CEO SUMMARY: Researchers at the Mayo Clinic showed that only 12% of patients referred to Mayo physicians for a second opinion got a confirmation that their original diagnosis was complete and correct. In 21% of the cases, the diagnosis was completely changed. Among patients who got additional work ups for a second opinion, some 80% to 94% of them got additional lab tests. The research also showed that pathologists and radiologists take steps to ensure diagnostic accuracy by confirming findings with colleagues.
DIAGNOSTIC ACCURACY is once again in the news. This time, a research study performed at the Mayo Clinic in Rochester, Minn., determined that, among other findings, in a sample of 286 patients referred to the clinic, 21% had their diagnosis completely changed.
This and other findings from the study have important implications for pathologists and lab administrators. It is evidence of the substantial value that laboratories could contribute by collaborating more closely with clinicians to improve diagnostic accuracy and reduce or eliminate recurring sources of diagnostic errors.
“One of the most challenging errors to improve in healthcare is diagnostic accuracy,” commented James Naessens, ScD, a health policy researcher with the Mayo Clinic. One of the authors of a recently-published article on misdiagnosis, Naessens’ research showed that only 12% of the patients who were part of a study got a confirmation that their original diagnosis was complete and correct.
One significant finding of the study is that more than one in five patients (21%) got a completely new diagnosis. As well, in 66% of the cases, the patients received a refined or redefined diagnosis.
Previous research on diagnostic errors shows that these mistakes contribute to about 10% of patients’ deaths and account for 6% to 17% of adverse events in hospitals.
An important finding from the research—and this finding was not in the published paper—is that during the 30 days when a patient was getting a second opinion at Mayo, some 80% to 94% of those patients got additional medical laboratory tests, Naessens said in an interview with THE DARK REPORT. “About 55% to 67% of these patients also got X-rays or other radiology exams, and 20% to 40% got CT scans,” he added.
“Clearly, these ancillary services were needed to supplement the information the Mayo Clinic providers had gathered initially to figure out what was going on with these patients,” explained Naessens. “To do that, our laboratory and radiology specialists were brought into play on the teams that were addressing the patients’ complaints.
“The point is that these patients not only got a second opinion, but also there was a rereading of the medical records, face-to-face meetings with patients, and there was a need for the additional work up of lab and imaging tests,” he added. “That additional information helped to change the diagnosis.”
The findings from Naessens’ research were published online on April 4 in the Journal of Evaluation in Clinical Practice. The paper is titled, “Extent of Diagnostic Agreement among Medical Referrals.”
Value In Second Opinions
For years, the patient safety movement has put an emphasis on the value of getting a second opinion. Now Naessens’ research has confirmed the need for a confirmatory diagnosis because, as the research shows, physicians don’t always have the answers. Often, unusual or complex symptoms will lead a physician to recommend a second opinion, or a patient may request one. Regularly, Naessens found, physicians rendering second opinions find that the original diagnosis was in error.
For the research, Naessens and colleagues examined the records of 286 patients referred from primary care providers to Mayo Clinic’s General Internal Medicine Division between Jan. 1, 2009, and Dec. 31, 2010. The referring diagnosis was compared to the final diagnosis.
The original diagnosis was confirmed in only 12% of the cases (34). In 21% of the cases (60), the diagnosis was completely changed; and in 66% of the cases (189), the patients received a refined or redefined diagnosis.
“Effective and efficient treatment depends on the right diagnosis,” Naessens said in an article published by the Mayo Clinic news department. “Knowing that more than 1 out of every 5 referral patients may be completely [and] incorrectly diagnosed is troubling, not only because of the safety risks for these patients prior to correct diagnosis, but also because of the patients we assume are not being referred at all.”
The results of Naessens’ research align with other research on diagnostic errors. Further, as Lenny Bernstein reported in The Washington Post, Naessens’ work provides evidence that the healthcare system still could improve its level of diagnostic accuracy.
For the Post’s article, Bernstein quoted Mark L. Graber, MD, the founder of the Society to Improve Diagnosis in Medicine, who said, “Diagnosis is extremely hard. There are 10,000 diseases and only 200 to 300 symptoms.”
In the article published in the Journal of Evaluation in Clinical Practice, Naessens wrote, “Unlike tangible errors involving systems and processes—such as medication administration errors, prescription errors, and wrong site surgery—diagnostic errors have avoided the spotlight because they are less easily understood, not viewed as a system problem, and not perceived as problematic by physicians.”
In September 2015, the National Academy of Medicine issued a report on diagnostic errors, “Improve Diagnosis in Health Care.” The report was a continuation of the landmark Institute of Medicine reports: “To Err is Human: Building a Safer Health System” (2000) and “Crossing the Quality Chasm: A New Health System for the 21st Century (2001).”
Inattention To Errors
In the academy’s report, experts showed that the occurrence of diagnostic errors has largely been unappreciated among many other efforts designed to improve healthcare quality and patient safety. “The result of this inattention is significant: the committee concluded that most people will experience at least one diagnostic error in their lifetime, sometimes with devastating consequences,” the academy said.
“Improving diagnosis will require collaboration and a widespread commitment to change among healthcare professionals, healthcare organizations, patients and their families, researchers, and policy makers,” commented the academy.
This point is one Naessens made in his interview with THE DARK REPORT. Pathologists and clinical laboratory scientists have an opportunity and a responsibility to work closely with other healthcare providers to ensure that patients get the correct diagnosis, he said. Also, he confirmed another point from the academy’s report: diagnostic error is a relatively under-measured and understudied aspect of patient safety.
A Call For More Research
“Diagnostic error is an area where we need more research and more information,” Naessens said. “I’ve been working in the quality measurement field for the last two decades here at Mayo, and diagnostic error is one of the areas that hasn’t gotten much attention, partly because it’s hard to measure. And, of course, we tend to focus on what you can measure.
“Yet, it’s an important area that underlies the issue of whether a patient is getting the right treatment,” he continued. “If we are going to determine if patients are getting the right treatment, then the first step is to ensure that they get the right diagnosis. If you get the wrong diagnosis you’re certainly not going to get the right treatment.”
Research published by Elizabeth McGlynn and others in 2003 in the New England Journal of Medicine showed that only 54.9% of the participants in the study received the recommended standard of care for their condition. (See TDR, July 7, 2003.)
All of which may indicate the need for systemic changes. “One of the things we wanted to point out in the study is that there seems to be an attempt by health insurers seeking to reduce costs to use narrow networks,” he said. “When payers do that, some of those narrow networks might put burdens on patients or they may create barriers for patients seeking to get second opinions or to get another work up for their conditions.
“In those instances when a provider has some doubt, there should be a way for those patients to get further information and second opinions,” he said. “Any restrictions on the ability of patients to get the additional information they need could be detrimental to diagnostic accuracy and to their health.”
Understanding The Findings From Research Study Of Diagnostic Errors At The Mayo Clinic
RESEARCHERS AT THE MAYO CLINIC UNDERTOOK A STUDY TO DETERMINE the number and types of diagnostic errors seen in patients referred to the clinic. The sample was 286 patients who had been referred to the Mayo Clinic General Internal Medicine by primary care practices from January 1, 2009 to December 31, 2010.
The study was published online on April 4 in the Journal of Evaluation in Clinical Practice, with the title, “Extent of Diagnostic Agreement among Medical Referrals.” The table below shows the categories of referral and final diagnosis. Researchers determined that in 12% (36/286) of cases, final diagnoses confirmed the diagnoses presented at the referral. Final diagnoses were better defined/refined in 66% (188/286) of cases. However, in 21% of cases (62/286) final diagnoses were distinctly different than referral diagnoses.
Examples Of Referral And Final Diagnosis By Categories
One In Five Patients Affected
“Primary care providers should be able to make sure their patients get any additional diagnostic work up when they feel that they don’t have all the expertise and they need to call in additional help,” he added. “Not every primary care visit has a 20% error rate, but the 20% rate means there should be plenty of opportunity for patients to get a confirmatory diagnosis. In fact, for some subgroups of patients, the rate could be higher than 20%.”
One of the most surprising aspects of the research, Naessens said, was that he did not think the rate of changed diagnoses would be as high as it was: 21%. “I thought it would be about 5% or maybe 10%. I wasn’t so surprised by that middle category, meaning where the diagnosis was refined or redefined, because we expected that a number of people were going to be referred as patients who had complex symptoms. For those patients, physicians often don’t know what’s going on and so, to figure it out, they need to work with experts in diagnosis such as we have here at the Mayo Clinic.
“The other surprising result was that there was no real concentration of problems in any particular diagnosis category except maybe for musculoskeletal issues,” added Naessens. “These patients were coming to Mayo from different sorts of areas and yet there were no particular diagnostic areas that stood out. They were all well represented.”
Model Of Collaboration
For clinical lab scientists and pathologists, Naessens commented that they have a model of collaboration designed to eliminate diagnostic errors. “We use laboratory and pathology as examples of how they do a lot of double checking on their work and rereading of pathology slides,” he said. “This method of operating your practice can go a long way toward minimizing errors.
“In part, that’s why we suggest that other providers should use pathology and radiology as examples of how secondary reviews are beneficial, from a learning perspective, to improve your approach in selected cases to the point where it would be beneficial to patients,” he added. “More of medicine should follow their lead. But physicians should also seek out pathologists and radiologists for second opinions, especially when the original physician is in doubt.
“In other words, physicians should not be afraid to work in teams because those different perspectives bring value, particularly in the more complex situations,” concluded Naessens.
Study Of Diagnostic Errors
This study of diagnostic errors shows the growing interest some clinicians have in measuring the actual rate of errors in medical care. Pathologists and lab administrators may want to collaborate with physicians in their hospitals to conduct similar studies.
Contact James Naessens at 507-284-5005 or firstname.lastname@example.org.