CureMD: Please give us an overview of your analysis report.
Dr. Singh: As the case details unfolded, it appeared to be a case of misdiagnosis. In our work we define misdiagnosis to occur when there are clear missed opportunities to make the correct and timely diagnosis. Our analysis focused on the complexity of making diagnosis using EHRs and how sometimes technology might get in the way of decision-making.
CureMD: As mentioned in your report, “EHRs are only tools and not a replacement for basic history taking, examination skills, and critical thinking.” In your opinion, to what extent can physicians rely on technology for accurate diagnosis of patients?
Dr. Singh: Well, we used to make mistakes in the paper-based system too, mostly when information was not available, or was not communicated, or we didn’t have the decision support tools to help us know what to do next. So I think the EHR remains an important tool to facilitate diagnosis.
However, while we’re implementing EHRs across U.S healthcare; the technology is still young and it has room to evolve. Physicians are spending more time interacting with the computer than interacting with the patient and that result in new types of hazards. There is evidence of the fact that the (Texas Health Presbyterian Hospital) ER physician was using templates which led to contradictory information in the note (example fever versus no fever). This may have been the reason that there were some initial discrepancies about how stable the patient was and what happened in the emergency room.
This is a great example of human error!
CureMD: A lot of emphasis has been given to the communication of information between physicians and staff for better patient care. However, there had been gap of communication between the nurse and ER physician in case of Eric Duncan. Where do you think the fault lies?
Dr. Singh: This is great example of human error. Clinicians are responsible for basic history-taking, exam, critical thinking, communication, and making judgments on the case. However, we can no longer separate technology out of the equation when a physician is working in a technology supported healthcare environment and making diagnostic decisions. So, the Electronic Health Record has to support this process in every way possible.
This is not only a coordination and EHR issue, but also brings to light many health policy and workflow issues.
In an ER, time is of essence, and there is quite a bit of chaos with large number of patients with multiple medical issues; so we want EHRs to provide cognitive support to physicians in these situations. Current EHRs lack innovation to do that. This (Eric Duncan case) is not only a coordination and EHR issue, but also brings to light many health policy and workflow issues. The ER physician was using a lot of templates and spending a lot of time with computer because we, physicians, are trying to meet documentation and billing requirements and ensure that we meet quality measures. For example, while documentation should ideally support decisions and diagnosis, it can get in the way when notes are so long and important information is lost. EHRs were implemented with the goal to make things better and safer. So, technology would ultimately have to evolve to support the diagnostic process but other non-technical factors that I mentioned will also need to change. It’s not one thing we can hold responsible and fix–it all needs to be addressed.
Dr.Sittig: This was a human challenge. The physicians were not prohibited to ask the same questions again. It’s not like the physician would have only received the information from listening to the nurse or reading her notes. But now physicians are relying more on computers for patient information.
CureMD: There were apparent red flags like “a black man with a foreign accent who travelled to Africa and has flu-like condition and fever.” However, the nurse failed to clearly identify symptoms of Ebola. In this situation, do you think the misdiagnosis would have been avoided, had the nurse identified the symptoms?
Dr. Singh: We relied on publicly available information to do the analysis. We looked at the information released by the Associated Press (AP) - who had medical records - and the timeline released by the hospital (Texas Health Presbyterian Hospital). Based on the available information, it’s clear that the hospital did receive information about Ebola. What they did with the information is unclear. Retrospectively yes it could have been avoided if the nurse or physician would have made the link. But remember that this is not the first case of misdiagnosis. As we mention in the paper, misdiagnosis has been around for a long time and only now getting public attention. Our research shows that about one in twenty adults can have a misdiagnosis every year in the U.S. This case got the attention because there are huge public health challenges associated with it. If there was more transparency in terms of case details and what happened in the hospital, then we would have been able to give you a better answer. We need to learn from these mistakes.
Dr. Sittig: Someone should do an independent investigation to examine what happened and publish a report describing their findings.
CureMD: EHRs are not designed to identify extra ordinary health epidemic or disaster like Ebola. Do you think the federal government should introduce guidelines or specifications to develop EHRs that can cater to public health emergencies?
Dr.Sittig: That would be a great thing for the government to do. There is still a lot of work to be done. That would take a tremendous amount of research and development effort, but that would be really good thing for the CDC to undertake and for the informatics community to work on with the EHR vendors.
People have put forward several ideas for recording patient history, like where they’ve been, and how you can code diseases that are endemic in different regions of the world. This can be a great research project, but I don’t think we are at the point where we can make a system like that right now.
We need a supportive socio-technical system to improve care, and EHRs are a part of the system.
CureMD: Unless we have a system like that we would have to rely on physicians’ critical thinking and judgment?
Dr. Singh: Oh yes, we will always need to rely on that (critical thinking). We are missing important common stuff right now. Ebola case was a rare situation; even though there were some red flags that suggested it. But we miss red flags even in situations related to diagnosis of cancer, cardiovascular conditions, and other infections. Thus, we are missing common conditions already. We need those basic skills but we also need a supportive socio-technical system to improve care, and EHRs are a part of the system.
Dr. Sitting: We will have to continue to reply on human thinking for quite some time.
CureMD: So are you saying technology is there to improve the healthcare system, but it’s not entirely responsible for bringing the complete change?
Dr. Singh: There are many other factors than technology, such as processes, policies, procedures, workflow, communication, external health rules and regulations all of which need to be considered. So, it’s not only the EHRs that have to be right, but the rest of the system and processes also have to be right. That’s why Dean and I advocate for a socio-technical system approach.
Dr. Sitting: As much as we would like computers to help us take care of patients and think for us, it is a long shot. There are three phases of work that need to be done: the first phase is to ensure that technology is working the way it’s supposed to, second phase is that people would have to use the technology appropriately and the third phase is to get the technology to help us monitor and prevent errors. Everyone wants to jump to the third phase and wants the computer to make a diagnosis, do the job for us. But we can’t skip the first two phases.
Dr. Singh: It’s important to look at these three phases differently. We are having conflicting media reports where the blame keeps shifting. People need to understand that there are many things beyond just getting the technology right. You could have the best Electronic Health Record system in the world, but if we don’t use it effectively then we don’t get better outcomes.
CureMD: In the end, I would like you to suggest few measures that would help make accurate and timely diagnosis of endemic diseases.
Dr. Sitting: While facing this deadly disease (Ebola), there are some questions that every hospital and practice staff should ask the patients. Have you been to West Africa recently? Have you been in contact with someone you think may have Ebola? These are some important protocols that should be added in the Electronic Health Records to make sure that every patient is appropriately screened.
In terms of policy, when the hospital staff communicates information to someone, they should ensure that the message has been received.