Expert Panel: [ARCHIVED] Diagnostic Error Reduction

When: Jan. 26, 2015 - Jan. 30, 2015 | Community: [ARCHIVED] Quality & Safety  

This Expert Panel is Archived.

While this Expert Panel is no longer active, we invite you to review and recommend past replies and resources. Membership for this Expert Panel is closed, but we hope you'll join us in one of the many communities on GHDonline.

Panelists of Diagnostic Error Reduction and GHDonline staff

Diagnostic Error Reduction

Posted: 19 Jan, 2015   Recommendations: 20   Replies: 41

Delays in diagnosis prevent patients from receiving the correct treatment in the timeframe that would result in the best health outcomes. Diagnostic errors can be fatal, and often involve common medical conditions. Incorrect or delayed diagnoses affect 1 in 20 adults in outpatient settings in the U.S., around 12 million people per year (Singh et al., BMJ Quality & Safety 2014).

A number of contributing factors lead to diagnostic errors at both the individual and system level, including: high volume of patients and time constraints that keep physicians from collecting comprehensive patient information; overconfidence in diagnostic accuracy, which can limit clinicians’ likelihood to reexamine difficult cases; and unreliable or un-integrated electronic health record (EHR) systems, especially in complex outpatient settings involving multiple locations, visits and providers, which can prevent necessary coordination between those providers. Correct diagnoses also rely on the effectiveness and efficiency of communication processes. While these errors are difficult to address partly because of their complex origins, recent research has informed their understanding and possible prevention.

From January 26th-30th, panelists will share their research about diagnostic errors in a range of settings. We are pleased to welcome our panelists for this discussion:

     • Gurpreet Dhaliwal, MD - Clinician-educator and Professor of Clinical Medicine at the University of California, San Francisco and site director of the internal medicine clerkships at the San Francisco VA Medical Center

     • Mark L. Graber, MD, FACP, - Senior Fellow at RTI International; Professor Emeritus of Medicine at the State University of New York at Stony Brook; Founder and President, Society to Improve Diagnosis in Medicine (

     • Urmimala Sarkar, MD MPH - Associate Professor of Medicine at UCSF in the Division of General Internal Medicine and a primary care physician at UCSF's General Medicine Clinic

     • Hardeep Singh, MD, MPH - General internist and patient safety researcher at the Center for Innovations in Quality, Effectiveness and Safety based at the Michael E. DeBakey Veterans Affairs Medical Center and an associate professor at Baylor College of Medicine

     • Dean Sittig, PhD - Professor at the School of Biomedical Informatics in The University of Texas Health Science Center at Houston and a member of the UT Houston-Memorial Hermann Center for Healthcare Quality and Safety

Our panelists will offer insight into the following questions:

      1. Measurement is the first step to improvement so how do you define and identify diagnostic errors in real-world practice?

      2. What are the most common causes for diagnostic error that you have seen in your work?

      3. What strategies could be successful in reducing diagnostic errors? How many of these have evidence for their implementation at this time?

      4. What role does use of information technologies and electronic health records play in diagnostic error reduction?

      5. How can doctors “think more effectively” to reduce diagnostic errors given the current work demands?

      6. What role do patients play in reducing error?

This panel is part of our US Communities Initiative, which is supported by the Agency for Healthcare Research and Quality (AHRQ), and aims to foster discussions between health care professionals on evidence-based practices, and translating these practices across disparate settings, to improve healthcare delivery in underserved populations in the US.

In an effort to understand the impact of our Expert Panels, we’ve created a short (4 question) survey. Your responses are greatly appreciated—please take the survey before the discussion begins:

We look forward to a rich discussion next week – please join the conversation and share your questions or comments!



Jessica Ludvigsen Replied at 2:41 PM, 21 Jan 2015

In preparation for next week's discussion, I want to share a number of resources that might be of interest.

Attached resources:

A/Prof. Terry HANNAN Replied at 3:18 PM, 21 Jan 2015

This is an EXCITING topic and I believe this set of panellists will provide outstanding input on a topic I deeply interested in. Not to undermine Jessica's resource postings here are some more.
1. Stead WW, Starmer JM. Beyond Expert-based Practice. Institute of Medicine (IOM) 2008. Evidence-Based Medicine and the Changing Nature of Health Care: . 2007 Annual Meeting Summary, Washington, DC: The National Academies Press, pp 94-105. 2008.
2.Weed LL, Weed L. Medicine in Denial. Version 1.0 ed: Createspace; 2011.
3.Topol EJ. The creative destruction of medicine: how the digital revolution will create better health care. New York, Basic Books 2012.

Mighty Casey Replied at 3:46 PM, 21 Jan 2015

Would like to add a podcast that features three UPenn MDs talking about cognitive diagnostic error, and how Kahneman's "Thinking Fast and Slow" can help inform clinician thinking when confronted with a horses/zebras fork in the road.

Attached resource:

haripriya kaur Replied at 5:37 PM, 21 Jan 2015

Hi , does anyone know how i can consult with someone about a strange case i am working up.It relates to elevated serum ammonia level s and atypical presentation of a patient.


Marie Connelly Replied at 6:04 PM, 21 Jan 2015

Many thanks for sharing these additional resources, Terry and Casey!

Haripriya, thanks for your question—while the GHDonline community does not provide clinical consultations, you may find some helpful starting points in our Discussion Brief on creating, working in and evaluating telemedicine projects:

A few additional resources attached here as we prepare for next week's Expert Panel on this important topic. Some of you may have noticed that the first link above actually goes to a different article—please find the correct link below. There's also an extensive reading list, provided by panelist Mark Graber, attached here for all to download.

Attached resources:

Sandeep Saluja Replied at 6:15 PM, 21 Jan 2015

The need for an unbiased case based discussion resource is well recognised.Further,there is an even greater need to have such a resource specifically targeting needs of low resource and remote areas.

William Martinez Replied at 11:05 AM, 22 Jan 2015

There are some physician social networks that claim to offer a space for case discussion, collective thought, and expert consultation. I have not used them so cannot speak about their effectiveness or endorse them but Haripriya you might want to check them out. There is a link to one of them below but there are others.

Attached resource:

DIMITRIOS PAPAVENTSIS Replied at 11:39 AM, 22 Jan 2015

Published International Standards on Risk Management include:

• ISO 14971:2007 Medical devices -- Application of risk management to medical devices
• ISO/TS 22367:2008 Medical laboratories -- Reduction of error through risk management and continual improvement
• ISO 31000:2009 Risk management -- Principles and guidelines
• ISO/IEC 31010:2009 Risk management – Risk assessment techniques
• ISO Guide 73 Risk management — Vocabulary
• (CLSI EP23-A) Laboratory Quality Control Based on Risk Management (2011)

Victoria Hill Replied at 5:26 AM, 24 Jan 2015

I've looked through resources and am really looking forward to the Error Reduction panel!

One additional resource "Checklists to Reduce Diagnostic Errors" Ely, J,, Graber, M.and Croskerry, P (attached).

Attached resource:

Hardeep Singh Replied at 10:07 PM, 25 Jan 2015

Hello and thanks for your interest in reducing diagnostic errors. My interest in misdiagnosis began when I was in private practice as a general internist, which motivated me to pursue evidence to understand and reduce the problem. Over the last 8 years, I have conducted multidisciplinary research on diagnostic errors, as well as on using electronic health records to deliver safe and effective patient care. I aim to leverage my research findings to influence national policy and practice improvement initiatives at the intersection of health IT and diagnostic safety. My full profile is at

Measurement is essential for improvement but we cannot measure what we cannot define. I have thus focused on research to better define and measure diagnostic errors especially in the outpatient setting. Even though our research consistently uses a definition based on presence of missed opportunities we don’t always find agreement between clinicians on errors or for the ideal diagnostic pathway for a given condition (see link). Determining this in hindsight is always hard so we use a definition to account for the evolving diagnostic process.

Our research has touched upon many facets of the diagnostic process, including the clinical patient-provider encounter, referral and testing processes including follow-up of abnormal test results, and patient engagement issues. I have developed methods to help identify diagnostic errors in the outpatient setting through trigger tools based on unexpected return visits and missed follow-up of abnormal findings. In future, these methods can be refined and brought from research into real-world practice.

Because clinician’s diagnosis making abilities cannot be separated from the complex and chaotic environment in which they make diagnosis, we recently proposed a multifaceted measurement framework for safer diagnosis that addresses the diagnostic process within the constraints of the system we clinicians practice in (see link).

I hope this overview sets the stage for discussion on question 1 and I look forward to learning from you and to contributing to this week’s discussion.

Attached resources:

Urmimala Sarkar Replied at 10:53 PM, 25 Jan 2015

Hello and thank you for joining this discussion. I am a general internist, and my interest in missed and delayed diagnosis came about when, while still in my residency training, I was not notified of an abnormal test result for my patient. This in turn led to a delayed cancer diagnosis. Luckily my patient did very well, but I never forgot that close call. Now my research is in the area of outpatient safety, including diagnosis. My full academic profile is here:

Over the course of the week, I will have the chance to chime in about all of the discussion topics, but the anecdote I shared about my patient made me focus on #6- the role of the patient. I believe timely and accurate diagnosis is the responsibility of the health system. However, time and again, I have noticed that patients who are engaged in care and can advocate for themselves can be the ones to prevent a diagnostic delay or recognize a mis-diagnosis. My patient was aware of the abnormal test result and shared it with me at a scheduled follow-up visit. Had my patient not done so, the diagnosis would have been more delayed with potentially devastating consequences. That is why I believe that although we cannot ever put the responsibility or burden on patients, we must maintain a focus on patient engagement if we want to comprehensively promote timely and accurate diagnosis.

I look forward to hearing others' perspectives on this.

Gurpreet Dhaliwal Replied at 7:58 AM, 26 Jan 2015

I am looking forward to participating as a panelist on this discussion about diagnostic error.

By way of introduction, I am a generalist internist at the San Francisco VA Medical Center. I split my practice as a clinician-educator between the emergency department and the inpatient wards. My academic area of interest is how doctors develop habits of continual improvement in their diagnostic skills. I am motivated by the notion that diagnostic excellence is the inverse of diagnostic error.

I am currently examining systems that allow individual physicians to track their own cases and learn from those outcomes in order to sharpen their diagnostic (and therapeutic) decision making for current and future encounters. This may appear to be an anathema to the question of measurement, because it is N-of-1 and qualitative, but that is how the human mind, the most important piece of diagnostic technology, improves.

Mark Graber Replied at 1:19 PM, 26 Jan 2015

Thanks to everyone interested in the quality of diagnosis and how to prevent diagnostic errors.

I'm a Nephrologist, and my first introduction to diagnostic error were all the cases of spurious hyperkalemia we saw on our consult services - situations where the patients' electrolytes were really fine and normal, but there appeared to be some major abnormality because of hemolysis, for example. Almost all of these were appreciated to be spurious, but every once in a while we'd see an order for Kayexelate + sorbitol to treat a problem that didn't really exist (those patients developed diarrhea), and once or twice we saw orders for glucose and insulin (and those patients could have died from arrhythmias to the extent that their potassium fell abruptly).

As a Chief of Medicine, I learned about many more diagnostic errors. This was in my own service, where I believed the quality of care was excellent! This contradiction continues to bother me, but I think its still accurate in a general sense: The quality of healthcare in the US is superb, but it could be even better if we could address the diagnostic errors that are occurring on a regular basis in our clinics, ER's, and hospitals.

The first step in addressing any problem is admitting that you have one. The dichotomy I just discussed plays out here too: As physicians, we all think we are practicing at a very high level, and we probably are. We have a hard time seeing our own diagnostic errors, because the vast majority are inconsequential or caught, and the more serious ones we may never hear about. But these errors exist, and summed up over the tens of thousands of clinicians in our country, and the millions of diagnoses made every year, the aggregate harm is substantial.

I believe that most of these errors are preventable. This will require understanding why and where they occur, and starting to address the cognitive- and system-related factors that contribute. I helped originate the Diagnostic Error in Medicine conference series to begin this process, and more recently the Society to Improve Diagnosis in Medicine ( Hopefully this discussion on GHDonline will continue to expand the community interested in addressing this problem - we're looking forward to learning from everyone who participates.

Mark L Graber, MD FACP
President, SIDM
Senior Fellow, RTI International
Professor Emeritus, SUNY Stony Brook, NY

Attached resource:

Dean Sittig Replied at 9:29 PM, 26 Jan 2015

It is great to see so many people interested in the topic of misdiagnosis. I come at the problem from a slightly different perspective than the others. I am a PhD trained in medical informatics from the University of Utah. Just to remind everyone, the field of medical informatics actually began in the late 1950's, right along with some of the earliest computing machines, as a way to help clinicians make clinical diagnoses. Ledley and Lusted published their seminal paper in Science in 1959, entitled, "Reasoning foundations of medical diagnosis; symbolic logic, probability, and value theory aid our understanding of how physicians reason". After many years of struggling to develop these types of computer-based diagnostic algorithms, the field turned more toward trying to help clinicians by providing them (the right people) with the right information at the right time in the right format to help them make the right decision. As anyone who has used the current generation of electronic health records (EHRs) can attest the field of medical informatics still has a significant amount of work to do.

Over the last 10 years I have focused most of my research effort on trying to understand why the existing EHR systems are not working as hoped and trying to figure out how to improve them. Much of this work revolves around an 8-dimension socio-technical model that we have developed to aid us in our work. Since joining Hardeep Singh's research group here in Houston, we have embarked on several interesting projects that have explored both what is going wrong with various aspects of clinician to clinician communication as well as trying to identify ways to fix these problems. Therefore, my take on the problem of misdiagnosis focuses mostly on how the EHR either helps or hinders the diagnostic process. This often takes the form of lost patient information. One of the key concepts that we are studying is the various safety issues that can result in this information being "lost", including errors or bugs in the computing systems themselves, errors in how clinicians are using these systems, all the way to errors in the way that EHR developers have attempted to build the systems to actually help the clinicians.

I'm looking forward to the discussions.

Attached resources:

Dr.Gedai Barakzai Dawar Replied at 11:32 PM, 26 Jan 2015

Hi I am Dr Gedai Family Physician
it was very interesting issue or experience in medicine.

Gurpreet Dhaliwal Replied at 8:45 AM, 27 Jan 2015

Question #2: What are the most common causes for diagnostic errors that you have seen in your work?

Diagnostic errors are multi-factorial. A few are no-fault (diseases presenting in a highly atypical manner), but most often it’s a combination of cognitive and systems factors, usually intertwined.

I recently made a delayed diagnosis of a large bowel obstruction (LBO) in an elderly man who had an equivocal clinical presentation and an equivocal abdominal CT reading. I initially favored bowel hypomotility. The 24 hour delay caused harm. When we reviewed my diagnostic error in the departmental M&M conference, there we some features that were no-fault-ish (atypical presentation) and some aspects that were systems-ish (errant interpretation of the CT by the radiologist), but it was also clear that there were some errors in my line of thinking about the probability of LBO and the ways in which it resembles and differs from small bowel obstruction. Despite having seen many LBO cases before, I now know more about the condition and why it should have been suspected despite the aforementioned factors.

The system has some role in virtually every diagnostic error and is in perpetual need of improvement. But we can’t forget that the same can be said of our knowledge and reasoning skills as well.

Trisha Finnegan Replied at 3:33 PM, 27 Jan 2015

Is anyone doing work in very low resource settings where the inability to identify or improper identification leads to errors? I am working to better understand challenges to accurate identification in these settings.

Thank you,
Trisha Finnegan

Hardeep Singh Replied at 6:52 PM, 27 Jan 2015

In response to Question 2, Dr. Graber published a seminal paper on systems and cognitive origins of diagnostic error in 2005 (see Graber link). From our recent research on outpatient diagnostic errors, we have learned that its almost always a complex interaction between systems and cognitive issues that leads to diagnostic errors. In a large study of 190 diagnostic errors, we found breakdowns in data gathering/synthesis such as history-taking and physical examination during clinical encounters in nearly 80% (see Types and origins paper) but there are many systems and cognitive factors that likely lead to these errors. Notably, most missed diagnoses were common conditions in primary care. Research in this area is still ongoing and informing us with more robust answers.

We do know that outpatient environments are busy and clincians are performing in time pressured conditions where there is quite a bit of uncertainty in patients symptoms and disease conditions. Diseases also evolve over time and many systems, organizational or policy factors affect the way we think and diagnose. Right now, there are a lot of administrative demands on all physicians, especially for primary care physicians who manage a lot of outpatient diagnosis (see WSJ editorial). While technology and electronic health records are being increasingly implemented in some health care systems, currently we are in a relatively early learning stage and technology can sometimes come inbetween the patient-doctor interaction (see paper on Ebola misdiagnosis).

Several provider factors including training issues etc. play a role. Of note, physician overconfidence in certain difficult to diagnose conditions was illustrated in the study sent out earlier and so can lack of feedback and learning from our own mistakes (see No News JGIM editorial).

Attached resources:

Hardeep Singh Replied at 6:58 PM, 27 Jan 2015

This link might work better for the No news JGIM Editorial mentioned above:
The reference is:
Singh H, Sittig DF. Were my diagnosis and treatment correct? No news is not necessarily good news. J Gen Intern Med 2014;29:1087–9.

Attached resource:

Urmimala Sarkar Replied at 7:25 PM, 27 Jan 2015

In our recent study of physician focus groups, participants mentioned several issues they perceive to cause diagnostic errors:

-Lack of clarity amongst specialists and primary care about responsibility for making and communicating the diagnosis

- Lack of follow-up of abnormal test results

- Constant time pressure which interferes with calm and clear thinking time

- Poor information flow between different health care systems and settings

- Poor communication with patients, at times due to cultural differences

Reference is Sarkar, Urmimala; Simchowitz, Brett; Bonacum, Doug; Strull, William; Lopez, Andrea; Rotteau, Leahora; Shojania, Kaveh G .A Qualitative Analysis of Physician Perspectives on Missed and Delayed Outpatient Diagnosis: The Focus on System-Related Factors, Joint Commission Journal for Quality and Patient Safety, October 2014, vol 40, no 10, p461-470

Attached resource:

Gurpreet Dhaliwal Replied at 8:46 AM, 28 Jan 2015

Question #3: What strategies could be successful in reducing diagnostic errors? How many of these have evidence of their implementation at this time?

Systems currently don’t do much to reduce diagnostic errors, but individual clinicians can. Physicians who track the outcomes of patients who remain in or leave their care can determine if a diagnosis they made was mistaken or delayed and set a goal of reducing such events in their own practice.

Electronic health records make that tracking easier, but it’s not easy. It takes a reminder to know when to check, a trigger for what to check and why (e.g., “check throat culture to stop antibiotics if negative”), dedicated time to do the checking (not built into current workflows), staff to help with follow through (e.g., contacting the patient), and the cognitive wherewithal to neither minimize nor overreact to errors.

Better information technology systems, valuation of this form of continuous quality improvement by employers, and a change in clinician mindset are all needed to make learning through systematic feedback on individual cases (not aggregate scorecard measures) a reality. Ivers et al. recently summarized the situation best: "There is substantial evidence that audit and feedback can effectively improve quality of care, but little evidence of progress in the field.” (

EDITA FALCOI Replied at 11:36 AM, 28 Jan 2015

The elusive nature of diagnosis error

1´what populations are we talking about?Different countries, different
hospitals, different medical schools
different results??
2-we cannot measure then: only know those who went to trial or were exposed
to the media, or discussed at M.and M rounds...most of them are anecdotes
informally talked about--
3-Do you remember the patient you sent home on Monday?..some times we don
t even get that information and our misdiagnosis goes us.
4-want to know if the investigations done at the present includes items as
age, grade. experience of the doctors.Found a CRICO report that mentioned
that 17% os complains in general surgery were due to residents failure to
diagnose postoperative complications .
5-cannot agree with Dr Graber when he says that faulty knowledge is uncommon
.Reports from MPS UK displays cases of obvious faulty knowledge: small
children do not have appendicitis, smallpox does not produce immuno
supression, etc.In my practice have wittnessed.the same mistake happening
once and again. Intususseption is diagnosed when the baby is bleeding, ,
Hirschsprung enterocolitis in practically unknow for most pediatricians.
More education is needed ..ask about inborn errors of metabolism,disorders
of the sexual development , pharmacogenetics and be surprised
Common conditions are also difficult to understand how
gout may be missinterpreted as an abcess(in the hand) or as flat feet!!

Thomas Tsai Replied at 12:15 PM, 28 Jan 2015

1) How should we start to conceptualize diagnostic error in order to compare quality across healthcare systems or providers? For example, many of the current AHRQ Patient Safety Indicators either focus on iatrogenic harm (eg. pneumothorax or accidedental puncture/laceration) or from outcomes indicators where the solutions are often standardization and following evidence-based protocols (e.g. checklists to decrease cather-associated infections). What would a PSI for dianostic error potentiall look like?

2) What are the right incentives to minimize diagnostic error? Should we start thinking about public reporting or even pay-for-performance based on dianostic error?

I understand these are both tough questions without easy or straight-forward answers, but interested in the thoughts of the panelists and of members from the GHD community!

Cheryl Lynne Galler Replied at 12:59 PM, 28 Jan 2015

Hi everyone. What I would like to say is that I believe we make a lot of errors due to the belief that a set of symptoms and signs are definitely a sign of condition X.

Cheryl Lynne Galler Replied at 1:08 PM, 28 Jan 2015

The reason I say this is because I had a patient who had been to several doctors and still had muscle pain in her foot. The muscle was pinching a nerve. So we surgically removed the cause of the nerve pressure and there was no improvement. We then decided to ignore the location and to evaluate systemic problems that could cause nerve pain. We ended up diagnosing early stage Parkinson disease. I think it would be a great game that could have a reward, for doctors who reported cases that were unusual in their end diagnosis because the symptoms lead them astray at first. It would teach others to look for the new combinations and to consider answers that are away from straight text book lists.

Mark Graber Replied at 7:25 PM, 28 Jan 2015

Trisha - Thanks for your question about identification problems. This is an issue in at least three settings:

Almost all healthcare institutions now insist on 2 forms of confirmation for medication administration, diagnostic procedures, and other situations. As an example, this has helped prevent the wrong patient from being sent down for a cardiac catheterization that was really intended for the patient in the next bed.
In the clinical laboratory and radiology: One of the more common errors encountered is that the wrong ID is applied to the blood sample or the film. This is one of those ‘human’ steps and is thus error prone. Bar coding of patients has dramatically reduced this problem for patients in the hospital.
In many electronic health record systems you select your patient from a very long electronic list, and its a little too easy to pick the wrong person. Incorporating the patient’s picture into the EMR is a helpful solution to this problem, as it offers a visual confirmation that you picked the right person.

In low resource settings, the fancier IT solutions will be less available, but the low-tech things like insisting on two forms of identification can still be used effectively.

Hardeep Singh Replied at 8:23 PM, 28 Jan 2015

Trisha, I would emphasize Mark's point that even with electronic solutions, we could face identification problems. Dean Sittig and I recently developed the "ONC SAFER Guides" for safe and effective EHR use and perhaps some of the items on the Patient Identification guide might be relevant here (see link).

There are additional SAFER Guides that others might find useful to address other diagnosis related issues such as test results and communication concerns that Urmimala Sarkar's study points to.

Attached resource:

Hardeep Singh Replied at 9:09 PM, 28 Jan 2015

Edita--thanks for your comment. You bring about complexities of measurement across settings and studies and yes, this is quite heterogenous. I refer you to the Zwaan research agenda and Graber incidence papers below for more details. However, the epidemiologic study of frequency we did (sent earlier) gets quite close to applying similar definition across US outpatient adult settings. We also report some MD demographics in the types/origins and confidence study sent earlier. Data from most countries (esp low resource settings) is lacking but I am aware of a large research project about to begin in UK. Overall, it would be fair to say that diagnosis concepts can be elusive and research in this area is still evolving as we have only recently begun to define most concepts rigorously.

Regarding knowledge, agree that there is some debate here. See for example a study Urmimala Sarkar and I did where PCPs endorsed knowledge gaps themselves.

Attached resources:

Hardeep Singh Replied at 11:10 PM, 28 Jan 2015

Thomas--regarding your question on what a future patient safety indicator for diagnostic error will look like, we have some ideas of conceptualization but not there yet. I think two areas that are fertile for further exploration are electronic detection of unexpected return visits that lead to hospitalization (see Singh BMJQS PC triggers paper where PPV was around 20%) and missed follow-up on abnormal findings (Murphy BMJQS Cancer triggers; PPVs upto 50% and don't miss Gordy Schiff's excellent editorial associated with it -Diagnosis and diagnostic errors: time for a new paradigm ).

Again, I must admit that work is still in progress, we need more indicators and for moving these indicators to practice we need to test them in other settings and make them more rigorous. Graber incidence paper above has a good overview of some methods, again mostly in research settings. In addition to better definitions and measurements, looking at clinical details in the record always takes time and we know how busy everyone is. All of these challenges need to be overcome for a good PSI.

Regarding are we ready for public reporting or even pay-for-performance based on diagnostic error--I strongly say no--not at this time. Unless the definition and measurement issues are clarified this would not be prudent. The IOM is working on a report in 2015 and I had the pleasure of presenting some of these measurement related issues to them (see Singh IOM Slides--specifically slide 13).

Nevertheless, measurement is ready for quality improvement (QI) purposes so we encourage everyone to start looking. Reflecting on Edita and Cheryl's comments, could we keep a better track of our own patients when there is some element of uncertainty involved or collect some of our own misdiagnosis for learning purposes and share with each other? Are others doing some of these QI type initiatives in their own practices? We would really like to hear from the group how you hear about your diagnostic mistakes, if at all, and how we can create some better feedback loops to frontline docs. Could we create better peer review systems so they could be leveraged for learning? See some of our ideas in the Peer Review paper attached. Dr. Graber's Jt comm article and a recent media article (in Moderhealthcare) nicely described some of Bob Trowbridge's work on championing and learning from reporting as well as some of the other issues discussed.

Would be great to hear thoughts from other members now that we have laid some foundational issues on the table.

Attached resources:

Urmimala Sarkar Replied at 1:50 AM, 29 Jan 2015

Today's question is what strategies improve diagnosis, and what is the evidence? Mark Graber has two review articles that I am sure he will add addressing this. For me it comes down to a single word: Feedback. What I see in common in most of the successful interventions is the opportunity for clinicians to learn from misdiagnoses. This happens naturally in settings where you see patients over time; less so in acute care settings. Peer learning opportunities like morbidity and mortality report, and informal dialog among colleagues with a learning culture are key.

How does diagnostic feedback happen where you work?

Trisha Finnegan Replied at 9:32 AM, 29 Jan 2015

Mark and Hardeep,
Thanks for your responses and dialogue on identification. I better understand current and best practice based on what you shared.

William Martinez Replied at 10:26 AM, 29 Jan 2015

As you said, I think a learning culture is key. Ultimately, many missed diagnoses are "caught" by colleagues within or outside the same institution. This sort of informal dialog between colleagues regarding a missed diagnosis seems uncommon. Worse sometimes rather than giving feedback, the physician or team that did not make the diagnosis is ridiculed behind their back. Establishing a culture where nonjudgmental feedback is expected and appreciated, and where ridiculing colleagues is rejected and unacceptable, would go a long way. The more it can be normalized the better. Would love to know if anyone has taken on this culture piece with regard to missed diagnoses or ideas about how to address it?

In my own practice (general internal medicine), I get notifications when my patients are readmitted within 30 days, are seen in the ED, if it was outpatient that they were admitted. These notifications always prompt me to review their care and try to understand what could have been done differently or what was missed. The hit rate on these for identifying diagnostic errors seems low though and I think the data on these triggers bares that out as well. So need more sensitive and specific triggers perhaps and additional methods of identifying diagnostic errors, then an established mechanism by which all identified cases are reviewed to establish "true positives" and then feedback to clinicians provided along with tracking for patterns and clinician outliers.

Gurpreet Dhaliwal Replied at 4:32 PM, 29 Jan 2015

I fully agree with Dr. Sarkar's and Dr. Martinez’s point about a learning culture. Diagnostic errors are often nuanced and context-dependent, and are best adjudicated through conversations. But a learning culture must be in place in order for those discussions to happen regularly and productively.

Culture change falls to the clinical microsystem leaders who must work toward normalizing frequent discussion among providers about each other’s work. Potential steps in this direction would be standing staff meetings that discuss real-world, local cases. Or encouraging routine secure electronic messaging between providers (e.g., a hospitalist group that always lets the emergency physicians know how a patient did, or an emergency medicine group that always notifies the hospitalist or PMD of a return visit).

Hardeep Singh Replied at 9:12 PM, 29 Jan 2015

The two papers Urmimala referred to that discuss evidence of interventions are enclosed. Bottom line--lots of talk about interventions in the literature but little action related to testing them. Its important to test these interventions because many interventions thought to be "game-changers" are later found to have unintended consequences. Nevertheless many actions to reduce diagnostic errors are the right things to do such as spending time with sick patients (and their data) to understand the clinical story and learning from mistakes to become better physicians.
Also, we need to have health care systems that support this learning culture.

Attached resources:

Mark Graber Replied at 10:51 PM, 29 Jan 2015

On the topic of patient safety indicators, I have to agree with Hardeep that we are NOT ready for public reporting or PFP.

On the other hand, there are MANY measures related to structure and process that should correlate with diagnostic quality, and I would advise healthcare organizations to consider some of these:

HCO’s can FIND diagnostic errors
HCO’s can use root cause analysis to promote understanding of diagnostic errors
HCO’s provide ways to learn about diagnostic error through FEEDBACK (Urmimala made this same point, and its a good one!)
HCO’s provide decision support software that supports diagnostic quality
HCO’s ensure that expertise is available when its needed (e.g., Radiology coverage on nights and weekends to read diagnostic imaging studies)
HCO’s have electronic systems in place to ensure critical diagnostic test results are addressed in a timely manner
HCO’s monitor the quality of diagnostic testing services from the clinical lab and Radiology
HCO’s encourage and facilitate second opinions

Hardeep Singh Replied at 11:36 PM, 29 Jan 2015

Mark mentions a good point about use of second opinions but docs often don't know when to seek one, and patients might not necessarily seek second opinions on their own. We sometimes have difficulty dealing with and communicating uncertainty about a diagnosis and don't realize we need help. Nevertheless, for second opinion on a difficult to diagnose condition, do you consult a computer? Personally, I would consult Dr. Dhaliwal :-) See a great write up in New York Times on diagnostic decision support tools and diagnostic excellence.

In general, on the topic of electronic health records and information technologies, we have yet to reach the potential. The Ebola misdiagnosis paper sent earlier discussed some complex human-computer interaction issues. As we move to more digital data, opportunities for improvement are plenty though. The two papers enclosed are excellent reads on this topic and discuss the breadth of diagnosis related issues in context of health IT.

Attached resources:

A/Prof. Terry HANNAN Replied at 9:38 PM, 30 Jan 2015

Dean I really like this staement. "Therefore, my take on the problem of misdiagnosis focuses mostly on how the EHR either helps or hinders the diagnostic process. This often takes the form of lost patient information". It reminded me of a common error in eHealth design (EHRs) which is where the e-system is almost a reproduction of the paper record. Therefore reproducing many of its limitations.
I agree we have a long way to go to improve these systems.
In addition to a reply to Dean's posting I would like to say thank you to all the other panellists who have expanded my learning on all the topics listed. Your contributions to the GHDonline community are invaluable.

Hardeep Singh Replied at 12:13 AM, 31 Jan 2015

Thanks to all--its been a great week of discussion. I'd like to add a couple of additional resources. One, a paper about patients role in the diagnostic process, and the other an editorial that summarizes the 2013 BMJQS Supplement on Diagnostic Error which had 10 papers on the topic. All papers are linked within the editorial and are open access.

We hope to see some of you at the AHRQ sponsored 8th Diagnostic Error in Medicine conference in Washington DC on Sept 27-29 -- the conference website will be linked later through the SIDM website so please be sure to check:

Attached resources:

Jessica Ludvigsen Replied at 9:35 AM, 2 Feb 2015

Many, many thanks to our exceptional panelists, and all of our community members who participated in this incredibly rich discussion. We greatly appreciate the insights everyone has shared, and look forward to continuing to discuss these important topics in 2015.

We will be working on a Discussion Brief to summarize the key points from this Expert Panel, and will share details as soon as that is available on the website.

In the meantime, we would be grateful for your feedback in our short, 9 question follow up survey. These surveys help us understand the impact of our Expert Panel discussions, and your feedback is incredibly valuable to us. Please take the survey now, by visiting:

Sudip Bhandari Replied at 9:12 AM, 20 Apr 2015

Many thanks again to our panelists and members who participated in this Expert Panel discussion last January!

To help us understand the longer-term impact of these Expert Panels and plan future events, we have created a very short, 5 question, follow-up survey. This survey will only take 2-3 minutes of your time—please take the survey now at:

Many thanks,

Leonardo Leonidas Replied at 10:03 AM, 20 Apr 2015


I hope you will find my 10 Commandments of help:

* Ten Commandments to Reduce *

* Diagnostic Errors*

1. Thou shalt First “Do No Harm.”

2. Thou shalt think of serious and treatable conditions and act on
them without delay.

3. Thou shalt remember that Diagnosis is History, History, History.
Then confirm with clinical examination and more History.

4. Thou shalt request a test only if it will change your plan or help
in predicting the outcome.

5. Thou shalt question “authority” such as your senior residents,
consultants, experts, or even National guidelines.

6. Thou shalt continue the debate and questioning even though the
data is “IN.”

7. Thou shalt maintain a high index of suspicion for uncommon
presentations of the common.

8. Thou shalt recognize your own beliefs, biases, prejudices, and
thinking style.

9. Thou shalt be wary of your hunches and intuitions. It is better to
use Evidence Based Medicine.

10. Thou shalt have an iPad* or a smartphone in your palm.

Given to his Son Len and Class 2001 Tufts University School of

From Dr. Leonardo Leonidas, Bangor, Maine 20 May 2001 Copyright

*Palm Pilot in the first edition.

Leonardo L. Leonidas, MD

Assistant Clinical Professor in Pediatrics (retired 2008)

Distinguished Career Teaching Award, 2009

Tufts University School of Medicine, Boston, USA

This Expert Panel is Archived.

While this Expert Panel is no longer active, we invite you to review and recommend past replies and resources. Membership for this Expert Panel is closed, but we hope you'll join us in one of the many communities on GHDonline.

Panelists of Diagnostic Error Reduction and GHDonline staff