This study was a retrospective evaluation of diagnostic yield of head CT referrals made from A&E Department of a tertiary care hospital located in northern-central India over three year period between January 2015 to December 2017 and was done by retrieving and systematically collating data from the CT report archives of the CT scan centre of the Radiology Department. Since, it was a retrospective study based purely on patient data hence ethical approval was not necessary.

Inclusion criteria: All the data of trauma and non trauma cases referred to the Radiology Department for Computed Tomography scan brain from Accident and Emergency Department between January 2015 to December 2017 was included in the study.

Exclusion criteria: To avoid a false high positive result follow-up scans with significant positive findings at first scan were excluded from the study.

Based on the diagnosis, the reports were first grouped into two major group’s trauma and non trauma referrals with further subdivision into positive and negative scan on the basis of presence of abnormality. Subsequently, the positive trauma reports were classified into major findings (three subgroups extra-axial haemorrhage, brain contusions and depressed skull fractures) and minor findings like undisplaced fractures of the skull or facial bones with or without subgaleal haematoma. The distinction was made only to separate out the more severe from the less severe findings; however no further classification was done on the basis of other parameters like presence or absence of cerebral oedema, midline shift or subfalcine or uncal herniation, as the results would have made the statistics complex and difficult to comment upon.

Further positivity and negativity rate of trauma referral cases was calculated depending upon their fulfilment of National Institute of Health and Care Excellence (NICE) guidelines for CT head in patients of head injury [2].

Non trauma reports were also further classified in the two major clinically important findings of haemorrhage and infarct. Minor findings like cerebral atrophy, non-specific white matter changes, sinusitis and mastoiditis were also recorded separately.

Results were expressed in terms of frequency and percentages and analysis was done in Microsoft Excel.

The audit in total collected 2185 traumatic and non traumatic referrals for NCCT head. Out of 2185, a total of 980 referrals for trauma were separated out from the rest of, other non traumatic causes [Table/Fig-1].

The result of the audit revealed, the out of 980 trauma cases that were sent for an urgent NCCT head scan; only 246 were having major positive findings as per the major criteria for trauma cases. Broadly, these were further classified into 3 subgroups-extra-axial haemorrhage (106), brain contusions (35) and depressed fractures (3) [Table/Fig-2]. Another 102 cases had only minor positive findings for trauma.

Out of the total 980 cases, only about 378 cases met the NICE criteria [2] for CT referral in head injury either directly or indirectly, out of these 378 cases, positive scans were 241 [Table/Fig-3]. The Glasgow Coma Scale (GCS) which is used as a quick examination tool in case of head injury or in cases of altered mental status was mentioned in bare 267 referral forms. An account of referring physician/surgeon was also made which brought out that a majority of the traumatic referrals were made by medical officers or surgery residents on call posted in the A&E Department (788 out of 980 trauma cases) [Table/Fig-4].

A total of 1205 number of referrals were grouped in to the major category of non traumatic referrals and 328 were found to be positive suggesting non traumatic patients had a slightly higher positive yield. A total of 817 cases were of recently developed weakness, slurring of speech, altered mental status suggestive of suspected stroke followed by headache which is another common presenting complaint in the Emergency Department. Yet again it was found that the number of referrals mentioning GCS in the referring note, were abysmally low, some of them with clinical notes of only the provisional diagnosis like altered mental status or moderate to severe headache. The incidental findings like non-specific demyelinating changes, sinusitis, and age related cerebral atrophy were also recorded which were presented in the vast majority of the patients (533 out of total 1205 non trauma cases) [Table/Fig-5].

As observed in trauma referrals, most of non trauma referrals were also by the A&E medical officers or medical residents (965 out of 1205). Referrals by physicians or neurosurgeons are relatively less (240 out of 1205). The cases referred for CT by medical officers or medical residents had poor yield as compare to specialist referrals [Table/Fig-6].

Just to avoid a false high positive result, about 217 cases (in past 3 years) were not included in the study which were follow-up scans and were already with clinically significant findings. Annual workload data of CT scan head showed that there was a nearly 15% increase in referrals over the past 3 years. The increase was primarily because of over-suspicion of stroke where the initial clinical work-up was lacking the details of a neurological examination.

The National Institute for Health and Care Excellence (NICE) has produced guidelines entitled ‘Head Injury in Infants, Children and Adults: Triage, Assessment, Investigation and Early Management’ [2] for early management of head injury. These guidelines include recommendation for initial assessment in A&E Department with indications for CT scan. As discussed by Hentel KD et al., due to various factors described, evidence-based guidelines and imaging criteria were not followed at A&E while referring patients for head CT leading to its inappropriate use and they have advocated greater scrutiny of imaging performed at A&E level [3]. A study done by Nawaz M et al., revealed that neuroimaging especially an urgent NCCT brain have a limited diagnostic role in conditions like headaches [4]. In present study, positive diagnostic yield for trauma and non trauma head scans were 25.1% and 27.2%, respectively, which were relatively lower as compared to similar earlier studies where positive results were ranging from 29% to 65.8% in trauma and 10% to 70.3% in non trauma [5-9].

In the present study the most common major positive finding was extra-axial haemorrhage (73.6%) followed by brain contusion (24.3%). Study by Waganekar A et al., also showed similar results with extra-axial haemorrhage being most common (68%) followed by brain contusion (32.9%). Similar results were also seen in other studies [9,10].

The present study demonstrated that the cases where NICE guidelines was followed, it led to a high diagnostic yield of 63.7% with percentage of positive cases meeting NICE criteria among total positive case is 97.96%. Similar findings were observed in other studies [2,11].

In this study, most of the scans were ordered by residents or medical officer from Emergency Department with poor diagnostic yield with positivity rate in trauma cases 17.4% and non trauma cases 19.4% as compared to specialist referral where positivity rate of scan was 56.8% and 58.8%, respectively. There are no known studies where comparisons of diagnostic yield out of referrals between a specialist/consultant and resident were done. However, Prevedello LM et al., suggested that experience is not related to ordering of head CTs [13]. As per Puri S et al., providers with more clinical experience are likely to consider previous imaging and discuss risks and benefits of imaging [14]. The attendance of a senior physician decreases the tendency for referral for imaging [15].

The exercise carried out by the Radiology Department of our hospital brought out that diagnostic yield of emergency head CT scans referred from A&E Department of our hospital is relatively low. On close scrutiny, it was found that the medical officers or the residents on call were more or less having partial knowledge of the good practice guidelines as mentioned above. Most of the normal reports that were generated were part of the documentation that was required to discharge the patient from the Emergency Department rather than a tool to evaluate and correlate the clinical findings with imaging. After analysing the results, following points were drawn behind the reason for low positive results in the urgent referrals:

Hence in view of the above observations, an audit Committee within the Department of Radiology is pivotal for a proper clinical audit and should be widely supplemented with participation of the departmental and managerial staff as well as referring physician’s and patients. Overuse of CT scan is a growing concern because of the financial implications and avoidable risks (radiation exposure) to the patient. Increase in the number of scans had actually increased the number of unnecessary scans to 15-30% which indirectly translates to an increase in the number of avoidable referrals as per our audit. Two major reasons that can be pin pointed are lack of adherence to the guidelines by the clinicians or practice of “defensive medicine” which is increasingly being done due to large scale corporatisation and increasing number of litigations in healthcare system [2,4].

As indicated by Baloescu C that to resolve issue of over imaging, an integrated approach at the level of health policy, government mandate, hospital administration, physician associations, research, and education are needed [16].

Hence, in view of above authors have following recommendations:

Small sample size and single centre study were the limitations. Another limitation was its retrospective nature as data was studied for limited time period.

Clinical audit is an essential process which any healthcare organisation can use to enhance and ensure the quality of the service it provides. The process in essence means sampling of performance and comparing the results with a pre-selected standard of good practice. If the standard is not upto the expected level, reasons for this are sought, changes implemented and a re-audit carried out to ensure improvement.