Description
In this episode of CIPI Connections, members of the CAP Informatics Committee, Alexis Carter, MD, FCAP, and Omar Baba, MD, share a gripping cautionary tale of unflagged abnormal laboratory values and the informatics solutions designed to avert similar pitfalls.
Council on Informatics and Pathology Innovation (CIPI)
Subscribe to CIPI Connections on Apple Podcasts, Spotify, or wherever you listen to podcasts.
Hosted on Ausha. See ausha.co/privacy-policy for more information.
Transcription
Welcome to CIPI Connections, the podcast of the College of American Pathologists Council on Informatics and Pathology Innovation. Here we connect you with the leaders and committees shaping the future of pathology. I'm Dr. M. E. de Baca, Chair of the College of American Pathologists Council on Informatics and Pathology Innovation, also known as CIPI. Today is the beginning of something brand new. The Informatics Committee will be bringing you horror stories in pathology informatics in multiple episodes. Today, Dr. Alexis Carter, a molecular pathologist and informatician, is joined by Dr. Omar Baba, the Informatics Committee Junior Member and a Pathology Informatics Fellow. Since it's the first in a series, Dr. Carter will initially share some housekeeping details about the series, and then she and Dr. Baba will be discussing the case from the lab. Take it away, Dr. Carter.
Hi everyone, welcome to Horror Stories and Pathology Informatics, Lessons Learned from Things Gone Awry, a series to help your healthcare organization, number one, keep your patients safe, and number two, avoid painful problems. I'm your host, Dr. Alexis Carter, a pathologist and the moderator for this series, and with me today is another pathologist, Dr. Omar Baba. As a reminder, these podcasts are available from your favorite podcast service under CIPI Connections, that's C-I-P-I Connections, and is also from the College of American Pathologists website at www.cap.org. Today's episode is Unflagged and Overlooked Lessons from a Missed Cortisol Result. As always, there are a few disclaimers needed before we get started. All of the situations discussed in this podcast are based on real events, so if any of our listeners think that the situation described can't or wouldn't happen, we can assure you that it can and did. Having said that, all information that could identify the personnel or the healthcare organization involved have been removed and replaced with fictitious characters and a fictitious location of Cabot Cove Memorial Hospital. Any locale-specific aspects? for example, regulations, that are more specific than the national United States level have been removed. And some details regarding the situation which do not affect the main point of the event have been changed. Finally, the person presenting the adverse event is not the person who contributed it for discussion. All of these things were done to ensure that we focus on the lessons learned and so that we can help others avoid these mistakes. Of course, this podcast does not represent legal or medical advice, and the lessons learned may not account for specific barriers that may be present at your own organization. All right, so now that we've done our disclaimers, let's get started. Again, here with me is Dr. Baba, who is going to present today's healthcare software failure on flagging of abnormal laboratory values and the lessons learned from it. So, Dr. Baba, let's start with the basics of the adverse event that we're going to discuss today. What happened?
So, our story today is about a 60-year-old male who was undergoing treatment at Cabot Cole Memorial Hospital for prostate cancer. After the initial radiotherapy and hormone treatment, the patient was scheduled for radical prostatectomy, which is a surgery to remove the prostate gland, and he was examined in the surgery clinic for preoperative testing. Now, during the pre-surgical appointment, the patient had a battery of tests performed among which was a serum cortisol level. Now cortisol is the body's main stress hormone produced by the adrenal glands. It turned out that in this patient the cortisol level was low indicating adrenal insufficiency which is a condition where the adrenal glands are not making enough cortisol. The pre-surgical labs were reviewed by the clinician but the low cortisol level was not identified in that encounter despite being there in the electronic health record. The patient was referred to the emergency department later that week for nausea, vomiting and abdominal pain, but the patient's low cortisol was missed too by the emergency care team and the patient's adrenal insufficiency was not diagnosed. Fortunately, the patient's surgery was postponed, allowing the patient to recover from a presumed viral gastroenteritis. Now, had the surgery not been postponed, the patient, undiagnosed, could have quickly progressed to adrenal crisis, which is a life-threatening emergency marked by severe blood pressure, shock, and even death without urgent treatment. One week after the initial cortisol lab test, the patient contacted the after-hours nursing service advice line to report increased dizziness and lightheadedness when standing up. The patient was told to follow up with their cardiologist in the morning. The next day, the cardiologist finally noted the low cortisol levels and referred the patient to the emergency department. The oncologist was made aware of the diagnosis of adrenal insufficiency. and the patient was treated accordingly, recovered, and successfully underwent the planned prostatectomy one month later.
Okay, wow. So what was the fundamental contributing factor to this particular adverse event?
So in this case, the low cortisol value was missed by the providers because there was no abnormal flag associated with the result. So an abnormal flag is a visual cue in the electronic medical record, or EMR, which is a patient chart basically, like a mark-colored highlight that alerts a provider to a result that requires attention. The providers have got used to relying on the abnormal flag to screen for laboratory values they should pay particular attention to which is common in busy clinical workflows. Cortisol has a diurnal reference range that is levels change throughout the day so levels should be interpreted depending on whether the sample was drawn in the morning or in the afternoon. For instance in morning samples the range is 5 to 23 micrograms per deciliter whereas for the afternoon samples it's 3 to 16 micrograms per deciliter. As a result, many laboratories display both reference ranges in the interpretive comment, which is a free text note that sits alongside the lab result. This, of course, does not allow for flagging of cortisol levels that are abnormal or outside of the reference range.
So with regard to that issue, could this issue have been completely avoided or just mitigated?
This issue definitely could have been avoided, and there are several informatics strategies for accomplishing this.
Okay, so what do you think could have been done to avoid or mitigate the fundamental contributing factor in this case?
This issue could have been avoided by either creating two separate orders for morning and afternoon levels of cortisol, which would lead to having two orders, morning cortisol and afternoon cortisol, making expected timing explicit. Another strategy is to include what we refer to an ask-on-order entry question when the clinicians place the order. The Ask on Order entry is a prompt asking the ordering clinician to specify timing, that is, morning or afternoon, which would trigger the correct reference range and abnormal flag. The system would indicate in this case which reference range should be displayed, thereby ensuring that there is always a discrete reference range and subsequently an abnormal flag when applicable. Alternatively, the nurse could also indicate whether the sample was collected in the morning or the afternoon, as part of the collection process using a different informatics configuration that would accomplish the same thing as the ask on order entry question or separating the morning and afternoon orders. This would also be beneficial to other laboratory tests that may have different reference ranges depending on the timing of the specimen collection, such as some antibiotics like vancomycin and gentamicin, which have different reference ranges for peak, that is right after dosing, or trough, just before the next dose.
Okay, so we've talked about the main contributing factor, but there are always secondary and even sometimes tertiary factors that can contribute to these issues. These can include inhibited communication due to poor relationships, you know, silos between different groups, such as clinicians and laboratorians in this case, or other systemic infrastructural issues. What secondary and tertiary issues do you think contributed to this adverse event and or any delays in its correction?
So one issue that contributed to this adverse event that we cannot always address with an informatics solution is the reliance on visual abnormal flags by providers as screening tools to quickly review laboratory results. Using flags as a shortcut to triage large result sets can be a risk. Another contributing factor is displaying the reference range in the interpretive comment rather than in a more structured format. Structured formats allow systems to act on the reference range whereas free text comments do not.
So with regard to providers' reliance on abnormal flags, which, you know, is understandable, do you think, what do you think could have been done to prevent it from happening?
I think the message has to come down from the hospital leadership that providers must carefully review all laboratory tests and consider the reason why they are ordering these tests. So every test result deserves a deliberate review and should be considered in light of the clinical question for that patient. It's not always possible or appropriate to flag a result as abnormal, especially for narrative or complex results. Narrative results, for example, cannot always be distilled down into a normal or abnormal binary. Sometimes a normal value is significant if it reflects a notable change or trend. On another note, a provider can miss a piece of critical information if they get distracted by exclusively paying close attention to results that are flagged as abnormal. If a test is ordered, it should have been ordered for a reason and deserves the full attention of the provider to answer the clinical question they were asking.
Yeah, those are definitely really good points. Let's go on to examine the use of displaying reference ranges in the interpretive comment. What do you think could have been done to prevent that from happening?
So displaying reference ranges in the interpretive comment can be appropriate, especially when there could be multiple possible interpretations for a value depending on different clinical variables that are not always easily captured with the ask on order entry questions. Sometimes it's not possible to easily capture complex clinical contexts, so a brief explanation can help clinicians interpret lab values. One example might be the absolute reticulocyte count in a CBC, where the expected count differs in anemia, hemolysis, blood loss, etc., so a comment can help tailor interpretation. It would be appropriate to display the reference range or expected value for these different clinical scenarios in order to allow the provider to correctly interpret the result for the patient's clinical condition. While there are definitely situations when using the interpretive comment to display reference ranges as appropriate, this practice should be avoided whenever possible. Ideally, structured fields should accompany results to allow downstream systems to interpret and display them correctly. At the very least, there should be a primary reference range that's displayed with the result in a more structured information. One of the dangers of only displaying the reference range is the interpretive comment is that the result value and the reference range may become uncoupled in downstream transmission to other healthcare institutions, public health agencies, or disease registries. When results move between systems, comments may be dropped and values lose their context. We have a lot of experience transmitting reference ranges in a structured format because it's a required data element per the Clinical Laboratory Improvement Amendments or CLIA. So there are US laboratory regulations that require reference ranges to accompany results. Messaging standards such as HL7 have a dedicated field for transmitting the reference range in a structured format. This ensures that the reference range will be displayed with the result value. Messaging standards can accommodate interpretive comments, but there is less certainty they will always be displayed by the receiving system. Comments may not always show up correctly when data are transmitted. This increases the likelihood that the result value could be displayed without the necessary contextual information to help interpret the result.
So, Dr. Baba, you know, there's a couple of other things that I'm thinking about, and I'm wondering if we can talk about those. So, you know, for example, many people may not realize, especially if they are not pathologists or working in the lab, that we have some restrictions on reference ranges. and that really the reference ranges drive whether or not an abnormal flag can be present, right? The abnormal flag displays when something is outside of its normal reference range automatically, right? And we've got, you know, high, low, abnormal, you know, critically high, critically low flags, and those are all automatically appended based on reference ranges. You know, so what are some additional things, you know, restrictions? So like with this cortisol all result, for example, you know, we typically can do reference ranges based on age and gender, but this case is special for a couple of reasons, right? Because it demonstrates something that we can't easily have automated flags or reference ranges set for. So do you have a couple of examples in your own experience where that might have happened?
One example I could think of is different reference ranges for pregnancy, you know, so collection time of day, pregnancy. So when the clinicians understand the limitations of lab systems, I think they become more lenient, I would say, in cutting us a break about displaying abnormal flags. Because what is considered abnormal, really, right? Like some pathology reports, genetic reports, these are mostly for the majority of the cases, they're like kind of free text. So how do you flag these as abnormal?
Right. And then how abnormal does it have to be in order to call it abnormal? You know, if you have a breast biopsy and what you see is a fibroadenoma, you know, that's... benign, but no one would classify it as normal. So, you know, do you flag those, you know, and flags can also, they can, you know, alert physicians sometimes that things may be worse than they actually are. You know, and the other thing is like, you know, we could potentially in the lab look at a collection date and time. We get collection dates and times on all samples, right? But we're also required to perform the order that the clinician gives us. And so having lab staff switching a lab order to cortisol morning, for example, for specimens that are collected during certain hours, as opposed to cortisol afternoon would require some significant operational changes that may not be necessarily helpful or doable in the lab. In addition to that, we don't always know. I mean, patients don't always know when they're pregnant, right? And, you know, their reference ranges can change, you know, fairly quickly sometimes after they become pregnant. And, you know, as a sort of amusing or not so amusing side note, you know, we had a situation in an organization that I worked in at one point where when we put into the interpretive comment for pregnant patients, this is the reference range. And we actually did have a patient who called her clinician very upset because she thought that the thyroid test result that we were reporting on was indicating that she was actually pregnant because of her value, which the, you know, the clinician explained to her and it was all fine. But, you know, so it's interesting trying to cover all the bases, you know, can be difficult. And, you know, we also want to say, yes, we do believe, you know, clinicians should look at every result, but we also understand that our clinician patient facing friends are just as busy. as we are. And it's hard. It's hard to do that. But on the flip side, you know, when you have these special situations, it's kind of hard to make everything automated. So this has been really educational. Let's sum it up for the listeners. What's the main lessons learned from this incident? And what do you recommend that our listeners should do to prevent this from happening?
So in this case, there was a delay in diagnosis for a patient with adrenal insufficiency due to a missed cortisol result. The low cortisol level was missed because there was no abnormal flag associated with the result. This happened because the laboratory had used the interpretive comment in free text rather than a dedicated discrete numerical field to display two possible reference regions for the cortisol depending on the timing of the system and draw. So comments can educate and add nuance but the efficient machine readable reference should be in a structured field. Whenever possible there should always be a reference range displayed in a structured format that allows for abnormal flagging. This can be accomplished by either creating separate orders for the different draw times or using ask and unorder entry prompts or draw timed questions for the nurse at the time of the collection to allow for display of the appropriate reference range. And lastly, providers need to be educated on the importance of reviewing all results for the test they order and Instead of relying on screening systems to not miss essential pieces of information in the care of patients, again, every ordered result should be reviewed, not just flagged ones.
Great. Thanks so much, Dr. Baba, for presenting today. For our listeners, we hope that you have found this podcast on unflagged and overlooked lessons from a missed cortisol result helpful. This podcast was produced by the College of American Pathologists, and the content was produced by the College of American Pathologists Informatics Committee. The Informatics Committee always welcomes questions about the podcast, as well as suggestions for future podcasts at informatics@CAP.org. If you would like to contribute an issue that happened to you, please contact the Informatics Committee again at informatics at CAP.org for instructions on how to anonymously contribute an educational issue and its lessons learned. Please do not spend specifics on the issue to this email address. We thank you so much for listening, and we look forward to sharing our next podcast with you soon. Bye-bye.
Thanks to Dr. Carter and Dr. Baba for sharing these important lessons with us. Stay tuned for future horror stories in pathology informatics. Thanks for joining us for insights, updates, and the people behind the innovation. This has been CIPI Connections, where ideas meet action in pathology.
Description
In this episode of CIPI Connections, members of the CAP Informatics Committee, Alexis Carter, MD, FCAP, and Omar Baba, MD, share a gripping cautionary tale of unflagged abnormal laboratory values and the informatics solutions designed to avert similar pitfalls.
Council on Informatics and Pathology Innovation (CIPI)
Subscribe to CIPI Connections on Apple Podcasts, Spotify, or wherever you listen to podcasts.
Hosted on Ausha. See ausha.co/privacy-policy for more information.
Transcription
Welcome to CIPI Connections, the podcast of the College of American Pathologists Council on Informatics and Pathology Innovation. Here we connect you with the leaders and committees shaping the future of pathology. I'm Dr. M. E. de Baca, Chair of the College of American Pathologists Council on Informatics and Pathology Innovation, also known as CIPI. Today is the beginning of something brand new. The Informatics Committee will be bringing you horror stories in pathology informatics in multiple episodes. Today, Dr. Alexis Carter, a molecular pathologist and informatician, is joined by Dr. Omar Baba, the Informatics Committee Junior Member and a Pathology Informatics Fellow. Since it's the first in a series, Dr. Carter will initially share some housekeeping details about the series, and then she and Dr. Baba will be discussing the case from the lab. Take it away, Dr. Carter.
Hi everyone, welcome to Horror Stories and Pathology Informatics, Lessons Learned from Things Gone Awry, a series to help your healthcare organization, number one, keep your patients safe, and number two, avoid painful problems. I'm your host, Dr. Alexis Carter, a pathologist and the moderator for this series, and with me today is another pathologist, Dr. Omar Baba. As a reminder, these podcasts are available from your favorite podcast service under CIPI Connections, that's C-I-P-I Connections, and is also from the College of American Pathologists website at www.cap.org. Today's episode is Unflagged and Overlooked Lessons from a Missed Cortisol Result. As always, there are a few disclaimers needed before we get started. All of the situations discussed in this podcast are based on real events, so if any of our listeners think that the situation described can't or wouldn't happen, we can assure you that it can and did. Having said that, all information that could identify the personnel or the healthcare organization involved have been removed and replaced with fictitious characters and a fictitious location of Cabot Cove Memorial Hospital. Any locale-specific aspects? for example, regulations, that are more specific than the national United States level have been removed. And some details regarding the situation which do not affect the main point of the event have been changed. Finally, the person presenting the adverse event is not the person who contributed it for discussion. All of these things were done to ensure that we focus on the lessons learned and so that we can help others avoid these mistakes. Of course, this podcast does not represent legal or medical advice, and the lessons learned may not account for specific barriers that may be present at your own organization. All right, so now that we've done our disclaimers, let's get started. Again, here with me is Dr. Baba, who is going to present today's healthcare software failure on flagging of abnormal laboratory values and the lessons learned from it. So, Dr. Baba, let's start with the basics of the adverse event that we're going to discuss today. What happened?
So, our story today is about a 60-year-old male who was undergoing treatment at Cabot Cole Memorial Hospital for prostate cancer. After the initial radiotherapy and hormone treatment, the patient was scheduled for radical prostatectomy, which is a surgery to remove the prostate gland, and he was examined in the surgery clinic for preoperative testing. Now, during the pre-surgical appointment, the patient had a battery of tests performed among which was a serum cortisol level. Now cortisol is the body's main stress hormone produced by the adrenal glands. It turned out that in this patient the cortisol level was low indicating adrenal insufficiency which is a condition where the adrenal glands are not making enough cortisol. The pre-surgical labs were reviewed by the clinician but the low cortisol level was not identified in that encounter despite being there in the electronic health record. The patient was referred to the emergency department later that week for nausea, vomiting and abdominal pain, but the patient's low cortisol was missed too by the emergency care team and the patient's adrenal insufficiency was not diagnosed. Fortunately, the patient's surgery was postponed, allowing the patient to recover from a presumed viral gastroenteritis. Now, had the surgery not been postponed, the patient, undiagnosed, could have quickly progressed to adrenal crisis, which is a life-threatening emergency marked by severe blood pressure, shock, and even death without urgent treatment. One week after the initial cortisol lab test, the patient contacted the after-hours nursing service advice line to report increased dizziness and lightheadedness when standing up. The patient was told to follow up with their cardiologist in the morning. The next day, the cardiologist finally noted the low cortisol levels and referred the patient to the emergency department. The oncologist was made aware of the diagnosis of adrenal insufficiency. and the patient was treated accordingly, recovered, and successfully underwent the planned prostatectomy one month later.
Okay, wow. So what was the fundamental contributing factor to this particular adverse event?
So in this case, the low cortisol value was missed by the providers because there was no abnormal flag associated with the result. So an abnormal flag is a visual cue in the electronic medical record, or EMR, which is a patient chart basically, like a mark-colored highlight that alerts a provider to a result that requires attention. The providers have got used to relying on the abnormal flag to screen for laboratory values they should pay particular attention to which is common in busy clinical workflows. Cortisol has a diurnal reference range that is levels change throughout the day so levels should be interpreted depending on whether the sample was drawn in the morning or in the afternoon. For instance in morning samples the range is 5 to 23 micrograms per deciliter whereas for the afternoon samples it's 3 to 16 micrograms per deciliter. As a result, many laboratories display both reference ranges in the interpretive comment, which is a free text note that sits alongside the lab result. This, of course, does not allow for flagging of cortisol levels that are abnormal or outside of the reference range.
So with regard to that issue, could this issue have been completely avoided or just mitigated?
This issue definitely could have been avoided, and there are several informatics strategies for accomplishing this.
Okay, so what do you think could have been done to avoid or mitigate the fundamental contributing factor in this case?
This issue could have been avoided by either creating two separate orders for morning and afternoon levels of cortisol, which would lead to having two orders, morning cortisol and afternoon cortisol, making expected timing explicit. Another strategy is to include what we refer to an ask-on-order entry question when the clinicians place the order. The Ask on Order entry is a prompt asking the ordering clinician to specify timing, that is, morning or afternoon, which would trigger the correct reference range and abnormal flag. The system would indicate in this case which reference range should be displayed, thereby ensuring that there is always a discrete reference range and subsequently an abnormal flag when applicable. Alternatively, the nurse could also indicate whether the sample was collected in the morning or the afternoon, as part of the collection process using a different informatics configuration that would accomplish the same thing as the ask on order entry question or separating the morning and afternoon orders. This would also be beneficial to other laboratory tests that may have different reference ranges depending on the timing of the specimen collection, such as some antibiotics like vancomycin and gentamicin, which have different reference ranges for peak, that is right after dosing, or trough, just before the next dose.
Okay, so we've talked about the main contributing factor, but there are always secondary and even sometimes tertiary factors that can contribute to these issues. These can include inhibited communication due to poor relationships, you know, silos between different groups, such as clinicians and laboratorians in this case, or other systemic infrastructural issues. What secondary and tertiary issues do you think contributed to this adverse event and or any delays in its correction?
So one issue that contributed to this adverse event that we cannot always address with an informatics solution is the reliance on visual abnormal flags by providers as screening tools to quickly review laboratory results. Using flags as a shortcut to triage large result sets can be a risk. Another contributing factor is displaying the reference range in the interpretive comment rather than in a more structured format. Structured formats allow systems to act on the reference range whereas free text comments do not.
So with regard to providers' reliance on abnormal flags, which, you know, is understandable, do you think, what do you think could have been done to prevent it from happening?
I think the message has to come down from the hospital leadership that providers must carefully review all laboratory tests and consider the reason why they are ordering these tests. So every test result deserves a deliberate review and should be considered in light of the clinical question for that patient. It's not always possible or appropriate to flag a result as abnormal, especially for narrative or complex results. Narrative results, for example, cannot always be distilled down into a normal or abnormal binary. Sometimes a normal value is significant if it reflects a notable change or trend. On another note, a provider can miss a piece of critical information if they get distracted by exclusively paying close attention to results that are flagged as abnormal. If a test is ordered, it should have been ordered for a reason and deserves the full attention of the provider to answer the clinical question they were asking.
Yeah, those are definitely really good points. Let's go on to examine the use of displaying reference ranges in the interpretive comment. What do you think could have been done to prevent that from happening?
So displaying reference ranges in the interpretive comment can be appropriate, especially when there could be multiple possible interpretations for a value depending on different clinical variables that are not always easily captured with the ask on order entry questions. Sometimes it's not possible to easily capture complex clinical contexts, so a brief explanation can help clinicians interpret lab values. One example might be the absolute reticulocyte count in a CBC, where the expected count differs in anemia, hemolysis, blood loss, etc., so a comment can help tailor interpretation. It would be appropriate to display the reference range or expected value for these different clinical scenarios in order to allow the provider to correctly interpret the result for the patient's clinical condition. While there are definitely situations when using the interpretive comment to display reference ranges as appropriate, this practice should be avoided whenever possible. Ideally, structured fields should accompany results to allow downstream systems to interpret and display them correctly. At the very least, there should be a primary reference range that's displayed with the result in a more structured information. One of the dangers of only displaying the reference range is the interpretive comment is that the result value and the reference range may become uncoupled in downstream transmission to other healthcare institutions, public health agencies, or disease registries. When results move between systems, comments may be dropped and values lose their context. We have a lot of experience transmitting reference ranges in a structured format because it's a required data element per the Clinical Laboratory Improvement Amendments or CLIA. So there are US laboratory regulations that require reference ranges to accompany results. Messaging standards such as HL7 have a dedicated field for transmitting the reference range in a structured format. This ensures that the reference range will be displayed with the result value. Messaging standards can accommodate interpretive comments, but there is less certainty they will always be displayed by the receiving system. Comments may not always show up correctly when data are transmitted. This increases the likelihood that the result value could be displayed without the necessary contextual information to help interpret the result.
So, Dr. Baba, you know, there's a couple of other things that I'm thinking about, and I'm wondering if we can talk about those. So, you know, for example, many people may not realize, especially if they are not pathologists or working in the lab, that we have some restrictions on reference ranges. and that really the reference ranges drive whether or not an abnormal flag can be present, right? The abnormal flag displays when something is outside of its normal reference range automatically, right? And we've got, you know, high, low, abnormal, you know, critically high, critically low flags, and those are all automatically appended based on reference ranges. You know, so what are some additional things, you know, restrictions? So like with this cortisol all result, for example, you know, we typically can do reference ranges based on age and gender, but this case is special for a couple of reasons, right? Because it demonstrates something that we can't easily have automated flags or reference ranges set for. So do you have a couple of examples in your own experience where that might have happened?
One example I could think of is different reference ranges for pregnancy, you know, so collection time of day, pregnancy. So when the clinicians understand the limitations of lab systems, I think they become more lenient, I would say, in cutting us a break about displaying abnormal flags. Because what is considered abnormal, really, right? Like some pathology reports, genetic reports, these are mostly for the majority of the cases, they're like kind of free text. So how do you flag these as abnormal?
Right. And then how abnormal does it have to be in order to call it abnormal? You know, if you have a breast biopsy and what you see is a fibroadenoma, you know, that's... benign, but no one would classify it as normal. So, you know, do you flag those, you know, and flags can also, they can, you know, alert physicians sometimes that things may be worse than they actually are. You know, and the other thing is like, you know, we could potentially in the lab look at a collection date and time. We get collection dates and times on all samples, right? But we're also required to perform the order that the clinician gives us. And so having lab staff switching a lab order to cortisol morning, for example, for specimens that are collected during certain hours, as opposed to cortisol afternoon would require some significant operational changes that may not be necessarily helpful or doable in the lab. In addition to that, we don't always know. I mean, patients don't always know when they're pregnant, right? And, you know, their reference ranges can change, you know, fairly quickly sometimes after they become pregnant. And, you know, as a sort of amusing or not so amusing side note, you know, we had a situation in an organization that I worked in at one point where when we put into the interpretive comment for pregnant patients, this is the reference range. And we actually did have a patient who called her clinician very upset because she thought that the thyroid test result that we were reporting on was indicating that she was actually pregnant because of her value, which the, you know, the clinician explained to her and it was all fine. But, you know, so it's interesting trying to cover all the bases, you know, can be difficult. And, you know, we also want to say, yes, we do believe, you know, clinicians should look at every result, but we also understand that our clinician patient facing friends are just as busy. as we are. And it's hard. It's hard to do that. But on the flip side, you know, when you have these special situations, it's kind of hard to make everything automated. So this has been really educational. Let's sum it up for the listeners. What's the main lessons learned from this incident? And what do you recommend that our listeners should do to prevent this from happening?
So in this case, there was a delay in diagnosis for a patient with adrenal insufficiency due to a missed cortisol result. The low cortisol level was missed because there was no abnormal flag associated with the result. This happened because the laboratory had used the interpretive comment in free text rather than a dedicated discrete numerical field to display two possible reference regions for the cortisol depending on the timing of the system and draw. So comments can educate and add nuance but the efficient machine readable reference should be in a structured field. Whenever possible there should always be a reference range displayed in a structured format that allows for abnormal flagging. This can be accomplished by either creating separate orders for the different draw times or using ask and unorder entry prompts or draw timed questions for the nurse at the time of the collection to allow for display of the appropriate reference range. And lastly, providers need to be educated on the importance of reviewing all results for the test they order and Instead of relying on screening systems to not miss essential pieces of information in the care of patients, again, every ordered result should be reviewed, not just flagged ones.
Great. Thanks so much, Dr. Baba, for presenting today. For our listeners, we hope that you have found this podcast on unflagged and overlooked lessons from a missed cortisol result helpful. This podcast was produced by the College of American Pathologists, and the content was produced by the College of American Pathologists Informatics Committee. The Informatics Committee always welcomes questions about the podcast, as well as suggestions for future podcasts at informatics@CAP.org. If you would like to contribute an issue that happened to you, please contact the Informatics Committee again at informatics at CAP.org for instructions on how to anonymously contribute an educational issue and its lessons learned. Please do not spend specifics on the issue to this email address. We thank you so much for listening, and we look forward to sharing our next podcast with you soon. Bye-bye.
Thanks to Dr. Carter and Dr. Baba for sharing these important lessons with us. Stay tuned for future horror stories in pathology informatics. Thanks for joining us for insights, updates, and the people behind the innovation. This has been CIPI Connections, where ideas meet action in pathology.
Share
Embed
You may also like
Description
In this episode of CIPI Connections, members of the CAP Informatics Committee, Alexis Carter, MD, FCAP, and Omar Baba, MD, share a gripping cautionary tale of unflagged abnormal laboratory values and the informatics solutions designed to avert similar pitfalls.
Council on Informatics and Pathology Innovation (CIPI)
Subscribe to CIPI Connections on Apple Podcasts, Spotify, or wherever you listen to podcasts.
Hosted on Ausha. See ausha.co/privacy-policy for more information.
Transcription
Welcome to CIPI Connections, the podcast of the College of American Pathologists Council on Informatics and Pathology Innovation. Here we connect you with the leaders and committees shaping the future of pathology. I'm Dr. M. E. de Baca, Chair of the College of American Pathologists Council on Informatics and Pathology Innovation, also known as CIPI. Today is the beginning of something brand new. The Informatics Committee will be bringing you horror stories in pathology informatics in multiple episodes. Today, Dr. Alexis Carter, a molecular pathologist and informatician, is joined by Dr. Omar Baba, the Informatics Committee Junior Member and a Pathology Informatics Fellow. Since it's the first in a series, Dr. Carter will initially share some housekeeping details about the series, and then she and Dr. Baba will be discussing the case from the lab. Take it away, Dr. Carter.
Hi everyone, welcome to Horror Stories and Pathology Informatics, Lessons Learned from Things Gone Awry, a series to help your healthcare organization, number one, keep your patients safe, and number two, avoid painful problems. I'm your host, Dr. Alexis Carter, a pathologist and the moderator for this series, and with me today is another pathologist, Dr. Omar Baba. As a reminder, these podcasts are available from your favorite podcast service under CIPI Connections, that's C-I-P-I Connections, and is also from the College of American Pathologists website at www.cap.org. Today's episode is Unflagged and Overlooked Lessons from a Missed Cortisol Result. As always, there are a few disclaimers needed before we get started. All of the situations discussed in this podcast are based on real events, so if any of our listeners think that the situation described can't or wouldn't happen, we can assure you that it can and did. Having said that, all information that could identify the personnel or the healthcare organization involved have been removed and replaced with fictitious characters and a fictitious location of Cabot Cove Memorial Hospital. Any locale-specific aspects? for example, regulations, that are more specific than the national United States level have been removed. And some details regarding the situation which do not affect the main point of the event have been changed. Finally, the person presenting the adverse event is not the person who contributed it for discussion. All of these things were done to ensure that we focus on the lessons learned and so that we can help others avoid these mistakes. Of course, this podcast does not represent legal or medical advice, and the lessons learned may not account for specific barriers that may be present at your own organization. All right, so now that we've done our disclaimers, let's get started. Again, here with me is Dr. Baba, who is going to present today's healthcare software failure on flagging of abnormal laboratory values and the lessons learned from it. So, Dr. Baba, let's start with the basics of the adverse event that we're going to discuss today. What happened?
So, our story today is about a 60-year-old male who was undergoing treatment at Cabot Cole Memorial Hospital for prostate cancer. After the initial radiotherapy and hormone treatment, the patient was scheduled for radical prostatectomy, which is a surgery to remove the prostate gland, and he was examined in the surgery clinic for preoperative testing. Now, during the pre-surgical appointment, the patient had a battery of tests performed among which was a serum cortisol level. Now cortisol is the body's main stress hormone produced by the adrenal glands. It turned out that in this patient the cortisol level was low indicating adrenal insufficiency which is a condition where the adrenal glands are not making enough cortisol. The pre-surgical labs were reviewed by the clinician but the low cortisol level was not identified in that encounter despite being there in the electronic health record. The patient was referred to the emergency department later that week for nausea, vomiting and abdominal pain, but the patient's low cortisol was missed too by the emergency care team and the patient's adrenal insufficiency was not diagnosed. Fortunately, the patient's surgery was postponed, allowing the patient to recover from a presumed viral gastroenteritis. Now, had the surgery not been postponed, the patient, undiagnosed, could have quickly progressed to adrenal crisis, which is a life-threatening emergency marked by severe blood pressure, shock, and even death without urgent treatment. One week after the initial cortisol lab test, the patient contacted the after-hours nursing service advice line to report increased dizziness and lightheadedness when standing up. The patient was told to follow up with their cardiologist in the morning. The next day, the cardiologist finally noted the low cortisol levels and referred the patient to the emergency department. The oncologist was made aware of the diagnosis of adrenal insufficiency. and the patient was treated accordingly, recovered, and successfully underwent the planned prostatectomy one month later.
Okay, wow. So what was the fundamental contributing factor to this particular adverse event?
So in this case, the low cortisol value was missed by the providers because there was no abnormal flag associated with the result. So an abnormal flag is a visual cue in the electronic medical record, or EMR, which is a patient chart basically, like a mark-colored highlight that alerts a provider to a result that requires attention. The providers have got used to relying on the abnormal flag to screen for laboratory values they should pay particular attention to which is common in busy clinical workflows. Cortisol has a diurnal reference range that is levels change throughout the day so levels should be interpreted depending on whether the sample was drawn in the morning or in the afternoon. For instance in morning samples the range is 5 to 23 micrograms per deciliter whereas for the afternoon samples it's 3 to 16 micrograms per deciliter. As a result, many laboratories display both reference ranges in the interpretive comment, which is a free text note that sits alongside the lab result. This, of course, does not allow for flagging of cortisol levels that are abnormal or outside of the reference range.
So with regard to that issue, could this issue have been completely avoided or just mitigated?
This issue definitely could have been avoided, and there are several informatics strategies for accomplishing this.
Okay, so what do you think could have been done to avoid or mitigate the fundamental contributing factor in this case?
This issue could have been avoided by either creating two separate orders for morning and afternoon levels of cortisol, which would lead to having two orders, morning cortisol and afternoon cortisol, making expected timing explicit. Another strategy is to include what we refer to an ask-on-order entry question when the clinicians place the order. The Ask on Order entry is a prompt asking the ordering clinician to specify timing, that is, morning or afternoon, which would trigger the correct reference range and abnormal flag. The system would indicate in this case which reference range should be displayed, thereby ensuring that there is always a discrete reference range and subsequently an abnormal flag when applicable. Alternatively, the nurse could also indicate whether the sample was collected in the morning or the afternoon, as part of the collection process using a different informatics configuration that would accomplish the same thing as the ask on order entry question or separating the morning and afternoon orders. This would also be beneficial to other laboratory tests that may have different reference ranges depending on the timing of the specimen collection, such as some antibiotics like vancomycin and gentamicin, which have different reference ranges for peak, that is right after dosing, or trough, just before the next dose.
Okay, so we've talked about the main contributing factor, but there are always secondary and even sometimes tertiary factors that can contribute to these issues. These can include inhibited communication due to poor relationships, you know, silos between different groups, such as clinicians and laboratorians in this case, or other systemic infrastructural issues. What secondary and tertiary issues do you think contributed to this adverse event and or any delays in its correction?
So one issue that contributed to this adverse event that we cannot always address with an informatics solution is the reliance on visual abnormal flags by providers as screening tools to quickly review laboratory results. Using flags as a shortcut to triage large result sets can be a risk. Another contributing factor is displaying the reference range in the interpretive comment rather than in a more structured format. Structured formats allow systems to act on the reference range whereas free text comments do not.
So with regard to providers' reliance on abnormal flags, which, you know, is understandable, do you think, what do you think could have been done to prevent it from happening?
I think the message has to come down from the hospital leadership that providers must carefully review all laboratory tests and consider the reason why they are ordering these tests. So every test result deserves a deliberate review and should be considered in light of the clinical question for that patient. It's not always possible or appropriate to flag a result as abnormal, especially for narrative or complex results. Narrative results, for example, cannot always be distilled down into a normal or abnormal binary. Sometimes a normal value is significant if it reflects a notable change or trend. On another note, a provider can miss a piece of critical information if they get distracted by exclusively paying close attention to results that are flagged as abnormal. If a test is ordered, it should have been ordered for a reason and deserves the full attention of the provider to answer the clinical question they were asking.
Yeah, those are definitely really good points. Let's go on to examine the use of displaying reference ranges in the interpretive comment. What do you think could have been done to prevent that from happening?
So displaying reference ranges in the interpretive comment can be appropriate, especially when there could be multiple possible interpretations for a value depending on different clinical variables that are not always easily captured with the ask on order entry questions. Sometimes it's not possible to easily capture complex clinical contexts, so a brief explanation can help clinicians interpret lab values. One example might be the absolute reticulocyte count in a CBC, where the expected count differs in anemia, hemolysis, blood loss, etc., so a comment can help tailor interpretation. It would be appropriate to display the reference range or expected value for these different clinical scenarios in order to allow the provider to correctly interpret the result for the patient's clinical condition. While there are definitely situations when using the interpretive comment to display reference ranges as appropriate, this practice should be avoided whenever possible. Ideally, structured fields should accompany results to allow downstream systems to interpret and display them correctly. At the very least, there should be a primary reference range that's displayed with the result in a more structured information. One of the dangers of only displaying the reference range is the interpretive comment is that the result value and the reference range may become uncoupled in downstream transmission to other healthcare institutions, public health agencies, or disease registries. When results move between systems, comments may be dropped and values lose their context. We have a lot of experience transmitting reference ranges in a structured format because it's a required data element per the Clinical Laboratory Improvement Amendments or CLIA. So there are US laboratory regulations that require reference ranges to accompany results. Messaging standards such as HL7 have a dedicated field for transmitting the reference range in a structured format. This ensures that the reference range will be displayed with the result value. Messaging standards can accommodate interpretive comments, but there is less certainty they will always be displayed by the receiving system. Comments may not always show up correctly when data are transmitted. This increases the likelihood that the result value could be displayed without the necessary contextual information to help interpret the result.
So, Dr. Baba, you know, there's a couple of other things that I'm thinking about, and I'm wondering if we can talk about those. So, you know, for example, many people may not realize, especially if they are not pathologists or working in the lab, that we have some restrictions on reference ranges. and that really the reference ranges drive whether or not an abnormal flag can be present, right? The abnormal flag displays when something is outside of its normal reference range automatically, right? And we've got, you know, high, low, abnormal, you know, critically high, critically low flags, and those are all automatically appended based on reference ranges. You know, so what are some additional things, you know, restrictions? So like with this cortisol all result, for example, you know, we typically can do reference ranges based on age and gender, but this case is special for a couple of reasons, right? Because it demonstrates something that we can't easily have automated flags or reference ranges set for. So do you have a couple of examples in your own experience where that might have happened?
One example I could think of is different reference ranges for pregnancy, you know, so collection time of day, pregnancy. So when the clinicians understand the limitations of lab systems, I think they become more lenient, I would say, in cutting us a break about displaying abnormal flags. Because what is considered abnormal, really, right? Like some pathology reports, genetic reports, these are mostly for the majority of the cases, they're like kind of free text. So how do you flag these as abnormal?
Right. And then how abnormal does it have to be in order to call it abnormal? You know, if you have a breast biopsy and what you see is a fibroadenoma, you know, that's... benign, but no one would classify it as normal. So, you know, do you flag those, you know, and flags can also, they can, you know, alert physicians sometimes that things may be worse than they actually are. You know, and the other thing is like, you know, we could potentially in the lab look at a collection date and time. We get collection dates and times on all samples, right? But we're also required to perform the order that the clinician gives us. And so having lab staff switching a lab order to cortisol morning, for example, for specimens that are collected during certain hours, as opposed to cortisol afternoon would require some significant operational changes that may not be necessarily helpful or doable in the lab. In addition to that, we don't always know. I mean, patients don't always know when they're pregnant, right? And, you know, their reference ranges can change, you know, fairly quickly sometimes after they become pregnant. And, you know, as a sort of amusing or not so amusing side note, you know, we had a situation in an organization that I worked in at one point where when we put into the interpretive comment for pregnant patients, this is the reference range. And we actually did have a patient who called her clinician very upset because she thought that the thyroid test result that we were reporting on was indicating that she was actually pregnant because of her value, which the, you know, the clinician explained to her and it was all fine. But, you know, so it's interesting trying to cover all the bases, you know, can be difficult. And, you know, we also want to say, yes, we do believe, you know, clinicians should look at every result, but we also understand that our clinician patient facing friends are just as busy. as we are. And it's hard. It's hard to do that. But on the flip side, you know, when you have these special situations, it's kind of hard to make everything automated. So this has been really educational. Let's sum it up for the listeners. What's the main lessons learned from this incident? And what do you recommend that our listeners should do to prevent this from happening?
So in this case, there was a delay in diagnosis for a patient with adrenal insufficiency due to a missed cortisol result. The low cortisol level was missed because there was no abnormal flag associated with the result. This happened because the laboratory had used the interpretive comment in free text rather than a dedicated discrete numerical field to display two possible reference regions for the cortisol depending on the timing of the system and draw. So comments can educate and add nuance but the efficient machine readable reference should be in a structured field. Whenever possible there should always be a reference range displayed in a structured format that allows for abnormal flagging. This can be accomplished by either creating separate orders for the different draw times or using ask and unorder entry prompts or draw timed questions for the nurse at the time of the collection to allow for display of the appropriate reference range. And lastly, providers need to be educated on the importance of reviewing all results for the test they order and Instead of relying on screening systems to not miss essential pieces of information in the care of patients, again, every ordered result should be reviewed, not just flagged ones.
Great. Thanks so much, Dr. Baba, for presenting today. For our listeners, we hope that you have found this podcast on unflagged and overlooked lessons from a missed cortisol result helpful. This podcast was produced by the College of American Pathologists, and the content was produced by the College of American Pathologists Informatics Committee. The Informatics Committee always welcomes questions about the podcast, as well as suggestions for future podcasts at informatics@CAP.org. If you would like to contribute an issue that happened to you, please contact the Informatics Committee again at informatics at CAP.org for instructions on how to anonymously contribute an educational issue and its lessons learned. Please do not spend specifics on the issue to this email address. We thank you so much for listening, and we look forward to sharing our next podcast with you soon. Bye-bye.
Thanks to Dr. Carter and Dr. Baba for sharing these important lessons with us. Stay tuned for future horror stories in pathology informatics. Thanks for joining us for insights, updates, and the people behind the innovation. This has been CIPI Connections, where ideas meet action in pathology.
Description
In this episode of CIPI Connections, members of the CAP Informatics Committee, Alexis Carter, MD, FCAP, and Omar Baba, MD, share a gripping cautionary tale of unflagged abnormal laboratory values and the informatics solutions designed to avert similar pitfalls.
Council on Informatics and Pathology Innovation (CIPI)
Subscribe to CIPI Connections on Apple Podcasts, Spotify, or wherever you listen to podcasts.
Hosted on Ausha. See ausha.co/privacy-policy for more information.
Transcription
Welcome to CIPI Connections, the podcast of the College of American Pathologists Council on Informatics and Pathology Innovation. Here we connect you with the leaders and committees shaping the future of pathology. I'm Dr. M. E. de Baca, Chair of the College of American Pathologists Council on Informatics and Pathology Innovation, also known as CIPI. Today is the beginning of something brand new. The Informatics Committee will be bringing you horror stories in pathology informatics in multiple episodes. Today, Dr. Alexis Carter, a molecular pathologist and informatician, is joined by Dr. Omar Baba, the Informatics Committee Junior Member and a Pathology Informatics Fellow. Since it's the first in a series, Dr. Carter will initially share some housekeeping details about the series, and then she and Dr. Baba will be discussing the case from the lab. Take it away, Dr. Carter.
Hi everyone, welcome to Horror Stories and Pathology Informatics, Lessons Learned from Things Gone Awry, a series to help your healthcare organization, number one, keep your patients safe, and number two, avoid painful problems. I'm your host, Dr. Alexis Carter, a pathologist and the moderator for this series, and with me today is another pathologist, Dr. Omar Baba. As a reminder, these podcasts are available from your favorite podcast service under CIPI Connections, that's C-I-P-I Connections, and is also from the College of American Pathologists website at www.cap.org. Today's episode is Unflagged and Overlooked Lessons from a Missed Cortisol Result. As always, there are a few disclaimers needed before we get started. All of the situations discussed in this podcast are based on real events, so if any of our listeners think that the situation described can't or wouldn't happen, we can assure you that it can and did. Having said that, all information that could identify the personnel or the healthcare organization involved have been removed and replaced with fictitious characters and a fictitious location of Cabot Cove Memorial Hospital. Any locale-specific aspects? for example, regulations, that are more specific than the national United States level have been removed. And some details regarding the situation which do not affect the main point of the event have been changed. Finally, the person presenting the adverse event is not the person who contributed it for discussion. All of these things were done to ensure that we focus on the lessons learned and so that we can help others avoid these mistakes. Of course, this podcast does not represent legal or medical advice, and the lessons learned may not account for specific barriers that may be present at your own organization. All right, so now that we've done our disclaimers, let's get started. Again, here with me is Dr. Baba, who is going to present today's healthcare software failure on flagging of abnormal laboratory values and the lessons learned from it. So, Dr. Baba, let's start with the basics of the adverse event that we're going to discuss today. What happened?
So, our story today is about a 60-year-old male who was undergoing treatment at Cabot Cole Memorial Hospital for prostate cancer. After the initial radiotherapy and hormone treatment, the patient was scheduled for radical prostatectomy, which is a surgery to remove the prostate gland, and he was examined in the surgery clinic for preoperative testing. Now, during the pre-surgical appointment, the patient had a battery of tests performed among which was a serum cortisol level. Now cortisol is the body's main stress hormone produced by the adrenal glands. It turned out that in this patient the cortisol level was low indicating adrenal insufficiency which is a condition where the adrenal glands are not making enough cortisol. The pre-surgical labs were reviewed by the clinician but the low cortisol level was not identified in that encounter despite being there in the electronic health record. The patient was referred to the emergency department later that week for nausea, vomiting and abdominal pain, but the patient's low cortisol was missed too by the emergency care team and the patient's adrenal insufficiency was not diagnosed. Fortunately, the patient's surgery was postponed, allowing the patient to recover from a presumed viral gastroenteritis. Now, had the surgery not been postponed, the patient, undiagnosed, could have quickly progressed to adrenal crisis, which is a life-threatening emergency marked by severe blood pressure, shock, and even death without urgent treatment. One week after the initial cortisol lab test, the patient contacted the after-hours nursing service advice line to report increased dizziness and lightheadedness when standing up. The patient was told to follow up with their cardiologist in the morning. The next day, the cardiologist finally noted the low cortisol levels and referred the patient to the emergency department. The oncologist was made aware of the diagnosis of adrenal insufficiency. and the patient was treated accordingly, recovered, and successfully underwent the planned prostatectomy one month later.
Okay, wow. So what was the fundamental contributing factor to this particular adverse event?
So in this case, the low cortisol value was missed by the providers because there was no abnormal flag associated with the result. So an abnormal flag is a visual cue in the electronic medical record, or EMR, which is a patient chart basically, like a mark-colored highlight that alerts a provider to a result that requires attention. The providers have got used to relying on the abnormal flag to screen for laboratory values they should pay particular attention to which is common in busy clinical workflows. Cortisol has a diurnal reference range that is levels change throughout the day so levels should be interpreted depending on whether the sample was drawn in the morning or in the afternoon. For instance in morning samples the range is 5 to 23 micrograms per deciliter whereas for the afternoon samples it's 3 to 16 micrograms per deciliter. As a result, many laboratories display both reference ranges in the interpretive comment, which is a free text note that sits alongside the lab result. This, of course, does not allow for flagging of cortisol levels that are abnormal or outside of the reference range.
So with regard to that issue, could this issue have been completely avoided or just mitigated?
This issue definitely could have been avoided, and there are several informatics strategies for accomplishing this.
Okay, so what do you think could have been done to avoid or mitigate the fundamental contributing factor in this case?
This issue could have been avoided by either creating two separate orders for morning and afternoon levels of cortisol, which would lead to having two orders, morning cortisol and afternoon cortisol, making expected timing explicit. Another strategy is to include what we refer to an ask-on-order entry question when the clinicians place the order. The Ask on Order entry is a prompt asking the ordering clinician to specify timing, that is, morning or afternoon, which would trigger the correct reference range and abnormal flag. The system would indicate in this case which reference range should be displayed, thereby ensuring that there is always a discrete reference range and subsequently an abnormal flag when applicable. Alternatively, the nurse could also indicate whether the sample was collected in the morning or the afternoon, as part of the collection process using a different informatics configuration that would accomplish the same thing as the ask on order entry question or separating the morning and afternoon orders. This would also be beneficial to other laboratory tests that may have different reference ranges depending on the timing of the specimen collection, such as some antibiotics like vancomycin and gentamicin, which have different reference ranges for peak, that is right after dosing, or trough, just before the next dose.
Okay, so we've talked about the main contributing factor, but there are always secondary and even sometimes tertiary factors that can contribute to these issues. These can include inhibited communication due to poor relationships, you know, silos between different groups, such as clinicians and laboratorians in this case, or other systemic infrastructural issues. What secondary and tertiary issues do you think contributed to this adverse event and or any delays in its correction?
So one issue that contributed to this adverse event that we cannot always address with an informatics solution is the reliance on visual abnormal flags by providers as screening tools to quickly review laboratory results. Using flags as a shortcut to triage large result sets can be a risk. Another contributing factor is displaying the reference range in the interpretive comment rather than in a more structured format. Structured formats allow systems to act on the reference range whereas free text comments do not.
So with regard to providers' reliance on abnormal flags, which, you know, is understandable, do you think, what do you think could have been done to prevent it from happening?
I think the message has to come down from the hospital leadership that providers must carefully review all laboratory tests and consider the reason why they are ordering these tests. So every test result deserves a deliberate review and should be considered in light of the clinical question for that patient. It's not always possible or appropriate to flag a result as abnormal, especially for narrative or complex results. Narrative results, for example, cannot always be distilled down into a normal or abnormal binary. Sometimes a normal value is significant if it reflects a notable change or trend. On another note, a provider can miss a piece of critical information if they get distracted by exclusively paying close attention to results that are flagged as abnormal. If a test is ordered, it should have been ordered for a reason and deserves the full attention of the provider to answer the clinical question they were asking.
Yeah, those are definitely really good points. Let's go on to examine the use of displaying reference ranges in the interpretive comment. What do you think could have been done to prevent that from happening?
So displaying reference ranges in the interpretive comment can be appropriate, especially when there could be multiple possible interpretations for a value depending on different clinical variables that are not always easily captured with the ask on order entry questions. Sometimes it's not possible to easily capture complex clinical contexts, so a brief explanation can help clinicians interpret lab values. One example might be the absolute reticulocyte count in a CBC, where the expected count differs in anemia, hemolysis, blood loss, etc., so a comment can help tailor interpretation. It would be appropriate to display the reference range or expected value for these different clinical scenarios in order to allow the provider to correctly interpret the result for the patient's clinical condition. While there are definitely situations when using the interpretive comment to display reference ranges as appropriate, this practice should be avoided whenever possible. Ideally, structured fields should accompany results to allow downstream systems to interpret and display them correctly. At the very least, there should be a primary reference range that's displayed with the result in a more structured information. One of the dangers of only displaying the reference range is the interpretive comment is that the result value and the reference range may become uncoupled in downstream transmission to other healthcare institutions, public health agencies, or disease registries. When results move between systems, comments may be dropped and values lose their context. We have a lot of experience transmitting reference ranges in a structured format because it's a required data element per the Clinical Laboratory Improvement Amendments or CLIA. So there are US laboratory regulations that require reference ranges to accompany results. Messaging standards such as HL7 have a dedicated field for transmitting the reference range in a structured format. This ensures that the reference range will be displayed with the result value. Messaging standards can accommodate interpretive comments, but there is less certainty they will always be displayed by the receiving system. Comments may not always show up correctly when data are transmitted. This increases the likelihood that the result value could be displayed without the necessary contextual information to help interpret the result.
So, Dr. Baba, you know, there's a couple of other things that I'm thinking about, and I'm wondering if we can talk about those. So, you know, for example, many people may not realize, especially if they are not pathologists or working in the lab, that we have some restrictions on reference ranges. and that really the reference ranges drive whether or not an abnormal flag can be present, right? The abnormal flag displays when something is outside of its normal reference range automatically, right? And we've got, you know, high, low, abnormal, you know, critically high, critically low flags, and those are all automatically appended based on reference ranges. You know, so what are some additional things, you know, restrictions? So like with this cortisol all result, for example, you know, we typically can do reference ranges based on age and gender, but this case is special for a couple of reasons, right? Because it demonstrates something that we can't easily have automated flags or reference ranges set for. So do you have a couple of examples in your own experience where that might have happened?
One example I could think of is different reference ranges for pregnancy, you know, so collection time of day, pregnancy. So when the clinicians understand the limitations of lab systems, I think they become more lenient, I would say, in cutting us a break about displaying abnormal flags. Because what is considered abnormal, really, right? Like some pathology reports, genetic reports, these are mostly for the majority of the cases, they're like kind of free text. So how do you flag these as abnormal?
Right. And then how abnormal does it have to be in order to call it abnormal? You know, if you have a breast biopsy and what you see is a fibroadenoma, you know, that's... benign, but no one would classify it as normal. So, you know, do you flag those, you know, and flags can also, they can, you know, alert physicians sometimes that things may be worse than they actually are. You know, and the other thing is like, you know, we could potentially in the lab look at a collection date and time. We get collection dates and times on all samples, right? But we're also required to perform the order that the clinician gives us. And so having lab staff switching a lab order to cortisol morning, for example, for specimens that are collected during certain hours, as opposed to cortisol afternoon would require some significant operational changes that may not be necessarily helpful or doable in the lab. In addition to that, we don't always know. I mean, patients don't always know when they're pregnant, right? And, you know, their reference ranges can change, you know, fairly quickly sometimes after they become pregnant. And, you know, as a sort of amusing or not so amusing side note, you know, we had a situation in an organization that I worked in at one point where when we put into the interpretive comment for pregnant patients, this is the reference range. And we actually did have a patient who called her clinician very upset because she thought that the thyroid test result that we were reporting on was indicating that she was actually pregnant because of her value, which the, you know, the clinician explained to her and it was all fine. But, you know, so it's interesting trying to cover all the bases, you know, can be difficult. And, you know, we also want to say, yes, we do believe, you know, clinicians should look at every result, but we also understand that our clinician patient facing friends are just as busy. as we are. And it's hard. It's hard to do that. But on the flip side, you know, when you have these special situations, it's kind of hard to make everything automated. So this has been really educational. Let's sum it up for the listeners. What's the main lessons learned from this incident? And what do you recommend that our listeners should do to prevent this from happening?
So in this case, there was a delay in diagnosis for a patient with adrenal insufficiency due to a missed cortisol result. The low cortisol level was missed because there was no abnormal flag associated with the result. This happened because the laboratory had used the interpretive comment in free text rather than a dedicated discrete numerical field to display two possible reference regions for the cortisol depending on the timing of the system and draw. So comments can educate and add nuance but the efficient machine readable reference should be in a structured field. Whenever possible there should always be a reference range displayed in a structured format that allows for abnormal flagging. This can be accomplished by either creating separate orders for the different draw times or using ask and unorder entry prompts or draw timed questions for the nurse at the time of the collection to allow for display of the appropriate reference range. And lastly, providers need to be educated on the importance of reviewing all results for the test they order and Instead of relying on screening systems to not miss essential pieces of information in the care of patients, again, every ordered result should be reviewed, not just flagged ones.
Great. Thanks so much, Dr. Baba, for presenting today. For our listeners, we hope that you have found this podcast on unflagged and overlooked lessons from a missed cortisol result helpful. This podcast was produced by the College of American Pathologists, and the content was produced by the College of American Pathologists Informatics Committee. The Informatics Committee always welcomes questions about the podcast, as well as suggestions for future podcasts at informatics@CAP.org. If you would like to contribute an issue that happened to you, please contact the Informatics Committee again at informatics at CAP.org for instructions on how to anonymously contribute an educational issue and its lessons learned. Please do not spend specifics on the issue to this email address. We thank you so much for listening, and we look forward to sharing our next podcast with you soon. Bye-bye.
Thanks to Dr. Carter and Dr. Baba for sharing these important lessons with us. Stay tuned for future horror stories in pathology informatics. Thanks for joining us for insights, updates, and the people behind the innovation. This has been CIPI Connections, where ideas meet action in pathology.
Share
Embed
You may also like
