Musculoskeletal chest pain prevalence in emergency department presentations: A retrospective case notes review

Musculoskeletal (MSK) causes of chest pain are considered common in emergency care, yet management is limited, reported outcomes are poor and prevalence data in New Zealand are lacking. The present study aims to estimate the prevalence of MSK chest pain in New Zealand EDs and describe the characteristics of MSK chest pain cases.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Accepted 10 November 2023
Key findings In the ED, the clinical imperative is to identify life-threatening causes of chest pain and prioritise risk stratification of cardiac chest pain.Once this has occurred, establishing a single definitive diagnosis for low-risk patients can be challenging during a single ED encounter.This is reflected in diagnostic categories such as atypical, nonspecific, or undifferentiated chest pain.This can lead to nonspecific management, for example, reassurance that the symptoms are not related to the cardiovascular system. 9MSK chest pain is often described as a self-limiting inflammation of the sternocostal or costochondral joints, 5 and management recommendations within the emergency environment are commonly limited to reassurance of the benign cause, rest and analgesia. 9,10][13] As a first step towards developing and testing interventions for improving outcomes for people with MSK chest pain, the purpose of the present study was to estimate the prevalence of MSK chest pain in EDs in the lower South Island of New Zealand.

Study design
A retrospective chart review of patients presenting with a primary complaint of chest pain in four EDs was performed.Ethical approval was granted by the University of Otago Human Ethics Committee (HD21/038).The study protocol was developed in line with recommendations by Vassar and Holzmann 14 for retrospective chart reviews.(patient self-referred) EDs.Patients in the Dunstan Hospital catchment are initially assessed by primary care providers (usually general practitioner [GP]) and referred to the hospital if they need secondary care.This alters the presenting case mix.Included in the study were all cases presenting to the included hospitals' ED or equivalent care facilities, with chest pain recorded in electronic notes as a diagnosis code, diagnosis description or triage presenting complaint between 1 March 2021 and 31 May 2021.We excluded people who: were under the age of 16, did not wait for treatment, did not have medical notes to review, or on case review did not have a primary complaint of chest pain.

Sample size estimation
Raw data indicated that there were 5718 total chest pain presentations over a calendar year in Dunedin, Southland and Lakes hospitals.When estimating the sample size, we assumed an expected MSK chest pain prevalence of 16% 6 and a potential data loss of 35% due to incomplete/excluded records or difficulties with diagnostic codes.Our calculation suggests the minimum required sample size was 319 for a margin of error of 5% in estimating the prevalence of MSK chest pain, with a confidence level of 95%.

Data extraction and categorisation
Electronic records for cases were identified by administrative hospital staff, extracted into a spreadsheet in de-identified format, further de-identified by an ED clinician (SB), and then shared with the research team.A research assistant (SD) and the primary investigator (EK) reviewed the extracted data and trialled coding of a small sample of cases (approximately 20).From this trial, we categorised cases based primarily on the organ system involved (e.g.[3]9 This method was determined to be the best fit for the study purpose and level of detail available in extracted data.Category definitions were guided by Dynamed diagnostic information (https://www.dynamed.com/)and the primary system affected.ACS included a spectrum of unstable angina, ST and non-ST elevation MI; other cardiac included all other cardiovascular conditions; MSK included conditions arising from MSK structures of the chest wall, spine and upper quadrant.
The research assistant received training from the primary investigator (EK) on how to navigate and Using the procedure manual, a pilot of 10% of all records (n = 142) was completed by the research assistant.This was followed by further discussion with the primary investigator (EK) and emergency consultant (SB).Resulting in further refinement of the procedure manual.For example, piloting led to a change from multiple to single categorisation, as it was observed that the case notes did not readily support multiple categorisation and that cases where the system(s) involved were unclear were better categorised as 'undifferentiated'.
The procedure for categorising cases was (i) review the diagnosis code; using specific codes indicating a clear diagnosis to categorise initially and (ii) if unclear, the clinical notes were reviewed and utilised the doctor's final 'impression' to categorise.Cases with an impression that was unclear or indicated multiple potential systems were categorised as undifferentiated.Cases in which the cause of chest pain was identified, yet not well represented by the categories were recorded as 'Other' (e.g.symptomatic anaemia and cellulitis).Repeat visitors were identified if the unique patient ID number was present more than once in the extracted data.Admissions were identified if a request for a bed was made, or based on the discharge destination where bed request data were not available.MSK cases were further subclassified by an emergency consultant (SB).
The research assistant completed full data extraction independently, categorising each case into the single best category.The research assistant identified challenging cases for independent review (EK) and a final categorisation.After a 1 month

Data analysis
The data were analysed independently (MDB), and descriptive statistics were produced using Microsoft Excel ® version 16.62.To estimate inter and intra-rater reliability, we used Cohen's kappa analyses and calculated those using IBM ® SPSS ® version 27.The strength of agreement from this would be based on categories outlined by Landis and Koch. 15

Results
Over the 3 months, 24 454 patients presented to Dunedin, Lakes and Southland hospital EDs, of which 1420 presented with chest pain (5.80%).After screening, 1344 cases from all four hospitals were categorised for analysis (Fig. 1).Demographic information is presented in

Discussion
The present study indicates that the MSK system is a common distinct source of chest pain presentations to EDs, worthy of further attention.Risk stratification of cardiac chest pain is the focus of early ED care, importantly however, the findings re-iterate that a majority (79%) of patients presenting to ED will ultimately be diagnosed with noncardiac or undifferentiated chest pain in line with findings elsewhere. 3,16,17SK causes were the second most common system responsible for chest pain presentations, second only to the cardiovascular system (represented in the present study in both the ACS and other cardiac categories).The estimated prevalence of 15% (11-31% across sites) is most likely a conservative estimate due to several factors: the emergency context prioritising identification of non-MSK causes; a large proportion of undifferentiated cases; study procedures requiring a clear impression of MSK chest pain; and noting that cases of clear trauma like motor vehicle accidents were not represented in these data.The MSK system is accepted as a common cause of chest pain in other studies. 9,18,19tudies examining the causes of noncardiac chest pain found MSK causes could be responsible for more than 44% of noncardiac chest pain, 7,17 whereas studies examining the causes of general chest pain presentationssimilar to the present studyestimate the prevalence of MSK chest pain between 4% and 30%. 1,3,5An overall prevalence estimate of 15% aligns well with a recent systematic review estimating the worldwide pooled prevalence of MSK chest pain in EDs to be 16%. 6Mandrekar and Venkatesan 6 describe the prevalence of MSK chest pain as varying with the population setting, with urban prevalence (13%) slightly less than the overall pooled prevalence (16%).This was also observed in the present study, with the most urban hospital involved (Dunedin) having the lowest prevalence (11%) of MSK chest pain.The authors call for more data from rural settings, age and gender-specific data, and MSK chest pain sub-types which are reported here.
The present study offers further insight into the features of MSK chest pain.Those with MSK chest pain were on average younger (Tables 1  and 4), consistent with reports that people with noncardiac chest pain are younger than people with cardiac causes; 9 and that people with MSK chest pain have fewer pre-existing comorbidities such as ischaemic heart disease and diabetes than those with cardiac chest pain. 20,21Despite a younger tendency, MSK chest pain still affects all ages as illustrated in Table 4, again consistent with other studies. 19,21In our data, the prevalence of MSK chest pain appears lower in those above age 65 years (7-10%) than in those below age 65 (14-30%).Notable differences in the prevalence of MSK chest pain were neither observed across gender (Table 3) in line with other studies; 19,22 nor across ethnicities.
MSK sub-types were most often nonspecific, reflective of other priorities in the emergency setting and the challenges with pinpointing the exact source of MSK pain in a brief ED encounter.Costochondritis was the most common specific sub-type, notably observed across age and gender (Tables 5 and 6).Only a small proportion of fractures were observed.These data offer more detail than previous reports. 6No increase in prevalence of MSK chest pain was observed in repeat visitor cases.Suggestions that patients with MSK chest pain generate more repeat visits due to recurrent symptoms and anxiety 9,22,23 are not supported by our data.There should be some caution when comparing the study sites to each other, due to the differing models of care involved (e.g. if people have seen their GP prior) and many other contextual differences.It is more notable that MSK chest pain accounted for at least 11% of chest pain presentations across a range of ED settings.
MSK chest pain remains challenging to recognise in ED settings.In all chest pain patients, the priority is to exclude potentially serious and lifethreatening causes.Work to expedite the identification of those with high-risk chest pain 24,25 supports the identification of low-risk groups, but this is not sufficient to positively identify MSK chest pain in these settings.A high level of undifferentiated cases (47%) observed in the present study and other research 1,9 highlights the challenge of identifying cause(s).Yet this is important as identifying MSK causes guides effective management, as well as potentially reducing unnecessary investigations and stress in those presenting with chest pain. 23hose with MSK chest pain can benefit from physiotherapy interventions, including manual therapy and stretching. 11,12,26,27However, those with MSK chest pain are rarely referred to physiotherapy. 17Positive identification of MSK chest pain facilitates specific management and referral to appropriate services like physiotherapy.Although excluding serious causes of chest pain should always be the first step of any chest pain presentation, more steps are required to deliver effective chest pain management.

Limitations
Several limitations of the present study arise from the retrospective design.Categorisation was primarily based on the documented final impression of the doctor, which was time-consuming and required judgement.This is observed in lower (but still acceptable) inter-rater reliability between the research assistant and an ED consultant.On closer examination, cases of disagreement appeared to primarily represent greater use of the category 'undifferentiated' by the research assistant.As a result, the undifferentiated category may be overrepresented, whereas other chest pain categories, including MSK chest pain, represent conservative estimates.This is consistent with the objective to provide a robust conservative estimate of the prevalence of MSK chest pain in ED settings, recognising that further MSK chest pain cases will be identified in evaluation beyond the ED setting (e.g. via cardiology and general practice).The methods described support categorisation of MSK chest pain only when clearly documented.Data analysis has been limited to descriptive statistics in line with the aims of the study and the retrospective design.
The 3 months covered by the study represented a manageable volume of data for analysis involving a review of the medical impression.Scoping work revealed that hospital diagnostic codes were insufficient to identify MSK causes, requiring a timeintensive review of individual case notes.Although a longer study period (e.g. a full year) may have been ideal, seasonal factors were considered unlikely to affect the results, and 1473 records across multiple hospital sites were considered a suitable data set for the study purpose.

Conclusion
The present study indicates that MSK causes represent 15% (range 11-31%) of people presenting to New Zealand EDs with chest pain.Despite being a common problem, MSK chest pain remains a challenge to identify and manage.These data provide a basis and rationale for further work to improve health services for people with MSK chest pain.

TABLE 1 .
Demographic characteristics of adult ED cases presenting with chest pain, categorised by cause

TABLE 2 .
Case categorisation of adult ED cases presenting with chest pain, by hospital site

TABLE 3 .
Case categorisation of adult ED cases presenting with chest pain, by gender

TABLE 4 .
Case categorisation of adult ED cases presenting with chest pain, by age

TABLE 5 .
Musculoskeletal (MSK) sub-type categorisation of adult ED cases, by gender

TABLE 6 .
Musculoskeletal (MSK) sub-type categorisation of adult ED cases, by age