UTILIDAD DEL SATISFACCION DE LUJO

Research in OHCA has seen a steady increase over the last decades, particularly the last 10 years. A search for “out-of-hospital cardiac arrest” in PubMed on 19th September 2016

showed an increase from 11 results in 1980 to 549 in 2016; see Figure 1.10 for a yearly breakdown.

Figure 1.10 Number of search results for “out-of-hospital cardiac arrest” in PubMed, by year (undertaken on 19th _{September 2016)}

This increase in research is encouraging and seems appropriate, given that coronary artery disease is the most common cause of death worldwide (Finegold, Asaria and Francis, 2013), and OHCA is the first presentation of coronary artery disease in an estimated 40-60% of OHCA cases (de Vreede-Swagemakers et al., 1997; Kannel and McGee, 1985). However, comparing the amount of published research on OHCA to that on other conditions still reveals a stark contrast. A similar basic PubMed search for “breast cancer” (19th September 2017) reveals 2,149 and 16,920 search results in 1980 and 2016, respectively. The reasons for this discrepancy is likely multifactorial and beyond the remit of this thesis. One factor that is relevant, however, is that much research in OHCA has to be undertaken outside of the hospital, which imposes important barriers and challenges.

1.4.1 Challenges of prehospital and cardiac arrest research

The challenges of prehospital research, when compared to research in hospital or out-patient clinic settings, are manifold. Firstly, there is geography. Clinical care is delivered by EMS

systems which often cover vast areas with varying terrain, climate and population characteristics (Symons and Shuster, 2004). As prehospital care is largely provided in patients’ homes, public places or the inside of a moving ambulance, the capture and collection of research data are challenging (McClelland et al., 2015). Add to this the time pressures of emergency care and the general unpredictability of medical emergencies and it becomes clear why setting up high-quality research studies is challenging. A systematic review of prehospital research in 1999 identified a total of 41 randomised controlled trials, of which only four had a sample size of over 1,000 (Brazier et al., 1999). Other authors note a general lack of volume, quality and funding of emergency and prehospital medicine research (Bounes et al., 2013). Nevertheless, factors such as improved paramedic education, information technology to disseminate and collect information and a general enthusiasm by prehospital providers (McClelland et al., 2015; Hargreaves, Goodacre and Mortimer, 2014) have helped important prehospital studies to be completed successfully (Austin et al., 2010; Van't Hof et al., 2008).

Research focusing on OHCA adds its own challenges, with Herlitz et al. (2007, p.213) remarking that “cardiac arrest is one of the most chaotic events in medicine”. Life-saving diagnostics and treatments need to be delivered in timeframes measured in seconds and minutes, limiting opportunities for data capture, randomisation procedures or adding yet unproven interventions into the complex resuscitation process (Perkins et al., 2015c). The major obstacle to interventional trial designs is that patients in OHCA, by the very nature of the condition, are unable to consent to participation in research (Coats and Goodacre, 2009). Many countries, including the US, Australia and the UK have implemented legal and ethical frameworks that allow patients to be enrolled in research without consent, under strict conditions (van Belle et al., 2015). In the UK, a number of ethical, legal and professional stipulations need to be followed (World Medical Association, 2013; General Medical Council, 2013; Department of Health, 2006). Davies et al. (2014) published a framework of questions and considerations that address the relevant requirements when considering emergency research without consent; see Box 1.2.

Box 1.2 Questions and considerations for research where waiver of consent is considered (Davies et al., 2014)

 Is this research needed?

 Is there uncertainty about treatment?

 Is there a need to recruit subjects who lack capacity?

 In the context of the research, is consent or consultation feasible?  Does treatment need to be given quickly?

 Might delay change the effect of treatment or the results?  Will procedures accommodate variations in capacity?

 Would the legal representative/consultee be likely to have capacity?

 Is it practical to consult a professional legal representative unconnected to the research?

 What should the patient or legal representative be asked later?

The alignment of ethics, law and professional guidelines, with support from research organisations (Nolan et al., 2015a), has enabled large randomised trials of interventions in OHCA in the UK (Taylor et al., 2016; Perkins et al., 2015c). However, academics, the public and prehospital providers continue to show concern about the ethical appropriateness of randomised research in OHCA (Hargreaves, Goodacre and Mortimer, 2014; Smith, 2014). Another challenge in OHCA research is the fact that the outcomes of survival to hospital discharge or good neurological outcome (see Section 1.3.1) are relatively rare and, as described in Section 1.3.2, heavily influenced by variations in patient factors. Furthermore, even small changes in OHCA survival rates are clinically important, given the significance of the alternative outcome of death (Nichol et al., 2016). This means that the sample size required for the detection of clinically significant differences in outcome is often in the range of 5,000 to 10,000 patients, which requires significant logistical and funding support for OHCA prehospital research projects (Taylor et al., 2016).

1.4.2 Out-of-hospital cardiac arrest registries and observational research

Given the challenges of conducting research without informed consent and the need for large sample sizes in OHCA research, it is not surprising that many countries have set up national OHCA registries (Goldberger and Nichol, 2013). These registries allow monitoring of epidemiologic data, including trends over time or regional variation in outcomes (Daya et al., 2015; Nehme et al., 2015b; Wang et al., 2015). Commonly, researchers have used the available data in prospective or retrospective observational studies, examining new interventions or changes in practice (Choi et al., 2016; Schober et al., 2016; Vyas et al., 2015).

In the UK, the University of Warwick hosts the OHCAO Registry, which collects data from participating NHS ambulance trusts; see Chapter 7 for more information (Perkins and Brace- McDonnell, 2015).

A major issue for observational research in OHCA is the fact that outcomes are determined by a number of factors other than the variable of interest (see Section 1.3.2), introducing the risk of confounding and bias (Fouche and Jennings, 2016; Lu, 2009). In an effort to improve reporting of confounding variables and the comparison of different treatments or systems for OHCA, a consensus-based guideline, the Utstein template, was published in 1991 (Cummins, 1991). The Utstein template has since been updated twice, reflecting advances in technology and understanding of OHCA, and has been adopted by most OHCA registries (Perkins et al., 2015a). The template includes mandatory core variables as well as supplemental variables relating to the key epidemiological factors (Section 1.1), interventions (Section 1.2), prognostic factors and outcomes (Section 1.3) discussed previously (Perkins et al., 2015a). See Figure 1.11 for a summary of the core and supplemental variables in the Utstein 2014 template.

Figure 1.11 Variables in the Utstein 2014 template (Perkins et al., 2015a), with permission from Resuscitation (license number 4134100961000)

CPR: Cardiopulmonary resuscitation, AED: Automated external defibrillator, ROSC: Return of spontaneous circulation, DNAR: Do Not Attempt Resuscitation, ECG: Electrocardiogram, STEMI: ST- elevation myocardial infarction, BP: Blood pressure, ECMO: Extracorporeal membrane oxygenation, IABP: Intra-arterial blood pressure

Knowledge of the Utstein variables allows comparison between individual study results and between different groups within observational studies. If variables are not equally distributed within a study, statistical methods such as multiple logistic regression or propensity score matching (see Chapter 7) can be used to adjust outcomes, eliminating or at least reducing confounding and bias (Hasegawa et al., 2015; Do Shin et al., 2012).

Another method of comparing the effects of interventions or different EMS systems, suggested by the Utstein group, is to compare survival only for the subgroup of patients with witnessed OHCA and ventricular fibrillation (VF) as the presenting rhythm (Perkins et al., 2015a). This approach has been criticized, as it results in reporting high survival rates (due to the underlying favourable prognostic factors), which are at odds with the general poor survival rates from resuscitation of unselected cases of OHCA (Timmermans, 1999). On the other hand, the argument for reporting the outcomes in this subgroup of witnessed VF OHCA is twofold. Firstly, the generally higher survival rate means that any beneficial effect from a studied intervention is likely to be more pronounced in this group (Lindner et al., 2011). Secondly, the OHCA subgroups of witnessed VF are relatively homogenous in regards to many of the patient or EMS factors discussed in Section 1.3.2 which influence survival (Perkins et al., 2015a). While the results are certainly not generalizable to all cases of OHCA, the internal validity of the study of OHCA interventions improves when analysing only cases of witnessed VF OHCA (Lu, 2009).