14 th Annual Mid-South Critical Access Hospital Conference. August 17-19, 2016 Nashville, Tennessee

Transcription

1 14 th Annual Mid-South Critical Access Hospital Conference August 17-19, 2016 Nashville, Tennessee

2 IMPROVING DIAGNOSIS IN HEALTHCARE: A GRAND CHALLENGE & OPPORTUNITY HARDEEP SINGH, MD, MPH HOUSTON VA CENTER FOR INNOVATIONS IN QUALITY, EFFECTIVENESS & SAFETY MICHAEL E. DEBAKEY VA MEDICAL CENTER BAYLOR COLLEGE OF MEDICINE 2

3 Center for Innovations in Quality, Effectiveness and Safety 3 From Encounter to Policy: An Integrated Approach to Improving Quality of Care

4 Multidisciplinary Team Reducing diagnostic errors Improving health ITrelated patient safety 4 Dean Daniel Traber Ashley Shailaja Roosan Medical Informatics Physician/ Health IT Social Work/ Qualitative Research Psychologist/ Analyst Sociologist/ Qualitative Research Medical Informatics Viraj Viral Elise Jessica Janet Donna Daniel Physician/ Health IT Physician/ Health IT Project Coordinator Research Coordinator Research Coordinator Project Coordinator Human Factors

5 5

6 Early Work Evaluated evidence of errors in integrated system w/detailed review of comprehensive EHR Common conditions missed in outpatient settings (5% or 1 in 20 US adults/year) High risk areas: Patient-physician interaction (history/exam) & tests follow-up Similar situation in ER settings Singh et al Arch Intern Med. 2007; Singh et al Arch Intern Med 2009

7 Estimates in Perceptual Specialties Radiology 2 4 % discrepancies on 2 nd readings Pathology 1.4% major discrepancies on 2 nd readings Dermatology 2% discrepancy rate on pigmented lesions Berner and Graber, Am J Med May

8 Estimates in Other Settings Patient Surveys MD Surveys Look backs Autopsies About 12% of patients relate a diagnostic error that affected themselves, a family member, or close friend 45% of pediatricians reported diagnostic errors that harmed patients at least once or twice per year 30% of subarachnoid hemorrhage misdiagnosed; 39% of dissecting AAA delayed diagnosis; Major unexpected discrepancies that would have changed management found in 10-20% Expert guess Arthur Elstein: 10-15% 8

9

10 10

11 Why Are Diagnostic Errors Hard to Fix? Confluence of cognitive science, informatics, human factors, social science, and the art of medicine Experts still debating definitions of diagnosis lack of standards, agreed upon definitions for most concepts involving diagnosis Operational definitions of diagnostic error harder often confused with screening and treatment errors

12 A "diagnosis" is not a static, fixed conclusion; it is a fluid, evolving conclusion based on serial observation and hypothesis building 12 Comments from frontline docs One moves from less certainty to more certainty more or less quickly depending on a number of factors Many of the complications introduced by both medicolegal and quality improvement efforts come from treating diagnosis as a black and white situation

13 Safety Begins with Measurement 13 We cannot improve what we cannot measure! We cannot measure what we cannot define!

14 IOM Definition of diagnostic error 14 The failure to a) establish an accurate and timely explanation of the patient s health problem(s) or b) communicate that explanation to the patient

15 Operational Definition of Diagnostic Error 15 Case analysis reveals unequivocal evidence of a missed opportunity to make a correct or timely diagnosis (do something different?) Missed opportunity is framed within the context of an evolving diagnostic process The opportunity could be missed by the provider, care team, system, and/or patient H Singh Jt Comm J Qual Patient Saf 2014

16 Defining Preventable Diagnostic Harm 16 MISSED OPPORTUNITIES HARM (from delayed or wrong treatment/test) NO MISSED OPPORTUNITIES A B C D Missed opportunities in diagnosis due to system and/or cognitive factors Preventable diagnostic harm Delayed/wrong diagnosis associated with patient harm but no clear evidence of missed opportunities Delayed/wrong diagnosis but no clear evidence of missed opportunities Adapted from Singh Jt Comm J Qual Patient Saf 2014

17 17 Diagnosis # cases % Pulmonary embolism % Poisoning, ADR, overdose % Lung cancer % Colorectal cancer % Acute coronary syndrome % Breast cancer % Stroke % Congestive heart failure % Fracture % Abscess % Pneumonia % Aortic aneurysm/dissection 9 1.5% Appendicitis 9 1.5% Depression 9 1.5% Schiff et al 2009

18 Contributing Factors 18 Premature closure Overconfidence Unintended consequence of policy Faulty data gathering Misinterpretation of test Inadequate follow-up Failed heuristic Knowledge deficit Affective bias Process failure Faulty synthesis Failure to detect physical finding Perception error Sample mix-up Wrong estimate of pretest probability Failure to follow-up abnormal test Limited access Communication failure Language barrier Faulty triggering Uninformed patient Thanks to Karen Cosby, MD

19 Diagnostic Accuracy and Confidence Physicians assessed 4 clinical vignettes (2 easy & 2 difficult) based on real-world cases Goals Assess how diagnostic accuracy is aligned with perception of confidence in that accuracy Meyer et al JAMA Intern Med 2013

20 20 Diagnostic Accuracy vs. Confidence

21 Research Summary 21 Common diseases missed despite red flags Failure to elicit key history or exam finding Overlooking documented critical information Absent differential diagnosis Diagnostic accuracy not always black & white Poor calibration: struggle between underdiagnosis and over-zealous diagnostic pursuits No single systems or cognitive solution Meyer et al JAMA IM 2013; Singh et al JAMA IM 2013; Sarkar et al BMJQS 2012

22 3 of 8 IOM Goals 22 GOAL 1 Facilitate more effective teamwork in the diagnostic process among health care professionals, patients, and their families GOAL 3 Ensure that health information technologies support patients and health care professionals in the diagnostic process GOAL 4 Develop and deploy approaches to identify, learn from, and reduce diagnostic errors and near misses in clinical practice

23 Potential Areas of Solutions 23 Measurement Information Technology Missed Test Results Communication and Teamwork Patient Engagement

24 To Enable Rigorous Measurement 24 Missed opportunity measurement must reflect real-world practice more than just what s in the doctors head systems, team members, and patients, all inevitably influence clinicians thought processes Singh BMJQS 2013

25 Safer Dx Framework for Measurement & Reduction 25 Singh & Sittig BMJQS 2015

26 How to Study: Retrospective Measurements 26 Signals from traditional administrative data are weaker If validated, could provide clues on possible missed opportunities that warrant additional clinical evaluation Stronger signals to bolster error measurement Review of high-risk cohorts (cancer ~ 1/3rd delays) Targeted record reviews using triggers such as unexpected return visits which led to hospitalization Singh et al Am J Gastro 2009; Singh et al JCO 2010 Singh et al Arch IM 2012 Singh et al BMJQS 2011

27 How to Study: Retrospective Measurements 27 Trigger & Review Strategy (High-Risk Records) Documentation-dependent Hard to identify contributory/cognitive factors Quality of data issues Provider Reports/Surveys Not yet covered comprehensively by AHRQ Common Formats Patient Reports? Singh et al JAMA IM 2012; Singh and Sittig BMJQS 2015; Singh et al Peds 2010;

28 We Must Measure Diagnostic Error 28 Now ready for Quality Improvement, Learning & Research purposes Measurement not ready for public reporting, performance measurement or penalties Need to push the basic science ahead Good data, standards and operational definitions More institutions need to do this Singh & Sittig BMJQS 2015

29 Intersection of Health IT & Diagnostic Safety 29 Our goals are to use health IT to measure and reduce diagnostic errors and harm, but.. Current Reality: Trying to ensure health IT itself is being used safely Adapted from Sittig & Singh N Engl J Med. 2012

30 Communication of Test Results 30 Evaluation of 1,163 outpatient abnormal lab & 1,196 abnormal imaging test result alerts 7% abnormal labs lacked timely follow-up 8% abnormal imaging lacked timely follow-up Why abnormal test results continue to get missed in health IT-based settings Singh et al Am J Med 2010 & Singh et al Archives of Int Med 2009

31 31 Ambiguous Responsibility a Huge Issue

32 32

33 And More Digital Data Is on the Way 33 Smartphone Wearables Patients can now continuously monitor their data real-time and send it to their docs

34 Multiple Socio-Technical Issues 34 Issue Software Content Usability Workflow Providers Organizational Examples no functionality for saving, tracking, and retrieving alerts too many unnecessary alerts poor signal to noise ratio on screen surrogate feature to forward alerts when providers out of office not used properly lack of knowledge/training policies for follow-up ambiguous Singh et al JAMA Int Med 2013

35 35 8-dimensional Socio-Technical Model of Safe & Effective Health IT Use Sittig Singh QSHC 2010

36 Health IT Safety Framework 3 Domains 36 Domain 1: Safe health IT: Events unique/specific to health IT Sittig & Singh N Engl J Med Nov 8;367(19):

37 37

38 Health IT Safety Framework 3 Domains 38 Domain 1: Safe health IT: Events unique/specific to health IT Domain 2: Using health IT safely: Unsafe or inappropriate use of technology Unsafe changes in the workflows that emerge from technology use Sittig & Singh N Engl J Med Nov 8;367(19):

39 The Famous Ebola Misdiagnosis Temperature of 100.1F103, dizziness, nausea, abdominal pain, and decreased urination Travel history recorded in the nurse s EHR notes Day 1-Blame Nurse Day 2-Blame EHR Day 3-None of the above Report 1 year later Upadhyay D, et al. Diagnosis 2014

40 Health IT Safety Framework 3 Domains 40 Domain 1: Safe health IT : Events unique/specific to EHRs Domain 2: Using health IT safely: Unsafe or inappropriate use of technology Unsafe changes in the workflows that emerge from technology use Domain 3: Using health IT to improve safety Leveraging health IT to identify unsafe care processes and potential patient safety concerns before harm Sittig & Singh N Engl J Med Nov 8;367(19):

41 41

42 Targeting a High Priority Area 42 Missed/delayed Cancer Diagnosis a safety concern Major reason: Lack of timely followup of cancer-related abnormal test results Singh et al JCO 2010 Singh et al Am J Gastro 2009

43 Big Data Safety Net 43 Electronic health record (EHR)-based triggers look for follow-up actions on clues (or red flags) to detect delays prospectively Basic versions: + hemoccult or microcytic anemia with no subsequent colonoscopy in 60 days suspicious chest-x ray with no follow-up CT scan in 30 days Murphy et al BMJQS 2013

44 44

45 45 Singh Sittig BMJ Qual Saf. doi: /bmjqs

46 Must Evaluate How We Are Doing 46 The Office of the National Coordinator for Health Information Technology (ONC)-sponsored Safety Assurance Factors for EHR Resilience (SAFER) project Proactive risk assessment and guidance 1 st draft of best practices and knowledge Self-assessment; not meant to be regulatory Focused on high-risk areas Nine guides all freely available Singh et al BMC Med Inf 2013

47

48 48 Opportunities for Patient Engagement

49 Multi-Site Physician Perspective 49 65% of MDs surveyed agreed that patients should receive direct notification of normal test results Only 21% comfortable with patients receiving direct notification of clinically significant abnormal test results Physician concerns include: 85% : anxiety about test results 89%: confusion about test results 85%: patients lack expertise necessary to interpret results 75%: patients will seek unreliable information Giardina et al Patient Edu & Counselling 2015

50 Releasing Results to Patients 50 The result was abnormal but I didn t realize it. There s a comment section but the doctor never leaves a comment. My triglycerides are high. Ok, what does that mean? What am I supposed to do? I had to figure out the sodium was low. There s a problem with low sodium, what can I do? I m not a doctor. I hope they ll call if it s problematic.

51 Patient engagement 51 Key in improving safety of diagnosis Many opportunities for improvement in test results follow-up We must preach No news is not good news

52 52

53 53

54 Review of Grand Challenges 54 Common diagnoses missed or delayed Critical information in EHRs overlooked Upto a third of abnormal results missed Chaotic clinical settings & inadequate time to solve diagnosis-related challenges Lack of assessment and feedback systems for improvement Meyer et al JAMA IM 2013; Singh et al JAMA IM 2013; Sarkar et al BMJQS 2012

55 55

56 Innovations to Support Diagnosis 56 Dimension Examples Software Better health IT tools/functions Content Usability Smarter alerts & diagnostic decision support Better user-interfaces; signal to noise ratio

57 Need Systems Approaches 57 Dimension Workflow People Examples Time interacting with patients Culture change Organization Policies for closed-loop test results follow-up External rules National entity for shared learning Evaluation & Measurement Data to separate reality vs. hype; Triggers & ONC SAFER Guides

58 Thank you and Acknowledgements 58 Funding Agencies: Department of Veterans Affairs National Institute of Health Agency for Health Care Research & Quality Multidisciplinary team at VA Health Services Research Center for Innovation Web:

59 14 th Annual Mid-South Critical Access Hospital Conference August 17-19, 2016 Nashville, Tennessee