Kavishwar Wagholikar, MBBS PhD1, Sunghwan Sohn, PhD1, Stephen Wu, PhD1,Vinod Kaggal1, Sheila Buehler, RN CNP2, Robert A. Greenes, MD PhD3,8, Tsung-Teh Wu, MD4, David Larson, MD MBA5, Hongfang Liu, PhD1, Rajeev Chaudhry, MBBS MPH6, Lisa Boardman, MD7
Statistics and Informatics, 2Internal Medicine, 4Anatomic Pathology, 5Colon and Rectal Surgery, 6Primary Care Internal Medicine, 7Gastroenterology and Hepatology, Mayo Clinic Rochester, MN 3Biomedical Informatics, Arizona State University, Scottsdale, AZ; 8Health Science Research, Mayo Clinic, Scottsdale, AZ
1Biomedical
Deficient data quality can impact the clinical decisions and consequently health outcomes
Stetson et al. Assessing Electronic Note Quality Using the Physician Documentation Quality Instrument (PDQI-9). Appl Clin Inform. 2012;3(2):164-74. Weiskopf et al. Methods and dimensions of electronic health record data quality assessment: enabling reuse for clinical research. J Am Med Inform Assoc. 2012 Jun 25.
2012 MFMER | slide-3/21
Wagholikar K, et al. Clinical Decision Support for Colonoscopy Surveillance Using Natural Language Processing. IEEE Healthcare Informatics, Imaging, and Systems Biology Conference. University of California, San Diego, CA; 2012.
h/o CRCsyndrome
no
yes
GI-consult if no addendum
Age
yes
>=50y
Colonoscopy now
h/o IBD
no
IBD-consult if no addendum
no
Risk
reportAbsent
no
yes
Age
yes
>=40y
Colonoscopy now
h/o CRCancer
10y
yes
3y
5y
>=40y
CRCancer
no
GI consult 1 year
<1yr
Age at
Recent cololoscopy +10y
yes no
Age at Recent
colonoscopy +5y
Colonoscopy. Preparation
adequate, good, excellent
h/o CRCancer
no >=1yr yes
CRrisk 8
no
Hyperplastic polyp
hyperplastic yes
poor, inadequate
h/o CRC
no
Recent.Poylp
PolypType
<10 mths
Repeat in 1 or 2 days
Recurrent loop
group1
group2
Methods
Studied workflow analysis Designed several processes for data reconciliation To determine the utility of these processes, quantitative analysis after selectively adding/removing each of reconciliation processes. Two parameters: i. accuracy gain of the patient parameter reconciled by the process, and ii. accuracy gain of the recommendations computed by the CDSS. 106 cases
Unertl KM,et al. Traversing the many paths of workflow research: developing a conceptual framework of workflow terminology through a systematic literature review. JAMIA. 2010 May-Jun;17(3):265-73.
2012 MFMER | slide-8/21
Workflow
Coded Patient response Questionnaire System Annual Questionnaire Registration Demographics Coded Problem List Free text Problem section in Clinical notes Endoscopists dictated free-text procedure note Pathology report templatized free-text Pathology dept. system
RFID tag on biopsy
EHR
Patient
Care providers
Endoscopist
Biopsy specimen
Pathologist
during procedure
CDSS
findings
GI Dept. system
R.Nurse
before procedure
indications
Colonoscopy TV
TV
TV
Colonoscopy TV
Endoscopy dictation
TV
Results
Conventionally, software engineering practices involve workflow studies, when the software is primarily for workflow support. Our study demonstrates that when there is lack of adequate data quality, workflow studies of the documentation phase can help discover data reconciliation strategies that can resolve the data deficiencies
Lack of NLP sensitivity can be overcome with multi-source data reconciliation. e.g. NLP on the endoscopy note for polyp size was effective when supplemented with structured GI findings. Findings of the workflow analysis obviated the laborious effort to develop NLP for some information. e.g. the RNs structured recording obviated NLP for detecting number of polyps from endoscopy note.
Conclusion
Workflow analysis was useful to identify data reconciliation strategies to address documentation gaps, which improved CDSS performance. Workflow-based data reconciliation can play an important role in designing new-generation CDSS based on complex guideline models and NLP.
Questions
Contact and related resources: waghskATgmail.com
References
Stetson et al. Assessing Electronic Note Quality Using the Physician Documentation Quality Instrument (PDQI-9). Appl Clin Inform. 2012;3(2):164-74. Weiskopf et al. Methods and dimensions of electronic health record data quality assessment: enabling reuse for clinical research. J Am Med Inform Assoc. 2012 Jun 25 Reiser SJ. The clinical record in medicine. Part 2: Reforming content and purpose. Ann Intern Med. 1991 Jun 1;114(11):980-5. Rosenbloom ST, et al. Data from clinical notes: a perspective on the tension between structure and flexible documentation. J Am Med Inform Assoc. 2011 Mar-Apr;18(2):181-6. Unertl KM,et al. Traversing the many paths of workflow research: developing a conceptual framework of workflow terminology through a systematic literature review. J Am Med Inform Assoc. 2010 May-Jun;17(3):265-73. Wagholikar K, et al. Clinical Decision Support for Colonoscopy Surveillance Using Natural Language Processing. IEEE Healthcare Informatics, Imaging, and Systems Biology Conference. University of California, San Diego, CA; 2012. Wagholikar K, et al. Clinical decision support with automated text processing for cervical cancer screening. J Am Med Inform Assoc. 2012; 19 (5), 833-839, Wagholikar K ,et al. Formative evaluation of the accuracy of a clinical decision support system for cervical cancer screening. J Am Med Inform Assoc. 2013. (in press)