Exploring Senior Emergency Physicians' Perspectives on the Inclusion of Clinical Biomarkers in Clinical Decision Rules for CT Head Use in Patients With Suspected Traumatic Brain Injury

ABSTRACT

Background

Clinical biomarkers are an objective test with the potential to improve CT head decision pathways by better risk stratifying patients and supporting Emergency department (ED) physician gestalt. Understanding variations in diagnostic approach, risk tolerance and the perceived need to upgrade guidelines and where will help inform further development of novel TBI pathways.

Objectives

This study explored senior emergency physician perspectives regarding current clinical pathways and the potential use of clinical biomarkers as a rule‐out test for intracranial pathology in head‐injured patients.

Method

This descriptive qualitative design study uses data from 11 individual senior ED clinician interviews. Thematic analysis was then performed.

Results

The three main themes identified were safe practice and clinical decision‐making, resources and logistics and challenges of changing practice. While most clinicians reported easy access to CT head scans, most also reported that waiting for CTs was associated with delays. The majority believed that at least some patient groups were over‐investigated in this context. Potential benefits of a rule‐out biomarker were viewed positively. However, risk tolerance for missed findings was low, ranging from 0.2% to 1%–2%. A strong evidence base demonstrating a biomarker performed with this precision was viewed by most as essential prior to its use.

Conclusion

Senior ED physicians viewed the potential inclusion of biomarkers in CT head guidelines favorably, particularly concerning decision support in low‐risk head injuries. However, the lack of expert consensus or strong evidence demonstrating high diagnostic precision in comparable emergency care models limits their introduction.

1. Background

Traumatic brain injury (TBI) is a global public health issue responsible for significant morbidity and mortality across all demographic groups [1]. Most injured patients present to emergency departments (ED) for assessment [2]. ED clinicians must determine which patients require a computed tomography (CT) head scan to rule out serious intracranial injuries [2, 3]. CT head clinical decision guidelines have been widely adopted into routine clinical practice to curtail exponential rises in CT head use [4, 5]. Despite this, CT head use rates continue to increase, contributing to ED workload and resource pressure [6, 7].

Clinical biomarkers are objective tests with the potential to improve CT head decision guidelines by better risk‐stratifying patients [8, 9, 10, 11] S100B, GFAP and UCH‐L1 have demonstrated promise as suitable biomarkers to rule out significant intracranial pathology in head‐injured patients [8, 9, 10, 11]. This would mean patients returning a negative biomarker test could be discharged from ED without a CT head. However, quality evidence regarding the safety and economic benefits in broader international emergency care settings is still coveted, delaying their implementation in clinical pathways [11, 12, 13].

Ultimately, deciding not to perform a CT head in a set of patients that previously would have a CT head performed is a large culture change in ED that is likely to be met with uncertainty and doubt [14]. Also, there is likely to be clinical variation in how ED physicians practice as well as different opinions and attitudes towards these issues [15, 16, 17]. Understanding variation in diagnostic approach, risk tolerance and perceived need to upgrade guidelines and where will help inform the design and development of novel clinical pathways for TBI.

This study aimed to explore senior emergency physician perspectives regarding current ED CT head decision guidelines and the potential inclusion of clinical biomarkers as a rule‐out test for intracranial pathology in head‐injured patients. Qualitative research in this area could help determine where in guidelines biomarkers could be best placed, given their use is still in early developmental phases, as well as uncover a deeper understanding of senior physician experience, opinions and risk tolerance.

2. Methods

2.1. Study Design

This was a descriptive qualitative study generating data from individual senior ED physician interviews. The consolidated criteria for reporting qualitative research (COREQ) checklist was used as a framework [18].

2.2. Participants

Senior Australasian physicians in emergency medicine—defined as either a Fellow of the Australasian College of Emergency Medicine (FACEM) or a senior ACEM trainee studying for post‐completion of the FACEM exam (in general indicating succession to FACEM within the following 12 months)—were invited to participate in this study. These physicians are the primary stakeholders and ultimate decision makers with reference to ED CT head requests from ED.

Study participants were recruited indirectly via the NZ EMN, a national network of emergency care researchers. Most NZ EMN members were either co‐investigators or aware of biomarker research being conducted by the authors and were therefore not included themselves. However, general email invitations were sent via members to colleagues in their respective departments, advising interested physicians to contact the study investigators. Individuals were not directly approached to avoid unintentional coercion, so the authors cannot report on those who declined to participate. Participation was voluntary and electronic written consent was obtained before the interview. Participants were sent a copy of the participant information and consent form prior to the interview. Participants completed a voluntary demographic form detailing their background and ED experience.

2.3. Interviews

Two study investigators (A.R. and P.L.) conducted one‐on‐one individual semi‐structured interviews via Zoom. The interview guide and questions used are displayed in Table 1. These questions were developed iteratively by the two investigators and an ED consultant with research experience. Two pilot interviews were conducted with ED consultants. Participants were informed about the study aim in the consent form. The semi‐structured approach enabled a focused approach to specific topics but was flexible and adaptable. The interview was collaborative with both the investigator and the participant achieving understanding and meaning around the topic [19]. This approach enabled discussion between the investigators and participants to enable a better understanding of the issues [19]. A.R. transcribed the interviews verbatim and checked them for quality by P.L. Participants were given the opportunity to review their transcripts. No follow‐up or repeat interviews occurred, but participants were invited to comment on the final report.

TABLE 1.

Interview questions and prompt.

Questions		Prompts
1	Tell me from your perspective, what you think about the way in which CT head scans are used for head injured patients in ED?	Impacts on patients/department? Are too many or too few ordered?
2	How do you decide which patients require a CT head scan following a head injury?	Any areas of uncertainty or clinical concern? Do you use a guideline?
3	Biomarkers are currently being developed and validated for use in Emergency Departments as rule out tests for intracranial injuries in patients with mild head injury (GCS 13–15). This would mean some patients that currently undergo CT head scanning, would be discharged from ED without a scan if their biomarker results fell below a certain threshold. Do you have any initial thoughts on this? Where do you think biomarkers could or could not be useful?	Would having a biomarker to use as a rule out test make you feel more comfortable about not performing a CT head scan?
4	We are particularly interested in learning what your risk tolerance would be with regards to biomarker use in head‐injured patients. What evidence would you want to see to feel comfortable discharging a head‐injured patient, after a negative biomarker result but without a CT head scan?	What level of risk would you be willing to accept in this situation? Are there any injuries you would be willing to miss?
5	Do you have any other comments?

2.4. Data Collection

All data, including demographic details, audio recordings and transcripts, were stored securely on a REDCap database hosted by the University of Otago (Version 10.1.2). Transcribed interviews were uploaded to NVivo 12 software (QSR International, Melbourne, Australia) for data analysis.

2.5. Data Analysis and Thematic Coding

Data were analysed using qualitative thematic analysis [19, 20, 21, 22]. This is a method of analysing text in a systematic way to organise qualitative data into themes. Investigators used an inductive approach to coding, allowing the codes and subsequent themes to arise from the narrative [20]. The investigators conducted this independently with regular meetings to discuss coding trees, emerging themes and review disagreements [20]. Data collection and analysis were conducted simultaneously to ensure a reflexive process [21, 22], review emerging themes and monitor for data saturation. Data saturation was assessed periodically after 7, 9 and 11 interviews. At this point, both investigators agreed on data saturation using the inductive thematic saturation and data saturation models described by Saunders et al. [23] Coding trees were then compared and discussed in detail until final themes and subthemes were agreed upon. A narrative interpretation of the data has been provided, and supporting quotations have been edited for readability.

2.6. Ethical Approval and Māori Consultation

Research was conducted following ethical approval obtained from the University of Otago Human Ethics Committee (HEC Reference: H22/043). Māori Consultation was reviewed by Ngāi Tahu Research Consultation Committee.

3. Results

Between 27 April and 21 June 2022, 11 interviews were performed. The median interview length was 17.48 min (range 13.03–25.24). Participant demographics and ED work experience are displayed in Table 2. Seven of the physicians interviewed were emergency specialists and four were senior registrars who were due to become specialists within the next 12 months. Six of the emergency physicians interviewed were working in major referral EDs and five worked in regional/rural settings.

TABLE 2.

Participant demographics, ED setting and ED work experience.

	N = 11	Frequency
Gender N (%)	Female	6 (54.5)
	Male	5 (45.5)
Ethnicity N (%)	NZ European	4 (36.4)
	Asian	2 (18.2)
	Other European	5 (45.5)
Emergency department setting N (%)	Major Referral	6 (54.4)
	Regional/Rural	5 (45.5)
Years working as an ED doctor N (%)	5–7 years	4 (36.3)
	8–10 years	3 (27.3)
	10–15 years	2 (18.2)
	15–25 years	2 (18.2)
Years working as an ED doctor in NZ or Australia N (%)	5–7 years	6 (54.5)
	8–10 years	2 (18.2)
	10–15 years	2 (18.2)
	15–25 years	1 (9.1)

3.1. Thematic Analysis

The three main themes identified were Safe practice and clinical decision‐making, resources and logistics and challenges of changing practice. Table 3 provides additional supporting quotations for the main themes and subthemes. Quotes have been edited for readability.

TABLE 3.

Themes, subthemes, and additional supporting quotations for thematic analysis.

Safe practice and clinical decision‐making	Current guideline performance	So, combination of all the various bits of the various scores and gestalt. I don't particularly follow any particular score, I think they all have their plus points, and they all have their low points. And generally, whenever you look at research, it looks at senior clinician gestalt versus a score, the gestalt is as good or better. I think that's probably because we do steal bits from all of the different scoring systems. (SMO‐03) I'm like a relatively new FACEM or a senior registrar, right, so I'm like very, I'm still establishing my own like clinical gestalt when it comes to these patients. And so, I'm very… I use the rules like particularly the Canadian CT head rules to like to support my clinical gestalt and avoid CTing people that don't need to be CT'd. (SMO‐08)
	Decision protection	So, if you can add a biomarker onto that and actually like that supports your thoughts that actually this is a low risk, low probability injury, and your pretest probability is already low, and then you get the biomarker that is negative well, then that's an early discharge, rather than admission. (SMO‐04) Particularly, I think if you're in that sort of fatigue decision making realm, you know, end of a shift or late night or something like that, if you may be doubting, you're sort of clinical judgement a little bit. (SMO‐10)
	Clinical consensus and Individual risk tolerance	I'm not a flagship runner, with my new tests, it would take something to be put into common practice before I would feel particularly confident using it as my primary management plan or investigation plan. (SMO‐07)
		Again, it depends, because I'm quite, I'm relatively I don't know, the opposite to risk adverse, if that makes sense, like risk happy. As long as the patient knows when to come back, I don't mind missing injuries and letting people go as long as they know that they need to come back and when they need to come back. (SMO‐06)
Resources and logistics	CT head accessibility	And in the daytime, depends on where you work it varies, sometimes it's very easy to get CT heads, and sometimes it's very difficult to get CT heads, and then there's quite a lot of delay to kind of those sorts of things. And that's a lot that can be a lot of ‘bed’ time, that's taken up with patients' kind of hanging around for results. (SMO‐05)
	Over investigation	And yeah, I'm really interested in in how it will perform in people on anticoagulants. That's really where I'm think we could get a real reduction in in imaging. (SMO‐03) But I think in general, young, sort of concussion type patients gets over investigated with, with CTs. And I think that, anecdotally, I feel like the pick‐up rate for any injury on those as almost none. And I think it falsely reassures people as well in that concussion setting, that or, I must be okay, or they must be okay, because the CT was normal. (SMO‐10)
	ED flow and overcrowding	And also, the turnaround, you know, we have massive issues with bed block. Erm, and you can wait a few, three or four hours for a scan. And if you can get a biomarker back in an hour and discharge, then that's., and then you add that up over bed hours, over a long period of time, that's potentially a huge number of bed hours that you gain, which in a small department is huge. (SMO‐03)
		So, if you've got like patients who have had a head injury, who presented their GP, or, you know, like, ***** hospital, as a regional service that like, refers to us a lot. And you know, you know, head injury, you need to CT head query bleed… And so like, the money that, and time that could be saved from that point of view by avoiding a transfer for a scan is another way. (SMO‐08)
Challenges of changing practice	Evidence based consensus	Because, because one of the concerns as well, I think we should aim for, is a local population‐based pathway. We can't guarantee that the Canadians are the same as the New Zealanders, right? So, so I think that's what we're heading for here. (SMO‐09)
	Culture Shift	I think it'll be another one of those things as well, that would probably take a little bit of time to shift culture, I guess, that's probably the way to think, because it's, it will be a massive change in how we manage these things. Yeah, I mean, I can… if it worked, I can totally see a change in what we do. (SMO‐05)
	Defining significant injury	… it's the people who would clinically need some kind of intervention that I would definitely not want to miss. (SMO‐06) I don't know. I mean, possibly, because obviously, like, like they do neurosurgical intervention, then that's that is a more important thing. But I still think for patients, it is important that they know that they've got some sort of traumatic brain injury, because that does have a longer‐term impact on lots of other things. (SMO‐11) No, it's radiologically apparent but not clinically significant, I, personally wouldn't be emotionally as affected by that I wouldn't lose sleep as much. And as long as you can have a face‐to‐face discussion with the patient and the family, and you feel like you can explain this in a decent way, then that's much less significant than someone who ends up having a disability and needing home health or something because of a missed injury. (SMO‐01)
	Population factors	And a 20‐year‐old with a certain injury versus an 80‐year‐old with a certain injury. And like there's a much… there's a different threshold for doing intervention for neurosurgery, and depending on age, and functioning and all those other things. So yeah, my feel would be no, my feel would be no, but I think it'd be a hard thing to quantify. (SMO‐05) Now that I sort of, potentially work more regionally, I guess there's the implications that comes with sort of that, you know, that sort of patient population where you don't have a neurosurgeon readily at hand, which sort of does make a little bit of a difference of your… I guess of your tolerance, or of your kind of need to know the answer… (SMO‐10)

3.1.1. Safe Practice and Clinical Decision‐Making

3.1.1.1. Current Guideline Performance

Interviewed ED clinicians felt confident in their ability to determine which patients presenting to ED with head injuries required a CT head scan. Most clinicians use the NICE CT head injury guidelines or the similar Canadian CT head rule as a clinical decision aid. These tools were viewed as well‐functioning decision aids that align with preferred practice and clinical gestalt.

I use the NICE criteria. So that's just what I trained with in *****, they have the really lovely algorithm, which I found very easy to follow. For me, personally, I feel reasonably comfortable discharging people without a CT, as long as they meet criteria. (SMO‐01)

Population groups causing more indecision were the elderly, those on anticoagulants, the intoxicated and those with high concussion symptom burden. Clinicians noted that these groups were perceived as more at risk, or clinical assessment was more challenging. Often, it was felt that current guidelines were less helpful in these settings, meaning clinicians would err on the side of caution.

I have a lower threshold for CTing drunk people, just because they're sometimes quite hard to assess. (SMO‐06)

I think it's pretty clinician and department dependent on how much risk the department is willing to tolerate and what their capacity is for observation, and the kind of grey zone of CT heads which is increasingly that over 65 age group, or people who are on dual antiplatelets, or are anticoagulated, with evidence of very minor head injury. (SMO‐08)

3.1.1.2. Clinical Decision Support

In general, the theoretical addition of biomarkers to clinical guidelines was well received, particularly in low‐risk groups such as the young presenting with concussive symptoms. It was also noted that biomarkers could be particularly helpful for the cohorts of patients where current guidelines are less useful, such as the elderly, the anticoagulated, the intoxicated and those with altered mental status. Having an objective marker to support discharge decisions and to aid shared‐care decision‐making with patients was also commonly reported as desirable.

So yeah, my initial reaction is, sounds like a great piece of work and if the economics support it, too, then I'm sure a blood test is cheaper than a CT scan, and if it was quicker, maybe even point of care, you know… so a point of care test in particular would help those smaller regional rural hospitals. (SMO‐09)

Especially the elderly patients. I think that would be really, really, really useful. Because every time they're on blood thinners, every time they bump there head they get a CT. That would be really useful. (SMO‐02)

… sometimes that you are almost certain that this is a concussion. I think for those people, I'd feel better having something objective to hang my hat on. (SMO‐01)

The situations where biomarkers were thought to add the least clinical value were in the setting of high pre‐test probability of a significant injury, such as patients with reduced level of consciousness or clinical evidence of skull fractures.

Like that, you're worried about the base of skull or, you know, like, just consistently GCS 14 and altered, you know, something that like… I guess there's always a threshold, right, where your clinical gestalt becomes so strong that a biomarker would feel irrelevant to you. (SMO‐03)

3.1.1.3. Clinical Consensus and Individual Risk Tolerance

Safe practice was generally viewed as practice that would be supported by peers, accepting that there is a spectrum of individual risk tolerance impacted by both personality and clinical experience.

Because the main thing when you become a consultant is not being an outlier in your department, and therefore, you know, not making your colleagues feel uncomfortable with your practice, I guess, you know, just making sure that the amount of risk that you're willing to tolerate and stuff is on par with your colleagues. (SMO‐08)

I think individual practices always have to meet some like collective standard of care…. I talk about this spectrum of risk adversity and risk tolerance quite a lot. And I think all… I suppose all clinicians but, but certainly, in my experience, ED clinicians will place themselves somewhere along the spectrum, and some are very risk averse, and that probably wouldn't be acceptable to some people, and they would just prefer to have the scan to know, and then there is a group of people who are very risk tolerant, and you know, gung ho, and hopefully most people are somewhere in the middle. (SMO‐10)

Several clinicians discussed anxiety around the implications of missed injuries. This can lead to defensive practice, which sometimes equates to ordering CT head scans in borderline circumstances. Fatigue also impacts decision‐making, particularly towards the end of shifts. It was noted that biomarkers could also support decisions in this context.

And you look back on your practice, and I think there's a degree of all of us practice defensively to some extent. And that is definitely a factor. We've all heard of bad cases. I wouldn't say it's an excessive factor. I think most of the time scans are done appropriately. But we err on the side of caution. (SMO‐01)

3.1.2. Resources and Logistics

3.1.2.1. CT Head Accessibility

CT head scans were viewed positively as a quick, accessible investigation that provides good objective evidence to support admission and discharge decisions. While clinicians generally reported easy access to CT heads, most reported that waiting for CT heads was associated with delays to patient flow and ED discharge in certain circumstances. Delays were also more apparent when resources were overstretched.

I guess the resource that it impacts most is the delay for the patient to go to CT and then wait for the results; so they're taking up a bed space whilst we're doing all of that. (SMO‐11)

3.1.2.2. Over Investigation

The majority believed that at least some patient groups were over‐investigated in this context, although clinicians differed on who these patient groups were. Groups thought to be at risk of over‐investigation were similar to those identified when guidelines were less helpful: the elderly, the anticoagulated, the intoxicated, and the concussed.

I think we definitely overuse them in the anticoagulated patients who truly have very, like, they don't even classify as minor head injuries. I think the guidelines are misused in that sense. (SMO‐07)

Especially intoxicated patients, because you cannot use the Canadian head CT. Yeah, we over investigate those patients, especially they come in intoxicated and under the influence of some drugs or something like that. (SMO‐02)

3.1.2.3. ED Flow and Overcrowding

There were mixed opinions regarding the impact biomarkers could have on resources and logistics. The majority felt there was potential to reduce ED resource use and facilitate early discharge. Regarding regional/rural centres without an onsite CT scanner, it potentially saves the cost of transferring patients to larger centres.

“But I think there is a big role for using serum biomarkers. And potentially there would be a, you know, you could potentially reduce the number of CTs we do.” (SMO‐04)

Yeah, and it would save the patients a lot of time waiting, because as much as I say they don't take up too much scanning resources, they are still sat in a bed for two, three hours waiting for it to happen. (SMO‐07)

There was concern that biomarkers' high sensitivity and low specificity could increase CT head use if not used in an appropriately targeted population. Also, over time, biomarkers could be misused outside the scope of a proposed guideline, leading to inadvertent increases in CT head. Whether patients did not get a CT head because of a clinical biomarker would lead to increased representation rates was also raised.

Conversely, like, obviously, we need to balance out the specificity here and, you know…. using current examples, like… your troponin is, is an excellent, sensitive, and specific test, whereas the D‐dimer is super sensitive and not specific. And my worry is that that's the way the biomarkers will end up. (SMO‐04)

Like, there's no rebound back. I mean, if the biomarker comes back negative, and the patient goes home and there's no returning visits with worsening, you know, consequence of bleeding, yeah, I totally would go for the biomarker. So no return visits because of the head injury. Yeah, that would be my big thing. (SMO‐03)

3.1.3. Challenges of Changing Practice

3.1.3.1. Evidence‐Based Consensus

Despite the general opinion that biomarkers could provide clinical benefit, none would consider introducing them to practice until they were deemed to be the clinical standard for emergency care. Most viewed a solid evidence base demonstrating a biomarker performed with precision, including evidence from local healthcare systems, as essential prior to its use.

So, you come in with a rule, you've got a decent statistical power, and then it gets externally validated to the point where the cohort of Emergency Physicians can look at it and say, this is standard of care, and this is safe. (SMO‐01)

3.1.3.2. Culture Shift

It was noted that any change to practice will take time to saturate into routine use. Clinicians tend to be comfortable making decisions that align with their current practice and knowledge base.

Do you know, just because it's the current standard for practice, and it's the way everyone practices, I suspect it would take quite a lot of evidence to change that. Even if it was proven to be almost perfect, I think it would still take a lot of time and evidence. (SMO‐07)

3.1.3.3. Defining Clinically Significant Injuries

Risk tolerance for missed findings was low, ranging from 0.2% to 1%–2%. However, there was considerable debate regarding what would classify as a significant injury and what would count as a ‘missed injury’ versus an ‘incidental/inconsequential’ finding. Universally, any intracranial injury that required neurosurgery was deemed critical, and any miss rate here would need to be very low. The need to know about small findings that did not require neurosurgical input was disputed.

The question is, what's actually a miss? Yeah. So that you know, you get ones when they come back, and then somebody does a scan, and there's a small contusion or a tiny subdural is that a miss? I don't actually think that is a miss. (SMO‐03)

3.2. Population Factors

There were also wider considerations of changing practice, such as the impact on other specialties, particularly the neurosurgeons and the patients themselves, who likely need to be involved in clinical decisions and the development of new pathways.

Yeah, I mean, I guess it's around what you've just mentioned, which is risk appetite and risk trade‐offs. And I know that you know, so you ask an emergency physician and a neurosurgeon and a patient, and they will have a different answer to that. (SMO‐09)

Furthermore, there are notable differences between tertiary/urban and regional/rural settings, which may have different thresholds for observation or more difficult access to healthcare. New national guidelines would need to account for this. Any change in practice would need to be equitable and approved by various stakeholders.

We serve a population that can live two hours away from our centre. So identifying these patients who may not need neurosurgical intervention but do need to stay in hospital for very close observation is incredibly important to our health outcomes where I am. (SMO‐08)

And it's not only giving patients good care, it's getting rid of variability of care, which is a big concern in New Zealand, around equity, and so on. Not just ethnic equity but also regional equity and rural equity. And then it also provides physicians with, you know, a kind of safety in their practice in that if you can define what is reasonable to do; it protects that practice as well. (SMO‐09)

4. Discussion

The identified themes in this qualitative paper demonstrate that CT head decision‐making is a complex, multifaceted process. Future incorporation of biomarkers in CT head decision guidelines is generally viewed favourably, with the potential to support senior ED physician decisions and reduce ED resource pressure. However, until the strength of evidence is of high enough quality to be viewed as the standard of care for EDs, its clinical uptake will be minimal. This study provides a unique insight into the perceptions and expectations of senior clinical decision‐makers in ED about changing clinical practice in this context.

Senior ED physicians are generally confident in making CT head decisions with caution applied to population groups where current guidelines are less useful. The population groups that caused the most clinical concern or uncertainty were also the same population groups where guidelines were thought to be less useful. Furthermore, these were also the groups that clinicians believed were potentially over‐investigated. Many other individual, departmental and population factors impact CT head decision‐making, such as healthcare access, department patient flow and CT scanner availability.

Even though this study demonstrates optimism towards including biomarkers in CT head decision guidelines, the lack of high‐level evidence and institutional consensus deters their use. There is limited comparable qualitative data in this field. A study reviewing the development of a new guideline for paediatric head injury noted important factors affecting uptake included the need for national consensus and incorporation into tertiary hospital guidance [24]. Another study reported that physicians are more likely to adhere to a guideline if they can trust that it has been developed through a thorough process by appropriately competent people [25]. Finally, an Australian qualitative study noted the importance of increased education on using clinical decision tools and awareness of ‘low‐value’ care to reduce CT head ordering in low‐risk cohorts [26].

Interestingly, it was noted that even when such evidence is available, translating biomarkers into routine practice is expected to be slow and challenging. Historically, there are significant delays between the availability of research‐based evidence and its implementation into routine care [27, 28]. The reasons for this are beyond the scope of this paper, but it is important to note that introducing any new biomarker into routine practice will take time to become commonplace. Therefore, this study adds significant insight into how senior ED clinicians make CT head decisions and their expectations regarding introducing a new test.

Objective clinical tests are rarely perfect. They often have a margin of error, resulting in false negative results or ‘missed’ pathology [29]. The miss rate ED clinicians would be comfortable accepting is largely unknown for TBI [30]; this study suggests that the acceptable rate is very low. A challenge evident from this study is defining a clinically significant injury that meets both patient and doctor expectations. Currently, a biomarker's diagnostic performance is measured against any radiologically significant injury. Clinicians in this study were concerned this may lead to unintended increases in CT head scanning, given the high sensitivity of the test. Establishing a clearer consensus definition of what clinically significant injuries need to be identified with CT head in ED may help develop more precise biomarker thresholds.

Given that current CT head guidelines are used and well supported by senior ED clinicians, incorporating biomarkers within them makes sense. Limited evidence exists on how biomarkers perform with existing clinical decision rules. The addition of GFAP or S100B to the Canadian CT head rule (CCHR) improved the performance of this decision rule, and the combination of the biomarker and the decision rule was superior to either in isolation [31, 32]. The addition of S100B to a cohort of patients who meet NICE CT head injury guideline criteria demonstrated the potential to reduce CT head requests by one quarter [33]. In Scandinavia, where routine S100B has been included in neurotrauma guidelines for some time, S100B's diagnostic performance was high. However, guideline compliance overall was poor [34]. It is important to acknowledge that current clinical decision guidelines have limitations, particularly a lack of evidence demonstrating true improvements in clinical practice [35]. Therefore, the addition of biomarkers to these rules may provide an opportunity to review existing guidelines and test their real clinical value.

ED physicians in this study thought biomarkers may provide benefits such as reduced CT head rates, reduced wait times and objective support towards discharge decisions. There was also interest in how these biomarkers would work in populations where the current guidelines are less effective and potentially over‐investigated. Further work should focus particularly on these areas as improvements here have the most potential to improve current CT head guideline performance.

It was noted that individual clinician risk tolerance impacts CT head decision‐making, a factor reported in prior literature. ED physicians who are identified as being more risk averse are more likely to request CT heads, CT coronary angiograms and CT abdomens for paediatric head injuries, chest pain and abdominal pain, respectively [17, 30, 36]. Adding an objective biomarker may be of particular value to more risk‐averse physicians, but this needs further exploration.

5. Limitations

Thematic analysis in this study, particularly the reflexive methodology, can be criticised as simplistic; however, it was chosen for the flexibility it allows with this research question [22]. This method can also be prone to confirmation bias, and, notably, both authors are conducting other research investigating biomarker use, which may have unintentionally influenced the analysis. This study is also vulnerable to selection bias, given participants volunteered. This tends to select those with greater interest or knowledge in the topic. Despite reaching data saturation, the sample size is still small. This likely reflects the targeted topics and selection of senior ED physicians. It does, however, mean that wider stakeholders' opinions are not included, such as junior ED physicians, ED nurses, neurosurgeons, patients or broader community groups. It also only includes physicians working in New Zealand, so generalisability could be limited. Despite this, this paper provides unique insight into an important developing area of emergency care that is hoped will hold value for the reader.

6. Conclusion

Senior ED physicians viewed the potential inclusion of biomarkers in CT head guidelines favorably, particularly with regard to decision support in low‐risk head injuries. However, the lack of expert consensus or strong evidence demonstrating high diagnostic precision in comparable emergency care models limits their introduction.

Ethics Statement

The research was conducted following ethical approval from the University of Otago Human Ethics Committee (HEC Reference: H22/043). Māori consultation was reviewed by Ngāi Tahu Research Consultation Committee.

Conflicts of Interest

The authors declare no conflicts of interest.

Rogan A. and Larsen P., “Exploring Senior Emergency Physicians' Perspectives on the Inclusion of Clinical Biomarkers in Clinical Decision Rules for CT Head Use in Patients With Suspected Traumatic Brain Injury ,” Emergency Medicine Australasia 37, no. 5 (2025): e70144, 10.1111/1742-6723.70144.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.