Elsevier

Resuscitation

Volume 54, Issue 2, 1 August 2002, Pages 125-131
Resuscitation

The identification of risk factors for cardiac arrest and formulation of activation criteria to alert a medical emergency team

https://doi.org/10.1016/S0300-9572(02)00100-4Get rights and content

Abstract

Aim: (1) To identify risk factors for in-hospital cardiac arrest; (2) to formulate activation criteria to alert a clinical response culminating in attendance by a Medical Emergency Team (MET); (3) to evaluate the sensitivity and specificity of the scoring system. Methods: Quasi-experimental design to determine prevalence of risk factors for cardiac arrest in the hospitalised population. Weighting of risk factors and formulation of activation criteria to alert a graded clinical response. ROC analysis of weighted cumulative scores to determine their sensitivity and specificity. Setting: An acute 700 bed district general hospital with 32 348 adult admissions in 1999 and a catchment population of around 365 000. Subjects: 118 consecutive adult patients suffering primary cardiac arrest in-hospital and 132 non-arrest patients, randomly selected according to stratified randomisation by gender and age. Results: Risk factors for cardiac arrest include: abnormal respiratory rate (P=0.013), abnormal breathing indicator (abnormal rate or documented shortness of breath) (P<0.001), abnormal pulse (P<0.001), reduced systolic blood pressure (P<0.001), abnormal temperature (P<0.001), reduced pulse oximetry (P<0.001), chest pain (P<0.001) and nurse or doctor concern (P<0.001). Multivariate analysis of cardiac arrest cases identified three positive associations for cardiac arrest: abnormal breathing indicator (OR 3.49; 95% CI: 1.69–7.21), abnormal pulse (OR 4.07; 95% CI: 2.0–8.31) and abnormal systolic blood pressure (OR 19.92; 95% CI: 9.48–41.84). Risk factors were weighted and tabulated. The aggregate score determines the grade of clinical response. ROC analysis determined that a score of 4 has 89% sensitivity and 77% specificity for cardiac arrest; a score of 8 has 52% sensitivity and 99% specificity. All patients scoring greater than 10 suffered cardiac arrest. Conclusion: Risk factors for cardiac arrest have been identified, quantified and formulated into a table of activation criteria to help predict and avert cardiac arrest by alerting a clinical response. A graded clinical response has resulted in a tool that has both sensitivity and specificity for cardiac arrest.

Sumàrio

Objectivo: (1) Identificar os factores de risco para a paragem cardı́aca intra-hospitalar; (2) formular critérios de activação duma Equipa de Emergência Médica (MET); (3) Avaliar a sensibilidade e especificidade do sistema de score. Métodos: Desenho “quasi-experimental” para determinar a prevalência de factores de risco para paragem cardı́aca (PCR) na população hospitalizada. Avaliar a importância de factores de risco e a formulação de critérios de activação clı́nica escalonada. Análise ROC de scores cumulativos para determinar a sua sensibilidade e especificidade. Contexto: Hospital distrital com 700 camas de agudos com 32348 admissões de adultos em 1999 e servindo uma população de cerca de 365000 indivı́duos. População: 118 adultos que sofreram PCR primária intra-hospitalar e 132 indivı́duos que não pararam, seleccionados de forma randomizada de acordo com randomização estratificada por sexo e idade. Resultados: Os factores de risco para paragem cardı́aca incluem: frequência respiratória anormal (P=0.013), indicador de respiração anormal (frequência anormal ou documentado encurtamento da respiração) (P<0.001), pulso anormal (P<0.001), redução da pressão arterial sistólica (P<0.001), temperatura anormal (P<0.001), redução de valores na oximetria de pulso (P<0.001), dor torácica (P< 0.001) e preocupação por parte da enfermagem ou do médico (P<0.001). Uma análise multivariante das PCR, identificou três associações positivas para PCR: indicador de respiração anormal (OR 3.49; CI: 1.69–7.21), pulso anormal (OR 4.07; 95% CI: 2.0–8.31) e pressão arterial sistólica anormal (OR 19.92; 95% CI: 9.48–41.84). Os factores de risco foram ponderados e tabulados. Os scores agregados determinam o grau de resposta clı́nica. Análise de ROC determina que um score de 4 tem 89% de sensibilidade e 77% de especificidade para a paragem cardı́aca; um score de 8 tem uma sensibilidade de 52% e especificidade de 99%. Todos os doentes com scores superiores a 10 sofreram PCR. Conclusão: Foram identificados factores de risco para paragem cardı́aca, quantificados e formulados numa tabela de critérios de activação para ajudar a prever e evitar a PCR através da activação precoce do MET. Uma resposta clı́nica escalonada constitui um intrumento com elevada sensibilidade e especı́ficidade para a identificação de PCR.

Resumen

Objetivos: (1) Identificar factores de riesgo de paro cardı́aco intrahospitalario; (2) formular criterios de activación para alertar una respuesta clı́nica que culmine en la atención por el Equipo de Emergencia Médica (MET); (3) evaluar la sensibilidad y especificidad del sistema de puntaje. Métodos: diseño cuasi experimental para determinar la prevalencia de factores de riesgo de paro cardı́aco en la población hospitalizada. Sopesar los factores de riesgo y formulación de criterios de activación para alertar una respuesta clı́nica graduada. El análisis ROC de puntajes acumulativos evaluados para determinar su sensibilidad y especificidad. Ambiente: Un hospital general distrital de 700 camas de agudos con 32348 admisiones de adultos en 1999 y una población asignada de alrededor de 365000 personas. Sujetos:118 pacientes adultos consecutivos que sufren paro cardiorrespiratorio primario intrahospitalario y 132 pacientes que no presentan paro cardı́aco, seleccionados al azar de acuerdo a la estratificación randomizada por edad y sexo. Resultados: Los factores de riesgo de paro cardı́aco incluyen: frecuencia respiratoria anormal (P=0.013), indicador de respiración anormal (frecuencia anormal o dificultad respiratoria) (P<0.001), pulso anormal (P<0.001), presión sistólica reducida (P<0.001), temperatura anormal (P<0.001), oximetrı́a de pulso reducida (P<0.001), dolor de pecho (P<0.001), e inquietud del doctor o de la enfermera al respecto de ese paciente (P<0.001). El análisis de las múltiples variables de paro cardı́aco identificó tres asociaciones positivas con paro cardı́aco: indicador de respiración anormal (OR 3.49; 95% CI: 1.69–7.21), pulso anormal (OR 4.07; 95% CI: 2.0–8.31) y presión sistólica anormal (OR 19.92; 95% CI: 9.48–41.84). Los factores de riesgo fueron evaluados y tabulados. El puntaje total determina el grado de respuesta clı́nica. El análisis ROC determinó que el puntaje de 4 tiene 89% de sensibilidad y 77% de especificidad para paro cardı́aco; un puntaje de 8 tiene 52% de sensibilidad y 99% de especificidad. Todos los pacientes con puntaje mayor que 10 sufrieron un paro cardiorespiratorio. Conclusión: Se han identificado, cuantificado los factores de riesgo, y han sido formulados en una tabla de criterios de activación para ayudar a predecir y prevenir un paro cardı́aco a través de alertar una respuesta clı́nica. Una respuesta clı́nica graduada ha resultado ser una herramienta que tiene tanto sensibilidad como especificidad para paro cardı́aco.

Introduction

Two-thirds of in-hospital cardiac arrests are potentially avoidable [1]. Poor survival [2] and international recognition that cardiac arrest is frequently avoidable [3], [4], [5], [6] has resulted in the development of acute response teams that aim to avert cardiac arrest [7], [8], [9], [10]. These teams predominantly rely upon predetermined changes in physiological parameters to trigger a clinical response. United Kingdom central government supports the extension of acute response teams as part of its prevention of illness strategy [11], [12], [13].

The root scoring system for the MET from Australia was developed empirically, and has been subject to further empirical modifications in Australia [7], [8] and the UK [9], [10]. The development of these systems has been related to the prevalence of individual abnormal signs without knowing their predictive value for cardiac arrest [14], [15]. In the simpler systems, where no methodology is described, they are presumed to be entirely experience based. Failure to validate existing MET scoring systems has been identified as a weakness [14], [16].

Section snippets

Setting

An acute 700 bed district general hospital with 32 348 adult admissions in 1999 and a catchment population of around 365 000.

Definitions

Adult patients were defined as>16 years of age.

Patients and case-note analysis

About 118 consecutive adult patients suffering primary cardiac arrest where resuscitation was attempted and a comparative sample formed from the chart records of 132 non-arrest adult patients present on acute wards on the day of investigation. Chart records of controls were chosen randomly by stratified randomisation

Results

The two groups were of comparable age (cardiac arrest patients median age 72.7 years, range 22–93 years; non-arrest patients median age 69.2 years, range 18–98 years) and gender distribution (proportions of females 46.6% for cases and 49.2% for controls).

Discussion

This study represents the first evidence-based approach to providing activation criteria for an acute response team rather than relying upon extrapolations from empirical models. The clinical antecedents to fatal deterioration identified in this study are consistent with the findings of other studies. The key difference is the quantification of these antecedents into a structured and weighted response. Assessment of sensitivity and specificity confirms the need for a graded clinical response

Conclusion

This study has quantified the importance of new symptoms, physiological changes and biochemical changes as risk factors for in-hospital cardiac arrest. A graduated clinical response begins to address the concerns regarding sensitivity and specificity of scoring systems. Activation criteria that quantify deterioration and provide a graduated and appropriate clinical response are a risk management tool that will assist in managing system failures within our hospitals.

Acknowledgements

Colonel N Ineson and Dr L Shaikh are thanked for their review of cardiac arrest cases and activation criteria. This research project is funded for 2 years from November 1999 by the Defence Secondary Care Agency's Clinical Effectiveness budget.

References (29)

  • A. Smith et al.

    Can some in-hospital cardio-respiratory arrests be prevented

    Resuscitation

    (1998)
  • R. Schein et al.

    Clinical antecedents to in-hospital cardiopulmonary arrest

    Chest

    (1990)
  • A.L. George et al.

    Pre-arrest morbidity and other correlates of survival after in-hospital cardiopulmonary arrest

    Am. J. Med.

    (1989)
  • Hodgetts TJ, Kenward G, Vlachonikolis I, Payne S, Castle N, Crouch R, Ineson N, Shaikh L. Incidence, location and...
  • C.L. Gwinnutt et al.

    Outcome after cardiac arrest in adults in UK hospitals: effect of the guidelines

    Resuscitation

    (1997)
  • F.L. Sax et al.

    Medical patients at high risk for catastrophic deterioration

    Crit. Care Med.

    (1987)
  • C. Franklin et al.

    Developing strategies to prevent in-hospital cardiac arrest: analyzing responses of physicians and nurses in the hours before the event

    Crit. Care Med.

    (1994)
  • H.B. Rubins et al.

    Discharge decision-making in a medical ICU: characteristics of unexpected readmission

    Crit. Care Med.

    (1983)
  • F. Hourihan et al.

    The medical emergency team: a new strategy to identify and intervene in high-risk patients

    Clin. Int. Care

    (1995)
  • F.F. Daly et al.

    The medical emergency team (MET): a model for the district general hospital

    Aust. New Zealand J. Med.

    (1998)
  • R. Morgan et al.

    Early warning scoring system for detecting developing critical illness

    Clin. Int. Care

    (1997)
  • D.R. Goldhill et al.

    The patient-at-risk team: identifying and managing seriously ill ward patients

    Anaesthesia

    (1999)
  • Department of Health. An organisation with a memory: Report of an expert working group on learning from adverse events...
  • Department of Health. Comprehensive Critical Care – A review of adult critical care services. London 2000, The...
  • Cited by (263)

    View all citing articles on Scopus
    View full text