The utility of patient symptom questionnaires in CT of the lumbosacral spine

Introduction

Spinal pain and radiculopathy afflict millions worldwide, imposing significant costs on healthcare systems (1). Computed tomography (CT) of the lumbosacral spine is a frequently ordered investigation for the workup of lower back pain and radiculopathy in the community. Whilst magnetic resonance imaging (MRI) is more sensitive for symptoms arising from soft-tissue structures (i.e., nerve roots) like radiculopathy, it can be difficult to access, owing both to its cost and duration of study (2).

There exists a broad spectrum of spinal pathologies that can result in chronic lower back pain or radiculopathy. Broadly, these may encompass degenerative disc disease, intervertebral disc herniation, stenosis of the central canal, and subarticular zone, neural foramen or facet joint degeneration. Lower back pain and radiculopathy can be managed conservatively, or with minimally invasive procedures such as facet joint, epidural or foraminal corticosteroid injections. If more invasive therapy is required, there are a range of spinal decompression surgical options (3-5). Whilst it is ultimately the role of the treating clinician to integrate information from the patient’s presentation and imaging and determine a diagnosis, the role of the radiologist in providing a clear and useful report is essential in guiding timely and appropriate treatment. Similarly, if a treating clinician provides comprehensive and accurate clinical information, then this is beneficial to the radiologist and improves the quality of the investigation (6).

Currently, there is limited evidence that lumbosacral CT scans can identify clinically relevant pain generators (7-9). This is partly because accurately pinpointing the precise level and source of pain as relevant to the patient is often challenging for the radiologist; imaging can identify multiple possible sources of pain as well as incidental findings (1,10-12). One factor that contributes to this difficulty is poor clinical information, which leads to radiology reports listing structural abnormalities as potential symptom generators, instead of suggesting the most likely aetiology for the clinical presentation. Prior studies have shown that only a minority of patients with radiologically severe lumbar canal stenosis are symptomatic (13). In contrast to pain, there is a moderate level of evidence that CT scans can assist in identifying causes for radiculopathy (14,15). Despite this, for the radiologist, it is still difficult to determine whether notable radiological findings are of clinical significance without further information of the patient (6). These dilemmas necessitate the exploration of complementary tools to optimize diagnostic accuracy and guide treatment decisions.

Patient questionnaires emerge as a promising avenue, capable of capturing crucial clinical information not apparent from imaging alone (10,16). A previous study by Balza et al. showed the addition of a symptom questionnaire in patients undergoing spinal injection for pain or radiculopathy improved the diagnostic confidence of interpreting MRI of the spine in localising a specific pain generator when compared to those without any additional clinical information (16).

Hence, the aim of this study is to examine the effectiveness of patient questionnaires in assisting reporting of lumbosacral CT scans for patients with lower back pain and those with lumbar spine-related sensory symptoms.

Methods

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Central Adelaide Local Health Network Human Research Ethics Committee (CALHN HREC) in accordance with the National Health and Medical Research Council (NHMRC) National Statement on Ethical Conduct in Human Research (2023) on 17/06/2024 (Reference number: 19490). A waiver of consent was granted by the ethical review board as the study involved a retrospective review of medical records, and there were no foreseeable risks from participation in the study. It was conducted at Jones Radiology, a private radiologic practice which has multiple clinic locations.

A retrospective audit was performed of 250 consecutive patients who underwent a CT of the lumbar spine between 12/01/2024–22/01/2024. Inclusion criteria included the patient being a lumbar spine CT patient, with lower back pain, or lumbar spine-related sensory symptoms. Exclusion criteria included age less than 18 years, spinal surgical intervention in the prior 6 months, or history of recent trauma in the clinical information.

As a routine for all outpatient CT examinations of the lumbar spine at Jones Radiology, at the time of appointment, patients were asked to fill out a symptom questionnaire either electronically or on paper as designed by the investigators of the study (Figure 1). To ensure patient compliance, the questionnaire was made to be as simple as feasible, with questions being straightforward and few. All questions were fit onto one A4 page. The questionnaire was written in English, which was widely-spoken in our patient cohort. The symptom questionnaire elicited information regarding the presence of pain, altered sensation, or both, as well as indicating graphically the locations if present. These questions were left open-ended and allowed patients to shade in areas which they felt were affected, rather than influencing their responses through providing dermatomes. In addition, information regarding previous spinal imaging or interventions that may have been done at other institutions was sought. Symptom questionnaires were scanned into the picture archiving and communication system (PACS) and were readily available to the reporting radiologist alongside the CT images.

The CT spines were reported by a pool of 30 consultant radiologists at a private outpatient radiology practice. Radiological reports and symptom questionnaires were interrogated for further information. The presence of lower back pain was noted. The presence of lumbar spine-related sensory symptoms such as altered sensation or limb pain was also noted. Whether the radiological report addressed the clinical request information, either by including relevant positive or negative findings, was categorised as either “yes” or “no”. Whether the radiological report made reference to the questionnaire was scored as “yes” or “no”. Whether questionnaires were congruent with the provided clinical request information was scored as “yes” or “no”. “No” was scored if the two sources differed on whether patient had lower back pain or lumbar spine-related sensory symptoms or if questionnaires reported different locations of pain or altered sensation in comparison to the clinical request information.

Data collection was completed by all three investigators. Investigators were not blinded to patient information. To structure and standardise data collection, data entry during the analysis stage consisted mostly of inputting discrete data, i.e., “yes” and “no” responses. To create inter-rater consistency and ensure standard application of the criteria, all three investigators initially evaluated a subset of cases together. After these first few cases, the remaining cases were split evenly between the three investigators. In individual cases where data collection was unclear, all investigators would first independently review and then reach a result through consensus discussion.

Results

From the 250 questionnaires administered, one case was removed from the analysis sample as peer-review feedback identified this patient did not have back pain or lumbar spine-related sensory symptoms, thus not fitting with the inclusion criteria. Of these 249 questionnaires, a total of 214 were completed and included for analysis, whilst 35 were either not completed or uploaded. Of these 214, 132 (61.7%) were female and 82 (38.3%) were male. Mean age for patients was 63.4 years. All 30 consultant radiologists at the practice were involved with reporting.

Of the 214 patients who completed the questionnaire, 211 of them had back pain (98.6%) compared to 3 who did not have pain (1.4%). One hundred sixty-four (76.6%) patients had lumbar spine-related sensory symptoms (altered sensation or limb pain) compared to 50 (23.4%) who did not. This is shown in Table 1. One hundred sixty-one (75.2%) had both lower back pain and lumbar spine-related sensory symptoms. Fifty (23.4%) patients had lower back pain without lumbar spine-related sensory symptoms, and 3 (1.4%) patients had lumbar spine-related sensory symptoms without pain. See Figure 2 for a Venn diagram showing this data. Fifty-three (24.8%) patients had either only lower back pain or only lumbar spine-related sensory symptoms.

In the 214 cases with completed questionnaires, 184 radiological reports referenced the questionnaires (86%), whereas 30 (14%) did not. See Table 2 for additional information regarding the number of questionnaires which matched the clinical request information.

Answers from the questionnaire were compared to the imaging request form, to ensure the symptomatology was consistent with the clinical request information (see Table 2). Of the 214 questionnaires that were completed, 33 (15.4%) had information that suggested alternate pathology to what was indicated on the questionnaire. In these 33 cases, 28 (84.8%) radiological reports made direct reference to the patient questionnaire. Of the 164 patients with lumbar spine-related sensory symptoms, 24 (14.6%) had incongruent clinical request information with questionnaires. Of these 24 patients, 19 (79.2%) had radiological reports which made reference to the questionnaire. Of the 50 patients with lower back pain as their sole symptom, 9 (18%) had incongruent clinical request information with questionnaires. For all 9 (100%) of these cases, radiological reports referenced the questionnaires.

In one case, the report’s findings correlated with the clinical request, but differed to what was described by the patient questionnaire.

Of the 214 patients, the questionnaire identified that 199 (93%) patients reported external lower back imaging. Of the 164 patients with lumbar spine-related sensory symptoms, 151 (92.1%) reported external lower back imaging, and of the 50 patients with lower back pain as their sole symptom, 48 (96%) reported external lower back imaging (see Table 3).

Discussion

We observed that there was a high rate of questionnaire completion and a high rate of usage by radiologists. Of the 249 patients, 214 (85.9%) completed the questionnaire. They were reviewed by radiologists in at least 184 cases (86%), as this number of reports directly mentioned the questionnaire. From this, it can be inferred that the questionnaire was accessible for both patients and radiologists. Whilst we note that compliance and radiologist usage may differ outside of the context of a study, we recognise that these findings suggest that the questionnaire could integrate well into regular workflows.

The use of a questionnaire for lumbar CT imaging has value in identifying information that may be inaccurate. 15.4% (n=33/214) of total completed questionnaires contained information which differed from the clinical request information. This was similarly seen for patients with lumbar spine-related sensory symptoms as well as patients with lower back pain as their only symptom with 14.6% (n=24/164) and 18% (n=9/50) having incongruent clinical request information and questionnaire data, respectively. Whilst there is the possibility that some patients may have made errors in filling out the questionnaire, we recognise that patients are best equipped to describe their symptoms and thus assume that it is more likely for clinical request information to be inaccurate for these cases.

We observed that for those with incongruent information between their request form and questionnaire, 78.7% (n=26/33) of radiological reports made reference to the patient questionnaire. Similarly high percentages were noted for those with lumbar spine-related sensory symptoms and those with lower back pain as their sole symptom with 79.2% (n=9/24) and 100% (n=9/9) of radiological reports making reference to the patient questionnaire in cases of incongruent information, respectively. Thus, we note that the patient questionnaire was beneficial for the reporting radiologists as it could provide them an additional perspective, which could then be incorporated into the final report.

We note that 93% of patients (n=200/215) reported external imaging for the lower back on the questionnaire. Similar percentages were noted for those with lumbar spine-related sensory symptoms (92.1%, 151/164) and for those with lower back pain as their sole symptom (96%, n=48/50). Routinely searching for external imaging across various health systems is inefficient and exerts a significant time burden, especially when previous imaging is not guaranteed. If the radiologist, or clerical staff, is made aware of the presence of previous imaging, this can streamline the process of retrieval, allowing external imaging to be accessed. Thus, any information on prior lower back imaging is highly valuable to the reporting radiologist and we note that, unless patients had previous scans with the same practice, this information would often not otherwise be available. Thus, we consider that this is an additional benefit for the use of a patient questionnaire.

A limitation of the study was that physical examination findings were not incorporated alongside the questionnaire, it is important to recall that patient symptomatology is only part of the clinical assessment, as physical examination plays a vital role, and this cannot be elicited without a face-to-face consultation, which is well outside the time constraints and practice of a busy radiologic centre.

Moreover, a limitation of the study design was the lack of a method to determine whether the questionnaire had been accessed by the radiologist. Apart from mention within the radiological report, there was no direct indication that the reporting radiologist had reviewed the symptom questionnaire or factored in the responses during interpretation of the imaging.

Furthermore, there were limitations of the questionnaire used. Whilst our patient population lives in an area where English is largely spoken and is the official language, the questionnaire required the patient to have competence with reading and writing English and health literacy in regards to their symptoms. Whilst simplicity was necessary to achieve compliance, information about the quality of pain, such as dull or sharp, or the presence of weakness was not ascertained. Moreover, the open-ended nature of identifying symptom location led to poor dermatomal classification. An alternative would be a series of questions referring to a specific dermatomal distribution as outlined by Balza et al., or incorporating validated questionnaires, such as the Zurich claudication questionnaire to facilitate an objective assessment (16,17). In addition, the symptom questionnaire is a subjective assessment, thus there was a potential for recall bias from patients.

Whilst the retrospective design of this study allowed for the analysis of the various descriptive benefits of patient questionnaires and evaluation of the effectiveness of patient questionnaires in current workflows, it was limited in providing proof of utility in contrast to a comparative design. Whilst it is outside of the scope of this study, we note that future research with a comparative design, for example by including a control group, could elucidate more data about the utility of patient symptom questionnaires and draw stronger conclusions. At present, we note that our findings are more exploratory rather than serving as definitive evidence of clinical utility.

Clinical information contained within the imaging request only provides a snapshot at a particular time. At the time of imaging, which is not infrequently days to weeks after initial assessment, symptomatology can and often will change, in terms of nature, distribution and severity, and this information is crucial to generating an accurate radiological report that is relevant to the patient. For example, lumbar intervertebral disc herniation, has a high rate of spontaneous resolution, depending on the type and size of herniation (18,19). Given that a symptom questionnaire is completed on the day of CT, it provides an additional view into symptom timelines when combined with clinical request information.

Clinical information is of high importance to the reporting radiologist; however, in private practice, radiologists do not have access to patient information apart from previous scans within the same practice or the clinical details provided in the request form. Questionnaires, such as the one used in this study, have a role in providing more information, as they are an organised account of symptomatology and can be easily uploaded to the PACS as a single image. Further studies on questionnaires for differing presentations have utility in further exploring the benefits that these have in diagnosis.

Whilst it is outside of the scope of this study, it is important to note that there are likely also benefit to patients and doctors, apart from radiologists, in the use of this symptom questionnaire. To the patient, it can provide them with autonomy in describing their symptoms, specifically noting the location affected and whether it is pain or altered sensation. To the referring doctor, it can provide them with a visual and written description of patient symptoms, further assisting their clinical assessment. For doctors who may become involved with patient care at a later stage, this questionnaire could provide a valuable snapshot of patient symptomatology as recorded at a specific point in time. However, we recognise that implementing such a questionnaire would require additional paperwork to be completed by patients and uploaded by staff.

Conclusions

Inclusion of a symptom questionnaire is beneficial to the reporting radiologist. Questionnaires can supplement clinical request information and provide a snapshot of the clinical presentation at the time of CT scan. We observed a high completion and usage rate of the questionnaire, which may indicate feasible use in regular workflows. Importantly, we noted that patient questionnaires help with providing radiologists accurate clinical information and are especially important for cases in which there are differing accounts of patient symptomatology, both for patients with lumbar spine-related sensory symptoms and for those with back pain as a sole symptom. Furthermore, they can inform radiologists when external imaging is available. These benefits can assist the radiologist in providing higher quality reports, ideally resulting in better patient outcomes.

Supplementary

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Central Adelaide Local Health Network Human Research Ethics Committee (CALHN HREC) in accordance with the National Health and Medical Research Council (NHMRC) National Statement on Ethical Conduct in Human Research (2023) on 17/06/2024 (Reference number: 19490), and individual consent for this retrospective audit was waived.

Data Sharing Statement

Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0548/dss