Understanding bias in EMS STEMI data: a national service-evaluation study of deprivation, behavioural pathways and inequality metrics in Wales

Understanding bias in EMS STEMI data: a national service-evaluation study of deprivation, behavioural pathways and inequality metrics in Wales | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Understanding bias in EMS STEMI data: a national service-evaluation study of deprivation, behavioural pathways and inequality metrics in Wales Adam Nicholls, Luke Watkins This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8307622/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Background Composite deprivation indices are widely used in health services research but may introduce endogeneity when health-related indicators are included in the deprivation measure itself. This study examined whether socioeconomic gradients in emergency cardiac events reflect true disease burden or are distorted by index composition and behavioural pathway selection. Methods We conducted a national evaluation using linked electronic patient care records from the Welsh Ambulance Services NHS Trust (2021–2025). Deprivation gradients in ST-elevation myocardial infarction (STEMI) and out-of-hospital cardiac arrest (OOHCA) were analysed across 999 and 111 pathways using three WIMD variants: full, Health-domain–excluded (“Minus Health”), and Income-only. Inequality was quantified using Slope and Relative Indices of Inequality (SII, RII), with quadratic terms testing for non-linearity. Results Among 12,241 EMS-attended cases (2,515 STEMI; 9,726 OOHCA), STEMI showed an inverse-U deprivation gradient that peaked in mid-quintiles, particularly for 111 users, while OOHCA displayed a strong, linear increase with deprivation. Removing the Health domain altered STEMI gradients but had minimal impact on OOHCA, suggesting measurement bias in the former. The 999 STEMI pathway showed a more consistent, monotonic gradient aligned with biological risk, whereas the 111 pathway was less stable and more affected by behavioural factors such as help-seeking and symptom appraisal. Conclusions Inequalities in EMS STEMI data reflect both true deprivation-linked disease and behavioural capture effects shaped by index design. Behaviour-independent comparators like OOHCA and the use of Health-excluded indices offer more valid assessments of inequality. Non-linear modelling is essential, as standard linear metrics (SII, RII) may obscure complex relationships. Crucially, these findings suggest that no dataset is behaviourally neutral: even within a single EMS system, pathway-specific behaviours influence visibility and gradient shape. Accurate interpretation of service-based health data must therefore account for both measurement structure and access dynamics to avoid misrepresenting need and misallocating resources. Emergency medical services Socioeconomic deprivation ST-elevation myocardial infarction Health inequalities Behavioural endogeneity Welsh Index of Multiple Deprivation Out-of-hospital cardiac arrest Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Background Composite measures of deprivation are widely used in public health research to quantify area-level disadvantage and assess its impact on health outcomes (Noble et al., 2006). In Wales, the official measure is the Welsh Index of Multiple Deprivation (WIMD) 2019, comprising eight weighted domains: Income, Employment, Education, Access to Services, Housing, Community Safety, Physical Environment, and Health (Welsh Government, 2019). Although WIMD is integral to health services research and policy planning, methodological limitations arise when it is applied to health outcomes. The Health domain incorporates indicators reflecting chronic disease burden, including GP-recorded conditions, cancer incidence, mental health diagnoses, limiting long-term illness, premature mortality, childhood obesity, and low birth weight. Because these indicators are conceptually close to many health outcomes, the risk of endogeneity bias is introduced — a form of circularity where part of what defines deprivation also defines the health outcome being studied (Hill et al., 2021; Jordan et al., 2004). This bias has been empirically demonstrated. Mohammed et al. (2025) showed that excluding the Health domain from WIMD reclassified 17% of Welsh LSOAs and attenuated odds ratios for the association between deprivation and diabetes. Similar results have been reported in England and Scotland (Adams & White, 2006; Bradford et al., 2023). However, most prior work has focused on chronic conditions, leaving uncertainty over whether index design also influences the representation of acute events. This is not trivial: indices developed to study long-term morbidity are now routinely used to analyse and plan services for acute disease, despite potentially different causal mechanisms. While chronic conditions reflect cumulative exposure to deprivation, acute cardiac events provide a sharper test of whether deprivation operates through biological risk, behavioural processes, or healthcare access. The INTERHEART study demonstrated that modifiable factors such as smoking, obesity, dyslipidaemia, hypertension, and diabetes account for over 90% of the global risk of myocardial infarction (Yusuf et al., 2004), and these risks are follow the socioeconomic gradient (Foster et al., 2018). ST-elevation myocardial infarction (STEMI) is therefore an acute manifestation of chronic cardiovascular pathology. Yet, its appearance in emergency medical system (EMS) data depends on socially patterned help-seeking, pathway choice, and system triage. By contrast, out-of-hospital cardiac arrest (OOHCA) is behaviourally independent once witnessed, as it almost invariably results in a 999 call. Comparing these two conditions offers a means to distinguish true deprivation-linked disease risk from artefacts of healthcare interface and pathway engagement. Beyond these conceptual concerns, there are also methodological limitations in how inequality is routinely quantified. Indices such as the Slope Index of Inequality (SII) and Relative Index of Inequality (RII) assume a linear, monotonic relationship between deprivation and health. Simulation studies (Renard et al., 2019; Moreno-Betancur et al., 2015) show that these measures can produce paradoxical or misleading results when the true association is non-linear — an issue that has received little empirical investigation in real-world health data. This evaluation addresses both gaps. Using national Welsh ambulance data, we examine how socioeconomic deprivation shapes the capture of acute cardiac events within the EMS. We analyse deprivation gradients in STEMI and OOHCA across the 999 and 111-transfer pathways, evaluating three variants of WIMD: the original version (with Health), a Health-domain–excluded version (“Minus Health”), and an Income-only version. By contrasting a behaviour-dependent condition (STEMI) with a behaviour-independent comparator (OOHCA), and by integrating linear and quadratic inequality models, we test whether observed inequalities reflect true disease burden, measurement artefact, or behavioural endogeneity within EMS data. Objective: To examine how socioeconomic deprivation influences the representation of acute cardiac events within the EMS, and whether observed inequalities reflect true disease burden or artefacts of measurement or behaviour. Using national Welsh ambulance data, deprivation gradients in STEMI and OOHCA were compared across 999 and 111 pathways, testing sensitivity to alternative WIMD specifications. Findings are intended to inform interpretation of routinely collected EMS data and improve understanding of how deprivation measures may influence statistical representations of acute health events, particularly in the context of EMS utilisation. Methods Study design This project 3 used routinely collected individual level ePCR data from the WAST between September 2021 and June 2025. The primary aim was to understand how socioeconomic deprivation, as measured by different variants of the WIMD, influences the capture and characterisation of acute cardiac events within the EMS. By linking ePCR records to area-level deprivation indices, this evaluation sought to assess whether the inclusion of health-related domains within composite deprivation measures affects the observed distribution and inequalities in EMS data. Analyses were repeated using the full WIMD, a Health-domain–excluded version (“Minus Health”), and an Income-only version to evaluate the sensitivity of results to methodological choices in deprivation measurement. Ethics approval and consent to participate This analysis was classified as a service evaluation under the governance of WAST. In line with Health Research Authority (HRA) guidance, NHS Research Ethics Committee (REC) approval was not required. In line with this, the need for ethical approval was waived according to these regulations. Cohorts STEMI Cohort The STEMI cohort included all ePCR records in which the paramedic closure code indicated ST-elevation myocardial infarction (STEMI), confirmed by an electrocardiogram. All clear comorbid diseases that may have preceded the STEMI were controlled for. Cases were analysed by pathway of entry: 999 pathway – direct emergency activation via a 999 call. 111 pathway – calls initiated through NHS 111 before ambulance dispatch. This allowed comparison of deprivation gradients across different access routes into the same emergency system. Out-of-Hospital-Cardiac Arrest (OOHCA) Cohort The OOHCA cohort included all ePCR records with a paramedic closure code of cardiac arrest, accompanied by a Cardiac Arrest Form. OOHCA represents a behaviour-independent comparator, as nearly all witnessed arrests generate a 999 response, providing a benchmark less influenced by help-seeking behaviour or pathway selection. For both cohorts, patients who died at scene were included if a closure diagnosis was recorded. Recognised Life Extinct (ROLE) cases without closure codes were excluded. Exclusion criteria can be found in Supplementary Table 1. Analytic Approach Deprivation gradients were quantified using two regression-based inequality metrics: Slope Index of Inequality (SII): absolute difference across deprivation quintiles, estimated by weighted linear regression of outcome prevalence on ridit scores. Relative Index of Inequality (RII): relative difference, estimated by weighted logistic regression using the same ridit structure. Analyses were conducted separately for: 999, 111, and combined STEMI datasets, and OOHCA as a contrast condition. Quadratic terms were included to test for non-linear relationships between deprivation and case share. A logistic regression comparing 111 vs 999 capture assessed whether pathway selection varied with deprivation after adjusting for age and sex. This design enabled differentiation between deprivation effects driven by true disease burden and those arising from pathway-mediated capture within the emergency system. Results Study Population A total of 12,241 EMS-attended cases were analysed: 2,515 STEMI and 9,726 out-of-hospital cardiac arrests (OOHCA). Across both conditions, two-thirds were male, and over three-quarters were aged ≥60 years. STEMI included a higher proportion of adults aged 40–59 (18% vs 12% in OOHCA). Among STEMI cases, 89% entered via the 999 pathway and 11% via 111 transfers. All had confirmed ambulance attendance. Combined STEMI analysis When 999 and 111 cases were combined, STEMI displayed a non-linear, inverse-U gradient across deprivation quintiles. Case share rose from 19% in the most deprived to 22% in the mid-quintiles, then declined to 16% in the least deprived (Figure 1). Insert Figure 1 here. This curvature persisted after age–sex standardisation, indicating that demographic structure did not account for the shape. Removing the WIMD Health domain accentuated the mid-quintile peak, while the Income-only version slightly weakened it. The similarity between the standardised and Income-only curves suggests that the Health domain was masking true deprivation differences, whereas demographic effects were minor. The remaining curvature therefore likely reflects behavioural and system-visibility effects rather than demographic or index artefacts. 999 Pathway For 999 calls (n = 2,226), the deprivation gradient was stable and quasi-monotonic across all WIMD variants. Case share declined steadily from most to least deprived quintiles, consistent with higher STEMI burden in deprived populations. Removing the Health domain or adjusting for age and gender produced little change, supporting interpretation that 999 capture reflects true biological and structural inequality, minimally influenced by index composition or behaviour (Figure 2). Insert Figure 2 here. 111 Pathway For 111 calls (n = 289), the distribution was curved and unstable: case share rose from 17% (most deprived) to 24% (mid-quintiles), then fell to 16% (least deprived). Age–gender standardisation smoothed but did not remove the curvature. The Minus-Health WIMD produced a more regular gradient, while the Income-only index introduced greater irregularity, indicating that economic deprivation alone does not predict 111 use. This residual curvature therefore likely reflects behavioural influences, such as symptom appraisal, health literacy, and escalation decisions - rather than demographic structure or material deprivation (Figure 3). Insert Figure 3 here. Index Reclassification To test whether the observed deprivation gradients reflected artefacts of index composition, WIMD was re-specified by removing the Health domain and by using the Income-only domain. Across all pathways, reclassification was minimal: Combined STEMI: 91% of cases remained in the same quintile when the Health domain was excluded, and 69% under the Income-only specification. 999 pathway: 90% unchanged (Minus Health), 69% unchanged (Income-only). 111 pathway: 92% unchanged (Minus Health), 68% unchanged (Income-only). (Figure 4). These results confirm that WIMD composition had negligible influence on deprivation assignment. The persistence of non-linear and pathway-specific gradients despite stable deprivation classification supports the interpretation that behavioural visibility, rather than measurement artefact, drives the observed curvature. Insert Figure 4 here. Behaviour- independent comparator (OOHCA) OOHCA, used as a behaviour-independent comparator, showed a strong, linear deprivation gradient. Cases increased from least to most deprived quintiles (SII = –0.096; RII = 0.63), demonstrating a monotonic relationship between deprivation and event capture (Figure 5). This contrast with STEMI patterns confirms that, when behavioural influence is minimal, deprivation acts in a direct and biologically plausible manner. Insert Figure 5 here. Pathway selection regression A logistic regression compared 111 versus 999 capture, including a quadratic deprivation term to test curvature. The unadjusted model (β₁ = 0.27, β₂ = –0.04) showed an inverse-U trend, with higher 111 use in mid-quintiles. After adjusting for age and sex (β₁ = 0.29, β₂ = –0.04), the pattern persisted, indicating that demographic composition did not explain the shape. Predicted probabilities showed a modest mid-quintile elevation (Figure 6), consistent with behavioural pathway selection rather than structural or demographic confounding. Insert Figure 6 here. Inequality metrics SII and RII values were consistent across WIMD variants but differed in magnitude, reflecting how index composition interacts with pathway behaviour. Removing the Health domain slightly flattened gradients, while the Income-only version modestly accentuated mid-quintile peaks. These shifts imply that apparent inequality partly reflects measurement design rather than true biological risk. Given that SII and RII assume linearity, their divergence, together with the observed curved, quasi-monotonic distributions, suggests endogeneity between index structure and behavioural capture. The Minus-Health WIMD was therefore adopted for main interpretation to minimise circularity with health-related indicators. Insert table 1 here. Table 1: Slope Index of Inequality and Relative Index of Inequality results (minus Health-only. Discussion This study examined whether deprivation gradients in EMS-attended STEMI represent true disease risk or artefacts of measurement and pathway behaviour. By comparing 999 and 111 entry routes and applying alternative versions of the Welsh Index of Multiple Deprivation (WIMD), we assessed how behavioural and analytic endogeneity shape visible inequalities in EMS data. Out-of-hospital cardiac arrest (OOHCA) provided a behaviour-independent comparator. STEMI displayed a non-linear, inverse-U relationship with deprivation which peaked in mid-quintile areas and declined among the least deprived. This persisted across WIMD variants and after demographic adjustment, indicating that neither index design nor population structure explained the pattern. By contrast, OOHCA followed a clear, monotonic gradient increasing with deprivation, consistent with biological risk when help-seeking behaviour does not intervene. A key finding was the divergence between pathways. The 999 pathway showed a stable, quasi-monotonic gradient robust to WIMD specification and standardisation, suggesting it most closely reflected true disease burden. The 111 pathway, in contrast, exhibited an unstable and flattened gradient that changed when the Health domain was excluded. This implies dual forms of endogeneity: behavioural endogeneity (who seeks help and when) and measurement endogeneity (overlap between healthcare use and deprivation metrics). These findings extend Mohammed et al. (2025), who identified domain endogeneity in chronic disease outcomes, by demonstrating behavioural and analytic endogeneity in acute event capture. Unlike prior work focusing on cross-system bias, this study identifies intra-system endogeneity – i.e. bias arising within a single EMS dataset due to heterogeneity in how patients access care. The 999 pathway reflects behavioural endogeneity alone, whereas 111 is affected by both behavioural and index-design dependencies. The divergence between OOHCA and STEMI indicates that the apparent attenuation, or even reversal, of the deprivation gradient in STEMI is not a real reduction in cardiac risk, but a visibility artefact. The most deprived populations likely experience higher true incidence but appear less often in EMS-confirmed STEMI data due to later or absent system activation, whereas mid-deprived groups combine elevated risk with greater help-seeking propensity. This explains the inverted-U pattern observed in STEMI cases, which contrasts with the consistent monotonic increase in OOHCA cases across deprivation. In OOHCA, help-seeking behaviour is minimal or absent due to the acute severity, allowing the biological gradient to manifest clearly. For STEMI, symptom recognition, health literacy, and healthcare engagement influence case visibility, introducing pathway-dependent behavioural biases that modify the observed gradient. Methodologically, our results show that common inequality metrics such as the Slope Index of Inequality (SII) and Relative Index of Inequality (RII) can misrepresent non-linear relationships. Both assume monotonicity and may flatten mid-quintile peaks, underestimating inequality magnitude. This empirically supports simulation findings by Moreno-Betancur et al. (2015) and Renard et al. (2019), highlighting the need for non-linear and graphical approaches when interpreting deprivation gradients. More broadly, the analysis reinforces that no dataset is behaviourally neutral. Visibility within EMS and other routine health systems reflects not only biological risk but also social and behavioural processes that govern access. For equity-focused planning, analysts should prioritise less circular deprivation indices (such as the Minus-Health WIMD) and interpret mid-quintile peaks as potential visibility artefacts. Interventions should target improved symptom recognition and escalation among groups most affected by pathway-mediated bias, particularly 111 users. These findings underscore the importance of careful consideration when using routine EMS data and composite deprivation indices for service monitoring and planning. By identifying how behavioural factors and deprivation index design influence the visibility of acute cardiac events, this evaluation informs more accurate interpretation of EMS data within the Welsh Ambulance Services NHS Trust. This understanding is crucial for ensuring that service evaluations and operational decisions reflect true population needs rather than artefacts of data capture or index construction, thereby supporting equitable and effective service delivery. Limitations This evaluation has several limitations. First, deprivation status was assigned using WIMD 2019, which is based on 2011 Census LSOA boundaries. As with all UK deprivation indices, this creates a temporal lag: the measure reflects historic area conditions and cannot fully capture more recent change, introducing unavoidable misalignment with contemporary community characteristics. Second, deprivation was assigned at the area level, which may obscure differences between individuals living in the same LSOA and does not capture personal socioeconomic circumstances. Finally, although this analysis was conducted within a single national EMS system, the methodological mechanisms identified here—behavioural endogeneity, pathway-mediated capture, and index-driven measurement effects—are structural features of routine EMS data and are therefore likely to operate in other settings, even if specific gradient shapes differ. In summary, inequalities in EMS STEMI data arise from both true deprivation-linked disease and behavioural capture effects shaped by index design. Behaviour-independent comparators such as OOHCA and the use of Health-excluded indices provide more valid assessments of inequality. Recognising and correcting for these behavioural and analytic feedbacks is essential to ensure that service data reflect genuine need rather than recorded contact, and that resource allocation follows risk, not visibility. Abbreviations EMS Emergency Medical Services STEMI ST-elevation Myocardial Infarction OOHCA Out-of-Hospital Cardiac Arrest WIMD Welsh Index of Multiple Deprivation SII Slope Index of Inequality RII Relative Index of Inequality ePCR Electronic Patient Care Record LSOA Lower-layer Super Output Area WAST Welsh Ambulance Service NHS Trust Declarations Acknowledgements Greg Lloyd and Julie Starling, Welsh Ambulance Service NHS Trust – for support with the project. Save a Life Cymru – for its contribution to public awareness and education in cardiac arrest, which informed contextual interpretation of the findings. All acknowledged individuals and organisations provided permission to be named. Author information Corresponding author: [email protected] Author contact: [email protected] Declarations Ethics approval and consent to participate This study was classified as a service evaluation under the governance of the Welsh Ambulance Service NHS Trust. In accordance with UK Health and Research Authority guidance, formal NHS Research Ethics Committee approval was not required. All data were fully de-identified prior to analysis. The project did not alter patient care pathways and involved analysis of existing operational data without randomisation or intervention, consistent with the definition of service evaluation rather than research designed to generate new generalisable knowledge. Consent for Publication Not applicable Availability of data and materials The datasets generated and/or analysed during the current study are not publicly available due to patient confidentiality and data governance restrictions under NHS Wales regulations but are available from the corresponding author upon reasonable request and with permission of the Welsh Ambulance Services NHS Trust. Competing Interests The authors declare that they have no competing interests. Funding No external funding was received. Authors contributions Adam Nicholls: conceived the study, designed the methodology, conducted all data analysis, interpreted the findings, and drafted and revised the manuscript in full. Luke Watkins: provided clinical oversight and contributed to clinical interpretation of the findings. All authors read and approved the final manuscript. References Adams, J., & White, M. (2006). Removing the health domain from the index of multiple deprivation 2004: Effect on measured inequalities in census measures of health. Journal of Public Health (Oxf), 28(4), 379–383. Bradford, D. R. R., Allik, M., McMahon, A. D., & Brown, D. (2023). Assessing the risk of endogeneity bias in health and mortality inequalities research using composite measures of multiple deprivation. Health & Place, 80, 102998. Hill, A. D., Johnson, S. G., Greco, L. M., O’Boyle, E. H., & Walter, S. L. (2021). Endogeneity: A review and agenda for the methodology–practice divide affecting micro and macro research. Journal of Management, 47(1), 105–143. Jordan, H., Roderick, P., & Martin, D. (2004). The index of multiple deprivation 2000 and accessibility effects on health. Journal of Epidemiology & Community Health, 58(3), 250–257. Mohammed, S., Bailey, G. A., Farr, I. W., Jones, C., Rawlings, A., Rees, S., et al. (2025). Using the Welsh Index of Multiple Deprivation in research: Estimating the effect of excluding domains on a routine health data study. BMC Public Health, 25, 1178. Moreno-Betancur, M., Latouche, A., Menvielle, G., Kunst, A., & Rey, G. (2015). Relative Index of Inequality and Slope Index of Inequality: A structured regression framework for estimation. Epidemiology. Noble, M., Wright, G., Smith, G., & Dibben, C. (2006). Measuring multiple deprivation at the small-area level. Environment and Planning A, 38(1), 169–185. Renard, F., Devleesschauwer, B., Speybroeck, N., & Deboosere, P. (2019). Monitoring health inequalities when the socio-economic composition changes: Are the slope and relative indices of inequality appropriate? Results of a simulation study. BMC Public Health. Welsh Government. (2019). Welsh Index of Multiple Deprivation (WIMD) 2019: Technical Report. Cardiff: Welsh Government. Yusuf, S., Hawken, S., Ôunpuu, S., et al. (2004). Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): Case-control study. Lancet, 364(9438), 937–952. Footnotes The project did not alter patient care pathways and involved analysis of existing operational data without randomisation or intervention, consistent with the definition of service evaluation rather than research designed to generate new generalisable knowledge. This project was classified as such under the governance of the Welsh Ambulance Services NHS Trust (WAST). Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 28 Jan, 2026 Reviewers agreed at journal 08 Jan, 2026 Reviewers invited by journal 06 Jan, 2026 Editor assigned by journal 05 Jan, 2026 Editor invited by journal 16 Dec, 2025 Submission checks completed at journal 16 Dec, 2025 First submitted to journal 16 Dec, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8307622","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":571663191,"identity":"00c3fb5f-62dc-4b39-9755-38ac244ddc06","order_by":0,"name":"Adam Nicholls","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABE0lEQVRIiWNgGAWjYFCCBBAhwWDAwNx4gKGCgYEPxGdsABI8BLUwNhxgOMPAwEakFgaIFsY2IrTwtyc/3cC4w0LenL2x4eDPeYfl2djbHzAX7qhjMDhzAKsWiTPPzG4wnpEw3NlzsOEw77bDhm08ZwyYZ545zGBwtgG7NTcSgFraJBIMbiQ2HGbcdhvIzmFg5m07wGBwHrsO+Rvp3+BaDv6cc9u+Tf75A6CWOpxaDG7kIGw5wNtwO7ENGHpALcw4HWZ45k3ZDaAyww1ngH7hOfY/uY0nx+Awb9thHkkc3pc7nr7txse2OnmD480HH/6oSbPtZz/+8DHQYXJ8ZxKwuwwEMKRAxuOMyFEwCkbBKBgFhAEA2oZn79CVdyoAAAAASUVORK5CYII=","orcid":"","institution":"Welsh Ambulance Services NHS Trust","correspondingAuthor":true,"prefix":"","firstName":"Adam","middleName":"","lastName":"Nicholls","suffix":""},{"id":571663192,"identity":"6f21bea4-464f-47f0-900c-5e5b754db359","order_by":1,"name":"Luke Watkins","email":"","orcid":"","institution":"Welsh Ambulance Services NHS Trust","correspondingAuthor":false,"prefix":"","firstName":"Luke","middleName":"","lastName":"Watkins","suffix":""}],"badges":[],"createdAt":"2025-12-08 12:23:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8307622/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8307622/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100093892,"identity":"60550309-576b-49aa-875f-dc3fde085315","added_by":"auto","created_at":"2026-01-13 01:27:43","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":508937,"visible":true,"origin":"","legend":"","description":"","filename":"20251105UnderstandingbiasinEMSSTEMIdatav04.docx","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/56fd39cc95b50fcd225d335d.docx"},{"id":100093894,"identity":"270ba336-649b-43be-ba6b-9903d30b8099","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6028,"visible":true,"origin":"","legend":"","description":"","filename":"ac11182549a34f99a9f0527ade4ba234.json","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/64332f3f320c11fc6df97179.json"},{"id":100093909,"identity":"878a624b-45dd-4816-afe8-9bc7fc83ff85","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":58208,"visible":true,"origin":"","legend":"","description":"","filename":"ac11182549a34f99a9f0527ade4ba2341enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/b0096956cbd233cf8b152224.xml"},{"id":100364837,"identity":"bdc768f2-fff1-4d9b-bd2a-1c82e4434f87","added_by":"auto","created_at":"2026-01-16 07:54:21","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":206467,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/4989391f04e0f398b1099896.png"},{"id":100093898,"identity":"633bdc04-ecfb-4884-a138-01ec7c3a5102","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":203399,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/518ccb77d7f228393fd49c41.png"},{"id":100364589,"identity":"6e1e74f8-6659-484e-b481-5fd8fab8835c","added_by":"auto","created_at":"2026-01-16 07:53:58","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":122768,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/e516c18933ba88a23e71de9e.png"},{"id":100364471,"identity":"fc21392c-f985-4dda-a955-dff3ab6f29ad","added_by":"auto","created_at":"2026-01-16 07:53:46","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9114,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/b290459b2b66abf2751fe224.png"},{"id":100093908,"identity":"043a4ef5-61e4-4321-aadc-fa86d7cbd725","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11016,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/cc5e3917de294c8ce3d54ca1.png"},{"id":100364478,"identity":"c63e9f44-be6b-43a3-ab96-af575bdc323e","added_by":"auto","created_at":"2026-01-16 07:53:47","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9475,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/dff023337e61fd3cf62d7097.png"},{"id":100365248,"identity":"b792ccf4-3430-44f1-8776-9cfcc7ad6871","added_by":"auto","created_at":"2026-01-16 07:54:54","extension":"emf","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16432,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.emf","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/a53d5c21a6ebd8a583deedcc.emf"},{"id":100093903,"identity":"e39dd6e5-be0a-4be5-b289-24f350847714","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":37451,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/c13d70843124a30296b5b78f.png"},{"id":100365705,"identity":"2d4f018e-4dcf-4bd2-ad11-228ae7346733","added_by":"auto","created_at":"2026-01-16 07:55:32","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":37364,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/f57a3d37fd662ce1c0d61e40.png"},{"id":100093905,"identity":"e0793bdd-aac3-41b9-a79f-d3d4e1052938","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18997,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/19b4be3711cd15fc85d82912.png"},{"id":100366321,"identity":"8cac28f3-4548-4870-91f3-9c466fcee15b","added_by":"auto","created_at":"2026-01-16 07:56:13","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11967,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/d7ad58a1d5d3691dbcd05d31.png"},{"id":100093915,"identity":"eafd1d54-5e5a-4d46-82ae-db30fdf1c173","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12732,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/cb18450117fcb3879108210b.png"},{"id":100364443,"identity":"2d3efe41-80cc-48fb-8452-1a25aad1e8e3","added_by":"auto","created_at":"2026-01-16 07:53:42","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10903,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/98ed5200f1b465b87e6bc3b5.png"},{"id":100093914,"identity":"c89b34a6-7ce4-45b5-91c6-3cf83c207019","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14658,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/bf820daf53e83e80825c2b22.png"},{"id":100365633,"identity":"c657af76-30a9-43b3-bf32-51d2603c45f9","added_by":"auto","created_at":"2026-01-16 07:55:26","extension":"xml","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":53297,"visible":true,"origin":"","legend":"","description":"","filename":"ac11182549a34f99a9f0527ade4ba2341structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/3135d77e5227b87102b5003a.xml"},{"id":100093912,"identity":"55d9156d-0210-4d32-a775-95c70d8838af","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"html","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64958,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/f5f0ffd4010594da49ace4f7.html"},{"id":100365251,"identity":"29ed0de0-1950-436d-8cf3-dd5c4fa24420","added_by":"auto","created_at":"2026-01-16 07:54:54","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":206467,"visible":true,"origin":"","legend":"\u003cp\u003eGradient of EMS-attended STEMI across deprivation quintiles (Combined 111 + 999) before and after age–sex standardisation using the Minus Health WIMD.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/3bfad7e370dbe65d853afb52.png"},{"id":100364736,"identity":"34442bb4-f230-42be-89bb-1e4b5e41d434","added_by":"auto","created_at":"2026-01-16 07:54:16","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":203399,"visible":true,"origin":"","legend":"\u003cp\u003eGradient of EMS-attended STEMI across deprivation quintiles (999) before and after age–sex standardisation using the Minus Health\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/9671a0fcd457ceddd1a73e5d.png"},{"id":100366247,"identity":"6d579a5d-0ecb-4635-b502-5fb117a0028d","added_by":"auto","created_at":"2026-01-16 07:56:09","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":122768,"visible":true,"origin":"","legend":"\u003cp\u003eGradient of EMS-attended STEMI across deprivation quintiles (Combined 111) before and after age–sex standardisation using the Minus Health WIMD.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/6ed01e8b85db88ca41519372.png"},{"id":100093897,"identity":"e793ce7d-3cff-444f-a85d-b5c0d2f8ce2b","added_by":"auto","created_at":"2026-01-13 01:27:44","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":9114,"visible":true,"origin":"","legend":"\u003cp\u003eQuintile Change after WIMD specification: 111 and 999 combined.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/a1dbfa1ccec71253ea61d764.png"},{"id":100364846,"identity":"74593798-8bca-4c3c-a646-7faa64c87795","added_by":"auto","created_at":"2026-01-16 07:54:23","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":11016,"visible":true,"origin":"","legend":"\u003cp\u003eMonotonic Gradient Out of Hospital Cardiac Arrest\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/6584d98539ff0cfdb3042ef8.png"},{"id":100364859,"identity":"c7ffb511-4d57-4cc1-8171-a218a5b83da5","added_by":"auto","created_at":"2026-01-16 07:54:24","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":9475,"visible":true,"origin":"","legend":"\u003cp\u003eRegression Analysis: Probability of being captured by 111 vs 999 adjusted for age and gender.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/93f59a9bb2ef264330bb4bc4.png"},{"id":100382191,"identity":"85973d66-aa89-4b06-bb06-e28aeafdbe30","added_by":"auto","created_at":"2026-01-16 10:41:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1135169,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8307622/v1/2a965324-4871-4784-95d7-745ad8ce6c95.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Understanding bias in EMS STEMI data: a national service-evaluation study of deprivation, behavioural pathways and inequality metrics in Wales","fulltext":[{"header":"Background","content":"\u003cp\u003eComposite measures of deprivation are widely used in public health research to quantify area-level disadvantage and assess its impact on health outcomes (Noble et al., 2006). In Wales, the official measure is the Welsh Index of Multiple Deprivation (WIMD) 2019, comprising eight weighted domains: Income, Employment, Education, Access to Services, Housing, Community Safety, Physical Environment, and Health (Welsh Government, 2019). Although WIMD is integral to health services research and policy planning, methodological limitations arise when it is applied to health outcomes. The Health domain incorporates indicators reflecting chronic disease burden, including GP-recorded conditions, cancer incidence, mental health diagnoses, limiting long-term illness, premature mortality, childhood obesity, and low birth weight. Because these indicators are conceptually close to many health outcomes, the risk of endogeneity bias is introduced \u0026mdash; a form of circularity where part of what defines deprivation also defines the health outcome being studied (Hill et al., 2021; Jordan et al., 2004).\u003c/p\u003e\n\u003cp\u003eThis bias has been empirically demonstrated. Mohammed et al. (2025) showed that excluding the Health domain from WIMD reclassified 17% of Welsh LSOAs and attenuated odds ratios for the association between deprivation and diabetes. Similar results have been reported in England and Scotland (Adams \u0026amp; White, 2006; Bradford et al., 2023). However, most prior work has focused on chronic conditions, leaving uncertainty over whether index design also influences the representation of acute events. This is not trivial: indices developed to study long-term morbidity are now routinely used to analyse and plan services for acute disease, despite potentially different causal mechanisms.\u003c/p\u003e\n\u003cp\u003eWhile chronic conditions reflect cumulative exposure to deprivation, acute cardiac events provide a sharper test of whether deprivation operates through biological risk, behavioural processes, or healthcare access. The INTERHEART study demonstrated that modifiable factors such as smoking, obesity, dyslipidaemia, hypertension, and diabetes account for over 90% of the global risk of myocardial infarction (Yusuf et al., 2004), and these risks are follow the socioeconomic gradient (Foster et al., 2018). ST-elevation myocardial infarction (STEMI) is therefore an acute manifestation of chronic cardiovascular pathology. Yet, its appearance in emergency medical system (EMS) data depends on socially patterned help-seeking, pathway choice, and system triage. By contrast, out-of-hospital cardiac arrest (OOHCA) is behaviourally independent once witnessed, as it almost invariably results in a 999 call. Comparing these two conditions offers a means to distinguish true deprivation-linked disease risk from artefacts of healthcare interface and pathway engagement.\u003c/p\u003e\n\u003cp\u003eBeyond these conceptual concerns, there are also methodological limitations in how inequality is routinely quantified. Indices such as the Slope Index of Inequality (SII) and Relative Index of Inequality (RII) assume a linear, monotonic relationship between deprivation and health. Simulation studies (Renard et al., 2019; Moreno-Betancur et al., 2015) show that these measures can produce paradoxical or misleading results when the true association is non-linear \u0026mdash; an issue that has received little empirical investigation in real-world health data.\u003c/p\u003e\n\u003cp\u003eThis evaluation addresses both gaps. Using national Welsh ambulance data, we examine how socioeconomic deprivation shapes the capture of acute cardiac events within the EMS. We analyse deprivation gradients in STEMI and OOHCA across the 999 and 111-transfer pathways, evaluating three variants of WIMD: the original version (with Health), a Health-domain\u0026ndash;excluded version (\u0026ldquo;Minus Health\u0026rdquo;), and an Income-only version. By contrasting a behaviour-dependent condition (STEMI) with a behaviour-independent comparator (OOHCA), and by integrating linear and quadratic inequality models, we test whether observed inequalities reflect true disease burden, measurement artefact, or behavioural endogeneity within EMS data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eObjective:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo examine how socioeconomic deprivation influences the representation of acute cardiac events within the EMS, and whether observed inequalities reflect true disease burden or artefacts of measurement or behaviour. Using national Welsh ambulance data, deprivation gradients in STEMI and OOHCA were compared across 999 and 111 pathways, testing sensitivity to alternative WIMD specifications.\u003c/p\u003e\n\u003cp\u003eFindings are intended to inform interpretation of routinely collected EMS data and improve understanding of how deprivation measures may influence statistical representations of acute health events, particularly in the context of EMS utilisation.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eStudy design\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis project\u003ca href=\"#_ftn1\" name=\"_ftnref1\" title=\"\"\u003e\u003c/a\u003e\u003csup\u003e3\u003c/sup\u003e used routinely collected individual level ePCR data from the WAST between September 2021 and June 2025. The primary aim was to understand how socioeconomic deprivation, as measured by different variants of the WIMD, influences the capture and characterisation of acute cardiac events within the EMS.\u003c/p\u003e\n\u003cp\u003eBy linking ePCR records to area-level deprivation indices, this evaluation sought to assess whether the inclusion of health-related domains within composite deprivation measures affects the observed distribution and inequalities in EMS data. Analyses were repeated using the full WIMD, a Health-domain\u0026ndash;excluded version (\u0026ldquo;Minus Health\u0026rdquo;), and an Income-only version to evaluate the sensitivity of results to methodological choices in deprivation measurement.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis analysis was classified as a service evaluation under the governance of WAST. In line with Health Research Authority (HRA) guidance, NHS Research Ethics Committee (REC) approval was not required. In line with this, the need for ethical approval was waived according to these regulations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCohorts\u003c/strong\u003e\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003eSTEMI Cohort\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eThe STEMI cohort included all ePCR records in which the paramedic closure code indicated ST-elevation myocardial infarction (STEMI), confirmed by an electrocardiogram. All clear comorbid diseases that may have preceded the STEMI were controlled for.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCases were analysed by pathway of entry:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e999 pathway \u0026ndash; direct emergency activation via a 999 call.\u003c/li\u003e\n \u003cli\u003e111 pathway \u0026ndash; calls initiated through NHS 111 before ambulance dispatch.\u0026nbsp;\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThis allowed comparison of deprivation gradients across different access routes into the same emergency system.\u0026nbsp;\u003c/p\u003e\n\u003col start=\"2\"\u003e\n \u003cli\u003eOut-of-Hospital-Cardiac Arrest (OOHCA) Cohort\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eThe OOHCA cohort included all ePCR records with a paramedic closure code of cardiac arrest, accompanied by a Cardiac Arrest Form. OOHCA represents a behaviour-independent comparator, as nearly all witnessed arrests generate a 999 response, providing a benchmark less influenced by help-seeking behaviour or pathway selection.\u003c/p\u003e\n\u003cp\u003eFor both cohorts, patients who died at scene were included if a closure diagnosis was recorded. Recognised Life Extinct (ROLE) cases without closure codes were excluded. Exclusion criteria can be found in Supplementary Table 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAnalytic Approach\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDeprivation gradients were quantified using two regression-based inequality metrics:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eSlope Index of Inequality (SII): absolute difference across deprivation quintiles, estimated by weighted linear regression of outcome prevalence on ridit scores.\u003c/li\u003e\n \u003cli\u003eRelative Index of Inequality (RII): relative difference, estimated by weighted logistic regression using the same ridit structure.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eAnalyses were conducted separately for:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e999, 111, and combined STEMI datasets, and\u003c/li\u003e\n \u003cli\u003eOOHCA as a contrast condition.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eQuadratic terms were included to test for non-linear relationships between deprivation and case share.\u003c/p\u003e\n\u003cp\u003eA logistic regression comparing 111 vs 999 capture assessed whether pathway selection varied with deprivation after adjusting for age and sex.\u003c/p\u003e\n\u003cp\u003eThis design enabled differentiation between deprivation effects driven by true disease burden and those arising from pathway-mediated capture within the emergency system.\u0026nbsp;\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eStudy Population\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 12,241 EMS-attended cases were analysed: 2,515 STEMI and 9,726 out-of-hospital cardiac arrests (OOHCA).\u003c/p\u003e\n\u003cp\u003eAcross both conditions, two-thirds were male, and over three-quarters were aged \u0026ge;60 years.\u003c/p\u003e\n\u003cp\u003eSTEMI included a higher proportion of adults aged 40\u0026ndash;59 (18% vs 12% in OOHCA). Among STEMI cases, 89% entered via the 999 pathway and 11% via 111 transfers. All had confirmed ambulance attendance.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCombined STEMI analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWhen 999 and 111 cases were combined, STEMI displayed a non-linear, inverse-U gradient across deprivation quintiles. Case share rose from 19% in the most deprived to 22% in the mid-quintiles, then declined to 16% in the least deprived (Figure 1).\u003c/p\u003e\n\u003cp\u003eInsert Figure 1 here.\u003c/p\u003e\n\u003cp\u003eThis curvature persisted after age\u0026ndash;sex standardisation, indicating that demographic structure did not account for the shape. Removing the WIMD Health domain accentuated the mid-quintile peak, while the Income-only version slightly weakened it. The similarity between the standardised and Income-only curves suggests that the Health domain was masking true deprivation differences, whereas demographic effects were minor. The remaining curvature therefore likely reflects behavioural and system-visibility effects rather than demographic or index artefacts.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e999 Pathway\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor 999 calls (n = 2,226), the deprivation gradient was stable and quasi-monotonic across all WIMD variants. Case share declined steadily from most to least deprived quintiles, consistent with higher STEMI burden in deprived populations. Removing the Health domain or adjusting for age and gender produced little change, supporting interpretation that 999 capture reflects true biological and structural inequality, minimally influenced by index composition or behaviour (Figure 2).\u003c/p\u003e\n\u003cp\u003eInsert Figure 2 here.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e111 Pathway\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor 111 calls (n = 289), the distribution was curved and unstable: case share rose from 17% (most deprived) to 24% (mid-quintiles), then fell to 16% (least deprived). Age\u0026ndash;gender standardisation smoothed but did not remove the curvature. The Minus-Health WIMD produced a more regular gradient, while the Income-only index introduced greater irregularity, indicating that economic deprivation alone does not predict 111 use.\u003c/p\u003e\n\u003cp\u003eThis residual curvature therefore likely reflects behavioural influences, such as symptom appraisal, health literacy, and escalation decisions - rather than demographic structure or material deprivation (Figure 3).\u003c/p\u003e\n\u003cp\u003eInsert Figure 3 here.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIndex Reclassification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo test whether the observed deprivation gradients reflected artefacts of index composition, WIMD was re-specified by removing the Health domain and by using the Income-only domain. Across all pathways, reclassification was minimal:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eCombined STEMI: 91% of cases remained in the same quintile when the Health domain was excluded, and 69% under the Income-only specification.\u003c/li\u003e\n \u003cli\u003e999 pathway: 90% unchanged (Minus Health), 69% unchanged (Income-only).\u003c/li\u003e\n \u003cli\u003e111 pathway: 92% unchanged (Minus Health), 68% unchanged (Income-only).\u003cbr\u003e\u0026nbsp;(Figure 4).\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThese results confirm that WIMD composition had negligible influence on deprivation assignment. The persistence of non-linear and pathway-specific gradients despite stable deprivation classification supports the interpretation that behavioural visibility, rather than measurement artefact, drives the observed curvature.\u003c/p\u003e\n\u003cp\u003eInsert Figure 4 here.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBehaviour- independent comparator (OOHCA)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOOHCA, used as a behaviour-independent comparator, showed a strong, linear deprivation gradient. Cases increased from least to most deprived quintiles (SII = \u0026ndash;0.096; RII = 0.63), demonstrating a monotonic relationship between deprivation and event capture (Figure 5). This contrast with STEMI patterns confirms that, when behavioural influence is minimal, deprivation acts in a direct and biologically plausible manner.\u003c/p\u003e\n\u003cp\u003eInsert Figure 5 here.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePathway selection regression\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA logistic regression compared 111 versus 999 capture, including a quadratic deprivation term to test curvature.\u003c/p\u003e\n\u003cp\u003eThe unadjusted model (\u0026beta;₁\u0026nbsp;= 0.27, \u0026beta;₂\u0026nbsp;= \u0026ndash;0.04) showed an inverse-U trend, with higher 111 use in mid-quintiles. After adjusting for age and sex (\u0026beta;₁\u0026nbsp;= 0.29, \u0026beta;₂\u0026nbsp;= \u0026ndash;0.04), the pattern persisted, indicating that demographic composition did not explain the shape.\u003c/p\u003e\n\u003cp\u003ePredicted probabilities showed a modest mid-quintile elevation (Figure 6), consistent with behavioural pathway selection rather than structural or demographic confounding.\u003c/p\u003e\n\u003cp\u003eInsert Figure 6 here.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInequality metrics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSII and RII values were consistent across WIMD variants but differed in magnitude, reflecting how index composition interacts with pathway behaviour. Removing the Health domain slightly flattened gradients, while the Income-only version modestly accentuated mid-quintile peaks. These shifts imply that apparent inequality partly reflects measurement design rather than true biological risk.\u003c/p\u003e\n\u003cp\u003eGiven that SII and RII assume linearity, their divergence, together with the observed curved, quasi-monotonic distributions, suggests endogeneity between index structure and behavioural capture.\u003c/p\u003e\n\u003cp\u003eThe Minus-Health WIMD was therefore adopted for main interpretation to minimise circularity with health-related indicators.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Insert table 1 here.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 1: Slope Index of Inequality and Relative Index of Inequality results (minus Health-only.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"https://myfiles.space/user_files/58895_8739fc6c57c1c19a/58895_custom_files/img1768233331.png\" width=\"839\" height=\"217\"\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study examined whether deprivation gradients in EMS-attended STEMI represent true disease risk or artefacts of measurement and pathway behaviour. By comparing 999 and 111 entry routes and applying alternative versions of the Welsh Index of Multiple Deprivation (WIMD), we assessed how behavioural and analytic endogeneity shape visible inequalities in EMS data. Out-of-hospital cardiac arrest (OOHCA) provided a behaviour-independent comparator.\u003c/p\u003e\n\u003cp\u003eSTEMI displayed a non-linear, inverse-U relationship with deprivation which peaked in mid-quintile areas and declined among the least deprived. This persisted across WIMD variants and after demographic adjustment, indicating that neither index design nor population structure explained the pattern. By contrast, OOHCA followed a clear, monotonic gradient increasing with deprivation, consistent with biological risk when help-seeking behaviour does not intervene.\u003c/p\u003e\n\u003cp\u003eA key finding was the divergence between pathways. The 999 pathway showed a stable, quasi-monotonic gradient robust to WIMD specification and standardisation, suggesting it most closely reflected true disease burden. The 111 pathway, in contrast, exhibited an unstable and flattened gradient that changed when the Health domain was excluded. This implies dual forms of endogeneity: behavioural endogeneity (who seeks help and when) and measurement endogeneity (overlap between healthcare use and deprivation metrics).\u003c/p\u003e\n\u003cp\u003eThese findings extend Mohammed et al. (2025), who identified domain endogeneity in chronic disease outcomes, by demonstrating behavioural and analytic endogeneity in acute event capture. Unlike prior work focusing on cross-system bias, this study identifies intra-system endogeneity \u0026ndash; i.e. bias arising within a single EMS dataset due to heterogeneity in how patients access care. The 999 pathway reflects behavioural endogeneity alone, whereas 111 is affected by both behavioural and index-design dependencies.\u003c/p\u003e\n\u003cp\u003eThe divergence between OOHCA and STEMI indicates that the apparent attenuation, or even reversal, of the deprivation gradient in STEMI is not a real reduction in cardiac risk, but a visibility artefact. The most deprived populations likely experience higher true incidence but appear less often in EMS-confirmed STEMI data due to later or absent system activation, whereas mid-deprived groups combine elevated risk with greater help-seeking propensity. This explains the inverted-U pattern observed in STEMI cases, which contrasts with the consistent monotonic increase in OOHCA cases across deprivation. In OOHCA, help-seeking behaviour is minimal or absent due to the acute severity, allowing the biological gradient to manifest clearly. For STEMI, symptom recognition, health literacy, and healthcare engagement influence case visibility, introducing pathway-dependent behavioural biases that modify the observed gradient.\u003c/p\u003e\n\u003cp\u003eMethodologically, our results show that common inequality metrics such as the Slope Index of Inequality (SII) and Relative Index of Inequality (RII) can misrepresent non-linear relationships. Both assume monotonicity and may flatten mid-quintile peaks, underestimating inequality magnitude. This empirically supports simulation findings by Moreno-Betancur et al. (2015) and Renard et al. (2019), highlighting the need for non-linear and graphical approaches when interpreting deprivation gradients.\u003c/p\u003e\n\u003cp\u003eMore broadly, the analysis reinforces that no dataset is behaviourally neutral. Visibility within EMS and other routine health systems reflects not only biological risk but also social and behavioural processes that govern access. For equity-focused planning, analysts should prioritise less circular deprivation indices (such as the Minus-Health WIMD) and interpret mid-quintile peaks as potential visibility artefacts. Interventions should target improved symptom recognition and escalation among groups most affected by pathway-mediated bias, particularly 111 users.\u003c/p\u003e\n\u003cp\u003eThese findings underscore the importance of careful consideration when using routine EMS data and composite deprivation indices for service monitoring and planning. By identifying how behavioural factors and deprivation index design influence the visibility of acute cardiac events, this evaluation informs more accurate interpretation of EMS data within the Welsh Ambulance Services NHS Trust. This understanding is crucial for ensuring that service evaluations and operational decisions reflect true population needs rather than artefacts of data capture or index construction, thereby supporting equitable and effective service delivery.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLimitations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis evaluation has several limitations. First, deprivation status was assigned using WIMD 2019, which is based on 2011 Census LSOA boundaries. As with all UK deprivation indices, this creates a temporal lag: the measure reflects historic area conditions and cannot fully capture more recent change, introducing unavoidable misalignment with contemporary community characteristics. Second, deprivation was assigned at the area level, which may obscure differences between individuals living in the same LSOA and does not capture personal socioeconomic circumstances. Finally, although this analysis was conducted within a single national EMS system, the methodological mechanisms identified here\u0026mdash;behavioural endogeneity, pathway-mediated capture, and index-driven measurement effects\u0026mdash;are structural features of routine EMS data and are therefore likely to operate in other settings, even if specific gradient shapes differ.\u003c/p\u003e\n\u003cp\u003eIn summary, inequalities in EMS STEMI data arise from both true deprivation-linked disease and behavioural capture effects shaped by index design. Behaviour-independent comparators such as OOHCA and the use of Health-excluded indices provide more valid assessments of inequality. Recognising and correcting for these behavioural and analytic feedbacks is essential to ensure that service data reflect genuine need rather than recorded contact, and that resource allocation follows risk, not visibility.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eEMS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eEmergency Medical Services\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eSTEMI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eST-elevation Myocardial Infarction\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eOOHCA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eOut-of-Hospital Cardiac Arrest\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eWIMD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eWelsh Index of Multiple Deprivation\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eSII\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eSlope Index of Inequality\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eRII\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eRelative Index of Inequality\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eePCR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eElectronic Patient Care Record\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eLSOA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eLower-layer Super Output Area\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 22.0588%;\"\u003e\n \u003cp\u003eWAST\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77.9412%;\"\u003e\n \u003cp\u003eWelsh Ambulance Service NHS Trust\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGreg Lloyd and Julie Starling, Welsh Ambulance Service NHS Trust \u0026ndash; for support with the project.\u003c/p\u003e\n\u003cp\u003eSave a Life Cymru \u0026ndash; for its contribution to public awareness and education in cardiac arrest, which informed contextual interpretation of the findings.\u003c/p\u003e\n\u003cp\u003eAll acknowledged individuals and organisations provided permission to be named.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCorresponding author: [email protected]\u003c/p\u003e\n\u003cp\u003eAuthor contact: [email protected]\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was classified as a service evaluation under the governance of the Welsh Ambulance Service NHS Trust. In accordance with UK Health and Research Authority guidance, formal NHS Research Ethics Committee approval was not required. All data were fully de-identified prior to analysis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe project did not alter patient care pathways and involved analysis of existing operational data without randomisation or intervention, consistent with the definition of service evaluation rather than research designed to generate new generalisable knowledge.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for Publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and/or analysed during the current study are not publicly available due to patient confidentiality and data governance restrictions under NHS Wales regulations but are available from the corresponding author upon reasonable request and with permission of the Welsh Ambulance Services NHS Trust.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo external funding was received.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAdam Nicholls: conceived the study, designed the methodology, conducted all data analysis, interpreted the findings, and drafted and revised the manuscript in full. Luke Watkins: provided clinical oversight and contributed to clinical interpretation of the findings. All authors read and approved the final manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAdams, J., \u0026amp; White, M. (2006). Removing the health domain from the index of multiple deprivation 2004: Effect on measured inequalities in census measures of health. Journal of Public Health (Oxf), 28(4), 379\u0026ndash;383.\u003c/li\u003e\n \u003cli\u003eBradford, D. R. R., Allik, M., McMahon, A. D., \u0026amp; Brown, D. (2023). Assessing the risk of endogeneity bias in health and mortality inequalities research using composite measures of multiple deprivation. Health \u0026amp; Place, 80, 102998.\u003c/li\u003e\n \u003cli\u003eHill, A. D., Johnson, S. G., Greco, L. M., O\u0026rsquo;Boyle, E. H., \u0026amp; Walter, S. L. (2021). Endogeneity: A review and agenda for the methodology\u0026ndash;practice divide affecting micro and macro research. Journal of Management, 47(1), 105\u0026ndash;143.\u003c/li\u003e\n \u003cli\u003eJordan, H., Roderick, P., \u0026amp; Martin, D. (2004). The index of multiple deprivation 2000 and accessibility effects on health. Journal of Epidemiology \u0026amp; Community Health, 58(3), 250\u0026ndash;257.\u003c/li\u003e\n \u003cli\u003eMohammed, S., Bailey, G. A., Farr, I. W., Jones, C., Rawlings, A., Rees, S., et al. (2025). Using the Welsh Index of Multiple Deprivation in research: Estimating the effect of excluding domains on a routine health data study. BMC Public Health, 25, 1178.\u003c/li\u003e\n \u003cli\u003eMoreno-Betancur, M., Latouche, A., Menvielle, G., Kunst, A., \u0026amp; Rey, G. (2015). Relative Index of Inequality and Slope Index of Inequality: A structured regression framework for estimation. Epidemiology.\u003c/li\u003e\n \u003cli\u003eNoble, M., Wright, G., Smith, G., \u0026amp; Dibben, C. (2006). Measuring multiple deprivation at the small-area level. Environment and Planning A, 38(1), 169\u0026ndash;185.\u003c/li\u003e\n \u003cli\u003eRenard, F., Devleesschauwer, B., Speybroeck, N., \u0026amp; Deboosere, P. (2019). Monitoring health inequalities when the socio-economic composition changes: Are the slope and relative indices of inequality appropriate? Results of a simulation study. BMC Public Health.\u003c/li\u003e\n \u003cli\u003eWelsh Government. (2019). Welsh Index of Multiple Deprivation (WIMD) 2019: Technical Report. Cardiff: Welsh Government.\u003c/li\u003e\n \u003cli\u003eYusuf, S., Hawken, S., \u0026Ocirc;unpuu, S., et al. (2004). Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): Case-control study. Lancet, 364(9438), 937\u0026ndash;952.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Footnotes","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e The project did not alter patient care pathways and involved analysis of existing operational data without randomisation or intervention, consistent with the definition of service evaluation rather than research designed to generate new generalisable knowledge. This project was classified as such under the governance of the Welsh Ambulance Services NHS Trust (WAST).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-public-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"pubh","sideBox":"Learn more about [BMC Public Health](http://bmcpublichealth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/pubh/default.aspx","title":"BMC Public Health","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Emergency medical services, Socioeconomic deprivation, ST-elevation myocardial infarction, Health inequalities, Behavioural endogeneity, Welsh Index of Multiple Deprivation, Out-of-hospital cardiac arrest","lastPublishedDoi":"10.21203/rs.3.rs-8307622/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8307622/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e \u003cb\u003eBackground\u003c/b\u003e \u003c/p\u003e \u003cp\u003eComposite deprivation indices are widely used in health services research but may introduce endogeneity when health-related indicators are included in the deprivation measure itself. This study examined whether socioeconomic gradients in emergency cardiac events reflect true disease burden or are distorted by index composition and behavioural pathway selection.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMethods\u003c/b\u003e \u003c/p\u003e \u003cp\u003eWe conducted a national evaluation using linked electronic patient care records from the Welsh Ambulance Services NHS Trust (2021\u0026ndash;2025). Deprivation gradients in ST-elevation myocardial infarction (STEMI) and out-of-hospital cardiac arrest (OOHCA) were analysed across 999 and 111 pathways using three WIMD variants: full, Health-domain\u0026ndash;excluded (\u0026ldquo;Minus Health\u0026rdquo;), and Income-only. Inequality was quantified using Slope and Relative Indices of Inequality (SII, RII), with quadratic terms testing for non-linearity.\u003c/p\u003e \u003cp\u003e \u003cb\u003eResults\u003c/b\u003e \u003c/p\u003e \u003cp\u003eAmong 12,241 EMS-attended cases (2,515 STEMI; 9,726 OOHCA), STEMI showed an inverse-U deprivation gradient that peaked in mid-quintiles, particularly for 111 users, while OOHCA displayed a strong, linear increase with deprivation. Removing the Health domain altered STEMI gradients but had minimal impact on OOHCA, suggesting measurement bias in the former. The 999 STEMI pathway showed a more consistent, monotonic gradient aligned with biological risk, whereas the 111 pathway was less stable and more affected by behavioural factors such as help-seeking and symptom appraisal.\u003c/p\u003e \u003cp\u003e \u003cb\u003eConclusions\u003c/b\u003e \u003c/p\u003e \u003cp\u003eInequalities in EMS STEMI data reflect both true deprivation-linked disease and behavioural capture effects shaped by index design. Behaviour-independent comparators like OOHCA and the use of Health-excluded indices offer more valid assessments of inequality. Non-linear modelling is essential, as standard linear metrics (SII, RII) may obscure complex relationships. Crucially, these findings suggest that no dataset is behaviourally neutral: even within a single EMS system, pathway-specific behaviours influence visibility and gradient shape. Accurate interpretation of service-based health data must therefore account for both measurement structure and access dynamics to avoid misrepresenting need and misallocating resources.\u003c/p\u003e","manuscriptTitle":"Understanding bias in EMS STEMI data: a national service-evaluation study of deprivation, behavioural pathways and inequality metrics in Wales","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-13 01:27:39","doi":"10.21203/rs.3.rs-8307622/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-01-28T14:21:26+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"154974345597493882073140264475013891734","date":"2026-01-08T13:13:15+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-01-06T12:51:43+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-05T11:45:36+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-12-17T04:44:25+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-12-16T09:14:53+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Public Health","date":"2025-12-16T09:02:08+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-public-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"pubh","sideBox":"Learn more about [BMC Public Health](http://bmcpublichealth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/pubh/default.aspx","title":"BMC Public Health","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"14020efc-5f23-4138-8aec-d0fe7dfce5d0","owner":[],"postedDate":"January 13th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-01-13T01:27:39+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-13 01:27:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8307622","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8307622","identity":"rs-8307622","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}