Use of an EMR-based Registry to Support Clinical Research
5 likes3,117 views
This document discusses using an electronic medical record (EMR) to support clinical research. It outlines how EMR data can be used throughout the research process, including determining study feasibility, including data in grant applications, patient recruitment, study interventions, data collection, and assessing study outcomes. While EMRs provide rich clinical data and can streamline aspects of research, the data was primarily collected for clinical care so requires validation for research purposes. Fully integrating research workflows into EMRs remains a challenge.
Use of an EMR-based Registry to Support Clinical Research
- 1. Use of an EMR-based Registry to Support Clinical Research December 14, 2012 John Sharp, MSSA, PMP, FHIMSS Manager, Clinical Research Informatics Quantitative Health Sciences
- 2. Clinical Data Repositories The Clinical Data Repository (CDR) is the “perfect infrastructure” to run clinical trials “we’re moving into an era of data. To have data analytics allow for population-based care management, we need to make sure we have the tools that really work for providers.” Steven R. Steinhubl, MD, director of Geisinger Health System’s Cardiovascular Wellness Center Clinical Innovation and Technology 10/30/12 Medical Informatics Grand Rounds 12/14/12 l 2
- 3. Outline • Determining study feasibility with EMR data • EMR data for inclusion in grant applications • EMR data in clinical trial recruitment • EMR use in study intervention • EMR data for collecting data • EMR data for study outcomes Medical Informatics Grand Rounds 12/14/12 l 3
- 4. Using EMR data in Research • Rich clinical data available • Large amounts of data – 4M patients, > 10 years • CKD Registry – almost 60,000 patients since 2005 • Largest registry of its kind for CKD But • Some missing data • Data not as clean as clinical trial data management • Some patients lost to follow up Medical Informatics Grand Rounds 12/14/12 l 4
- 5. Overcoming obstacles to the use of EMR data • Large volume of data • Inference tools • Data validation • Improving data entry which can enable research • Becoming familiar with data outputs • Standardize definitions, rules Medical Informatics Grand Rounds 12/14/12 l 5
- 6. CKD Registry Technical Architecture • Oracle Nightly • Oracle • Refresh Epic Extract Clarity • Views QHS weekly or monthly Medical Informatics Grand Rounds 12/14/12 l 6
- 7. STUDY FEASIBILITY Medical Informatics Grand Rounds 12/14/12 l 7
- 8. Study Feasibility • “We have developed and validated a comprehensive EHR-based CKD registry, which identified nearly 60,000 CKD patients with their attendant co- morbidities in the CCHS. The CKD Registry enabled us to identify nearly 18,000 patients with CKD Stages 3B and 4 who may be eligible for enrollment in our proposed randomized control trial.” • From NIDDK grant application Medical Informatics Grand Rounds 12/14/12 l 8
- 9. Should all proposals be required to do study feasibility? • Prevent studies which fail to recruit adequate number of subjects • Inclusion/exclusion criteria can be modified to select a broader or narrower cohort • Plan for recruiting at main campus only or including Family Health Centers, Regional Hospitals or other sites Medical Informatics Grand Rounds 12/14/12 l 9
- 10. USING EMR DATA IN GRANT APPLICATIONS Medical Informatics Grand Rounds 12/14/12 l 10
- 11. Grant Application from Registry - 1 • We have developed an EHR-based CKD registry at the Cleveland Clinic and validated comorbid conditions. Patients who had at least one face-to-face outpatient encounter with a Cleveland Clinic health care provider and a) had two eGFR values <60 ml/min/1.73 m2 more than 90 days apart as of January 1, 2005 and/or b) were designated with International Classification of Diseases (ICD-9) codes for kidney disease (used twice in an outpatient encounter) were included Medical Informatics Grand Rounds 12/14/12 l 11
- 12. Grant Application from Registry - 2 Mean age was 69.5 ±13.4 years, with 55% females and 12% African Americans. The kappa statistics to assess the extent of agreement between the administrative dataset derived from the EHR and actual EHR chart review showed substantial agreement (≥ 0.80) for all conditions except coronary artery disease and hypertension which had moderate agreement (<0.60) suggesting the reliability of the registry. Our CKD registry will be used to identify and recruit patients for this clinical trial. Medical Informatics Grand Rounds 12/14/12 l 12
- 13. PATIENT RECRUITMENT Medical Informatics Grand Rounds 12/14/12 l 13
- 14. Patient Recruitment – contact by phone, mail • Filter patients in the registry who fit inclusion/exclusion criteria • Generate address list for initial contact • Primary care physician – contact to allow to opt out of recruiting a patient • Phone contact follow up • Some Epic sites using silent alerts – when a patient fits criteria, message sent to study coordinator • Physician alerts – often ignored. See Embi, J Am Med Inform Assoc. 2012 Jun;19 Medical Informatics Grand Rounds 12/14/12 l 14
- 15. Patient Recruitment • Report which filters on eligible patients • Indicates location of next appointment • Includes patient phone number • Research coordinator can call patient and arrange a meeting to sign consent, do pre-screening questionnaire Medical Informatics Grand Rounds 12/14/12 l 15
- 16. Patient Recruitment Report Medical Informatics Grand Rounds 12/14/12 l 16
- 17. Recruitment Results Month Count Aug-12 23 Sep-12 6 Oct-12 20 Nov-12 11 All Hands Meeting l 12/17/2012 l 17
- 18. STUDY INTERVENTION Medical Informatics Grand Rounds 12/14/12 l 18
- 19. Study Intervention – Enhanced MyChart Medical Informatics Grand Rounds 12/14/12 l 19
- 20. Study Intervention • Patient use of enhanced MyChart vs. patient use of standard MyChart • Does it make a difference in outcomes? • Current studies show mixed results • PHR use, but not intensity of use, was associated with improved diabetes quality measure profiles. To maximize value, next-generation PHRs must be designed to engage patients in everyday diabetes self- management. J Gen Intern Med. 2012 Apr;27(4):420-4 All Hands Meeting l 12/17/2012 l 20
- 21. EMR FOR COLLECTION DATA Medical Informatics Grand Rounds 12/14/12 l 21
- 22. Data Collection for Research through the EMR • Registries – discrete data available • Use of Smart Data Elements – gather discrete data for a registry or clinical trial • Clinical trials – routinely transfer lab data electronically from Epic/Clarity to study database • Other tests/procedures, e.g., echo results, imaging studies, pathology Medical Informatics Grand Rounds 12/14/12 l 22
- 23. Issues in Data Collection through the EMR • Special forms created for a clinical trial not easily archived, may pop up in clinical workflow after study end • Determining which visits are research visits – transfer only that data • Health Status measures through MyChart – not routinely used, how to use at research visit? • Knowledge Program and Clinical Solutions addressing this Medical Informatics Grand Rounds 12/14/12 l 23
- 24. Data Collection from the EMR • For CKD study: – Collecting specific lab results – Collecting comorbidities – Collecting information on Patient Navigator interventions – Information on outpatient encounters – Outcomes will be focused on lab results – rate of progression of CKD Medical Informatics Grand Rounds 12/14/12 l 24
- 25. STUDY OUTCOMES Medical Informatics Grand Rounds 12/14/12 l 25
- 26. Data on Study Outcomes • Straightforward if the outcomes are discrete data, such as, lab results, procedure results • More challenging if the outcomes are within notes (encounter notes, pathology or microbiology notes, imaging notes) • May require interpretation and then entry into a separate database with web forms Medical Informatics Grand Rounds 12/14/12 l 26
- 27. Additional Issues in EMR data in Research • Data preparation for analysis • Lack of coding of data into standard ontologies, such as, LOINC, CDISC, MEDRA, WHODrug • Dynamic nature of EMR data, e.g., labs flagged as initial results vs. final • Development of Smart Data Elements may extend timeline • Device data not always available Medical Informatics Grand Rounds 12/14/12 l 27
- 28. Data Management – Inside or Outside the EMR Inside Outside • Fit with clinical workflow • More flexibility in form design, data collection • All data in one place • Need for custom • Import data from EMR, data less dynamic templates, forms • Health Status measures • Can code data through MyChart or other • Health status/QOL means measures via REDCap or other survey tool • Requires data integration Medical Informatics Grand Rounds 12/14/12 l 28
- 29. Epic 2010 – Research Studies Can implement multiple features • Study recruitment • Associating patients with studies • Associating encounters with studies • Associating orders with studies • Billing for research, including charge routing and review • Releasing information for research patients • Potential interface with Clinical Trial Management System Medical Informatics Grand Rounds 12/14/1229
- 30. Conclusions • EMRs can be used throughout the research pipeline • Use of the EMR can streamline data collection • However, EMR data use in research requires validation since data is collected primarily for clinical care • There is not yet seamless integration of research workflow into the EMR • Disease Registries outside the EMR can be utilized for many aspects of clinical trials Medical Informatics Grand Rounds 12/14/1230
- 31. In Press • Electronic Health Records: a new tool to combat chronic kidney disease? • Clinical Nephrology • SD Navaneethan, SE Jolly, J Sharp, A Jain, JD Schold, MJ Schreiber, JV Nally Medical Informatics Grand Rounds 12/14/12 l 31