Gregory W. Hruby, MA, Mary Regina Boland, MA, James J. Cimino, MD, Junfeng Gao, PhD, Adam B. Wilcox, PhD, Julia Hirschberg, PhD, Chunhua Weng, PhD Department of Biomedical Informatics Columbia University, New York City
Outline
Introduction
Problem: data access bottleneck Limitations of related work Study goal: understanding of the query mediation process
Introduction
From Jan 2005 Dec 2011, the Columbia Urology Department has published a total of 244 manuscripts 244 publications 18 clinical trials 41 basic science 123 retrospective outcomes 61 literature reviews 51% percent of Columbia Urologys contribution to the medical literature is represented by retrospective outcomes work
Thus the expansion of clinical data access to our researchers is an important priority for accelerating clinical and translational research
Long-term goal
Toward a structured reference interview Model and automate the structured reference interview
Short-term goal
This study attempts to decompose and characterize the negotiation space between the QE and the MR
QE MR
Data
Key Players
Urology
Research
CDR
Data
Between July 2011 and January 2012, we recorded and transcribed 31 discussions between one query expert (QE) and eight medical researchers (MRs) Example of 5 dialogue acts
QE: Alright. So we're going to be talking about your study so I guess briefly describe to me what you want to do.
MR: So, I haven't really put much thought into it, I just talked with a guy and he suggested that he had talked with umm a pathologist and with other urologists and it would be like very, very interesting to see like after cystectomies see if the urethra was involved
QE:Uh huh
MR: Umm because that could umm like possibly umm affect you know the outcomes of like long term outcomes of the of the like complications and overall prognosis, that's what he told me. But I haven't like
QE:So we're looking at the effect of urethral involvement, urethral or ureteral?
Results: Schema
1. State the Problem (e.g., Alright, So we're going to be talking about your study so I guess briefly describe to me what you want to do.)
2. Explain the Clinical Process (e.g., And then they are diagnosed with cancer after the image?)
3. Locate Data Elements in EHRs (e.g., You will have to look in the operative note.)
4. Discuss Study Design (e.g., Because we want to exclude any disease that could potentially have an effect on the GFR.)
5. Clarify Research Workflow (e.g., It's gonna be rare. So you're probably gonna have to update it as well.)
6. Explain Data Results to Researchers (e.g., " So follow-up is last time known alive. So this is corresponding to overall survival information.)
7. Review IRB and Privacy Policies (e.g., It is expedited because it is deidentified.)
8. Confirm Completed Process (e.g., Alright. I think we have enough information.)
Results: Schema
2. Explain the Clinical Process (e.g., And then they are diagnosed with cancer after the image?)
1. Patients Demographics
2. Temporal Aspect of clinical process
1. Initial Diagnosis of Disease
2. Primary Treatment of Disease
3. Follow-up/Surveillance of disease
4. Salvage Treatment of disease
3. Laboratory Tests
4. Radiograhical studies
5. Clinical findings
1. Disease Confounders and Comorbidities
2. Social History
3. Family History
4. Clinical Stage/Risk assessment/Disease Status of Diagnosis
5. Survival: Disease Specific, and Overall survival
6. Surgical Procedure
7. Pathology
8. Medical Therapy
9. Radiation Therapy
10. Other Treatment
11. Treatment Toxicities, Complications, and Adverse events
Data normalization
Normalized
Conversation B
DA 1
DA 2
DA 3
3.0 Locate Data Elements in EHRs 2.0 Explain the Clinical Process 1.0 State the Problem
Trend Line
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 103 106 109 112 115 118 121 124 127 130 133 Conversation Acts in a Normalized Session
Trend Line
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 97 100 103 106 109 112 115 118 121 124 127 130 133
Conclusions
Contribute early understanding of the mediated query process The query negotiation space is an iterative process necessary to reach an understanding Query mediation represents a process-based needs assessment and clarification
Acknowledgments
The research was supported by grants
R01 LM009886 Bridging the semantic gap between research eligibility criteria and clinical data from the National Library of Medicine (PI: Weng) 5T15LM007079: Columbia University Biomedical Informatics Training Program (PI: Hripcsak)
22