽

K. Rudeen ◽

S. Agrawal ◽

S. Thiruvenkadam ◽

V.P. Vaidya ◽

...

Keyword(s):

Lung Cancer ◽

Electronic Health Records ◽

Small Cell Lung Cancer ◽

Cancer Patients ◽

Cell Lung Cancer ◽

Small Cell ◽

Small Cell Lung ◽

Health Records ◽

Lung Cancer Patients ◽

Electronic Health

Download Full-text

Application of process mining to assess the data quality of routinely collected time-based performance data sourced from electronic health records by validating process conformance

Health Informatics Journal ◽

10.1177/1460458215604348 ◽

2016 ◽

Vol 22 (4) ◽

pp. 1017-1029 ◽

Cited By ~ 10

Author(s):

Lua Perimal-Lewis ◽

David Teubner ◽

Paul Hakendorf ◽

Chris Horwood

Keyword(s):

Electronic Health Records ◽

Data Quality ◽

Performance Metrics ◽

Process Mining ◽

Hospital Performance ◽

Process Models ◽

Patient Journey ◽

Health Records ◽

Electronic Health

Effective and accurate use of routinely collected health data to produce Key Performance Indicator reporting is dependent on the underlying data quality. In this research, Process Mining methodology and tools were leveraged to assess the data quality of time-based Emergency Department data sourced from electronic health records. This research was done working closely with the domain experts to validate the process models. The hospital patient journey model was used to assess flow abnormalities which resulted from incorrect timestamp data used in time-based performance metrics. The research demonstrated process mining as a feasible methodology to assess data quality of time-based hospital performance metrics. The insight gained from this research enabled appropriate corrective actions to be put in place to address the data quality issues.

Download Full-text

Electronic health record use among cancer patients: Insights from the Health Information National Trends Survey

Health Informatics Journal ◽

10.1177/1460458217704246 ◽

2017 ◽

Vol 25 (1) ◽

pp. 83-90 ◽

Cited By ~ 3

Author(s):

Yulia A Strekalova

Keyword(s):

Electronic Health Records ◽

Electronic Health Record ◽

Cancer Patients ◽

Health Information ◽

Information Seeking ◽

Health Record ◽

Online Information ◽

Health Records ◽

Electronic Health ◽

National Trends

Over 90% of US hospitals provide patients with access to e-copy of their health records, but the utilization of electronic health records by the US consumers remains low. Guided by the comprehensive information-seeking model, this study used data from the National Cancer Institute’s Health Information National Trends Survey 4 (Cycle 4) and examined the factors that explain the level of electronic health record use by cancer patients. Consistent with the model, individual information-seeking factors and perceptions of security and utility were associated with the frequency of electronic health record access. Specifically, higher income, prior online information seeking, interest in accessing health information online, and normative beliefs were predictive of electronic health record access. Conversely, poorer general health status and lack of health care provider encouragement to use electronic health records were associated with lower utilization rates. The current findings provide theory-based evidence that contributes to the understanding of the explanatory factors of electronic health record use and suggest future directions for research and practice.

Download Full-text

USE OF AN ELECTRONIC HEALTH RECORDS REGISTRY TO IDENTIFY CARE GAPS IN CARDIOVASCULAR CARE OF CANCER PATIENTS

Journal of the American College of Cardiology ◽

10.1016/s0735-1097(18)31239-7 ◽

2018 ◽

Vol 71 (11) ◽

pp. A698

Author(s):

Steven Philips ◽

Duwayne Willett ◽

Sandeep Das ◽

Evan Sara ◽

Vaishnavi Kannan ◽

...

Keyword(s):

Electronic Health Records ◽

Cancer Patients ◽

Health Records ◽

Care Gaps ◽

Electronic Health ◽

Cardiovascular Care

Download Full-text

Electronic Health Records-Based Cardio-Oncology Registry for Care Gap Identification and Pragmatic Research: Development and Usability Study (Preprint)

10.2196/preprints.22296 ◽

2020 ◽

Author(s):

Alvin Chandra ◽

Steven T Philips ◽

Ambarish Pandey ◽

Mujeeb Basit ◽

Vaishnavi Kannan ◽

...

Keyword(s):

Heart Failure ◽

Clinical Practice ◽

Electronic Health Records ◽

Real Time ◽

Cancer Patients ◽

Left Ventricular ◽

Left Ventricular Ejection ◽

Health Records ◽

Electronic Health ◽

Cardiovascular Care

BACKGROUND Professional society guidelines are emerging for cardiovascular care in cancer patients. How effectively the cancer survivor population is screened and treated for cardiomyopathy in contemporary clinical practice remains unclear. As EHRs are now widely used in clinical practice, we tested the hypothesis whether an EHR-based cardio-oncology registry can address these questions. OBJECTIVE To develop an electronic health records (EHR)-based pragmatic cardio-oncology registry and, as proof of principle, to investigate care gaps in cardiovascular care of cancer patients. METHODS We generated programmatically a de-identified, real-time, EHR-based cardio-oncology registry from all patients in our institutional Cancer Population Registry (n=8275, 2011-2017). We investigated: 1) left ventricular ejection fraction (LVEF) assessment before and after treatment with potentially cardiotoxic agents, and 2) guideline-directed medical therapy (GDMT) for left ventricular dysfunction (LVD), defined as LVEF<50%, and symptomatic heart failure with reduced LVEF (HFrEF), defined as LVEF<50% and problem list documentation of systolic congestive heart failure or dilated cardiomyopathy. RESULTS Rapid development of an EHR-based cardio-oncology registry was feasible. Identification of tests and outcomes was similar by EHR-based cardio-oncology registry and manual chart abstraction (98% sensitivity and 92% specificity for LVD). LVEF was documented prior to initiation of cancer therapy in 20% of patients. Prevalence of post-chemotherapy LVD and HFrEF was relatively low (9% and 2.5%, respectively). Among patients with post-chemotherapy LVD or HFrEF, those referred to cardiology had significantly higher prescription of GDMT. CONCLUSIONS EHR data can efficiently populate a real-time, pragmatic cardio-oncology registry as a byproduct of clinical care for healthcare delivery investigation.

Download Full-text