Objective Clinical procedural experience and confidence are both important when performing complex medical procedures. Since out-of-hospital endotracheal intubation (ETI) is a complex intervention, we sought to clarify clinical ETI experience among prehospital rescuers as well as their confidence in performing ETI and confidence-associated factors.
Design Population-based cross-sectional study conducted from January to September 2017.
Setting Northern Japan, including eight prefectures.
Participants Emergency life-saving technicians (ELSTs) authorised to perform ETI.
Outcome measures Annual ETI exposure and confidence in performing ETI, according to a five-point Likert scale. To determine factors associated with ETI confidence, differences between confident ELSTs (those scoring 4 or 5 on the Likert scale) and non-confident ELSTs were evaluated.
Results Questionnaires were sent to 149 fire departments (FDs); 140 agreed to participate. Among the 2821 ELSTs working at responding FDs, 2620 returned the questionnaire (response rate, 92.9%); complete data sets were available for 2567 ELSTs (complete response rate, 91.0%). Of those 2567 respondents, 95.7% performed two or fewer ETI annually; 46.6% reported lack of confidence in performing ETI. Multivariable logistic regression analysis showed that years of clinical experience (adjusted OR (AOR) 1.09; 95% CI 1.05 to 1.13), annual ETI exposure (AOR 1.79; 95% CI 1.59 to 2.03) and the availability of ETI skill retention programmes including regular simulation training (AOR 1.31; 95% CI 1.02 to 1.68) and operating room training (AOR 1.44; 95% CI 1.14 to 1.83) were independently associated with confidence in performing ETI.
Conclusions ETI is an uncommon event for most ELSTs, and nearly half of respondents did not have confidence in performing this procedure. Since confidence in ETI was independently associated with availability of regular simulation and operating room training, standardisation of ETI re-education that incorporates such methods may be useful for prehospital rescuers.
- advanced airway management
- operating room exposure
- perceived anxiety
- prehospital rescuers
- simulation training
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
- advanced airway management
- operating room exposure
- perceived anxiety
- prehospital rescuers
- simulation training
Strengths and limitations of this study
To the best of our knowledge, this is the first population-based survey to clarify the confidence level of prehospital medical rescuers in performing endotracheal intubation (ETI) and to investigate confidence-associated factors.
The response rate was extremely high (more than 90% for all relevant analyses), minimising the non-response bias.
As with any survey using self-administered questionnaires, our study is subject to self-reporting bias, leading to a possible overestimation of clinical ETI experience and confidence among emergency life-saving technicians (ELSTs).
Since this study was not designed to measure patient outcomes, it remains to be clarified how a lack of ETI experience and low confidence among ELSTs affect outcomes of patients with out-of-hospital cardiac arrest.
Out-of-hospital cardiac arrest (OHCA) is a major public health concern worldwide. According to data provided by the Fire and Disaster Management Agency, more than 100 000 cases of OHCA occur annually in Japan.1 Although endotracheal intubation (ETI) has long been considered the standard for definitive airway management in patients with OHCA,2 the effect of ETI during cardiopulmonary resuscitation remains controversial.3–11 Several studies have identified an association between ETI and increased mortality,3–7 whereas others have found a survival benefit of ETI during cardiopulmonary resuscitation.9 10 A recent randomised clinical trial11 failed to demonstrate whether ETI was superior or inferior to conventional bag-valve-mask ventilation for favourable neurological outcome in the OHCA population.
The ETI experience of the laryngoscopist is known to significantly influence the outcome of patients with OHCA.12 Although healthcare professionals involved in airway management are expected to have regular clinical ETI experience, past studies have indicated limited ETI opportunities for paramedics in the USA13 and the UK.14 Emergency medical service in Japan is quite different from that in those countries,3 15 16 and little is known regarding the out-of-hospital ETI experience of Japanese emergency life-saving technicians (ELSTs). Previous studies on this subject in Japan have been limited to single-centre reviews.17 18 To clarify the current situation in Japan, assessment of population-based data from a broader geographical area is necessary.
Out-of-hospital ETI is a complex procedure that can be fraught with errors and severe adverse events, even for skilled laryngoscopists.19–21 Self-confidence plays an important role when performing such high-risk, difficult medical interventions.22–26 For example, previous studies have shown that greater self-confidence correlates with better clinical performance.25 26 An example of this phenomenon from outside the medical field is that more confident athletes perform significantly better than less confident athletes.27 28 Therefore, it is important to understand self-perceived competency and its associated factors in performing out-of-hospital ETI. However, to date, there has been no thorough evaluation of self-confidence levels and associated factors related to airway management skills among prehospital rescuers.
In this study, we investigated (1) clinical ETI experience and (2) self-confidence levels related to ETI and confidence-associated factors among Northern Japanese ELSTs. The findings of this population-based survey reveal areas for improvement and the need for better training programmes for ETI skill retention.
Consent to participate
The board regarded return of the questionnaire as consent to participate.
Study design, setting and subjects
This cross-sectional study was conducted from January 2017 to September 2017 (planning phase, January to June; survey phase, July to September). The emergency medical service system in Japan has been described previously.3 15 16 29 Briefly, an ambulance crew typically consists of three emergency medical service personnel, including at least one ELST who has completed extensive training. These ELSTs are permitted to insert intravenous lines, use semi-automated external defibrillators and use supraglottic airway devices (SGAs) for patients with OHCA. Since 2004, under the direction of online medical control, ETI can be performed in patients with OHCA by specially trained ELSTs who have completed an additional 62 hours of training and performed 30 successful supervised ETI in operating rooms. Since 2011, ELSTs have also been allowed to use rigid video laryngoscopes after completion of additional training.30 ELSTs in Japan are not permitted to perform ETI except in patients with OHCA. The indications for ETI in patients with OHCA31 include: (1) impossibility of maintaining ventilation without ETI, such as foreign-body airway obstruction, and (2) cases in which the medical control doctor judges ETI to be required.
Our target subjects were all ELSTs authorised to perform ETI (defined as advanced-level ELSTs in this study) in Northern Japan, which includes eight prefectures (Hokkaido, Aomori, Iwate, Akita, Miyagi, Yamagata, Fukushima and Niigata). In Northern Japan, 149 fire departments (FDs) with dispatch centres provide emergency medical service for roughly 16.7 million inhabitants in an area of approximately 163 000 km2.
When selecting items for the questionnaire, we referred to relevant studies that similarly assessed ETI experience, proficiency and skill development among paramedics in the USA and UK.13 14 23 32–35 We also referred to previous reports that assessed competence and confidence with airway management skills among military advanced life support providers,24 emergency physicians,36 paediatric emergency medicine fellows,37 38 medical students39 40 and general practitioners.41 We then circulated drafts among the survey team members (an epidemiologist, anaesthesiologists, physicians specialising in emergency medicine and an ELST) before finalising the questionnaire. During the planning phase, the clarity and relevance of each survey item were checked using convenient samples from FDs in Fukushima and Koriyama. English versions of the Japanese questionnaires used in this study are included as online supplementary data S1 and S2.
Survey protocol and items
To target ELSTs allowed to perform ETI, a two-phase postal approach was used. Prepaid return envelopes with preprinted addresses were used throughout the process to increase the response rate, but no incentives were offered. Owing to the satisfactory response rate (see the Results section), no non-response follow-up techniques such as phone calls or reminder letters were used.
First, to obtain data about facility characteristics, self-administered questionnaires (see online supplementary data S3) were mailed to every director of Northern Japanese FDs (149 FDs in eight prefectures) in July 2017. These facilities were extracted from the website of the Japanese Fire Chiefs’ Association.42 A complete list of these FDs is included as online supplementary data S3.
The initial survey asked: (1) the number of ELSTs (both basic and advanced level), (2) the total number of ambulance dispatches and ambulance dispatches for OHCA in 2016 and (3) the availability of a rigid video laryngoscope and its product name. The definition of functional urban area of Organisation for Economic Co-operation and Development countries was used to identify urban areas.43 In brief, urban FDs were defined as those in cities with 50 000 or more inhabitants.
After completion of the initial survey, anonymous questionnaires (see online supplementary data S2) were sent to each responding FD in August 2017. All FD directors were asked to distribute and collect the surveys from advanced-level ELSTs allowed to perform ETI. The secondary survey requested participants’ demographic data, including age; sex; years of experience after achieving basic-level or advanced-level ELST status; provider or instructor status for American Heart Association-certified cardiopulmonary resuscitation courses, including Basic Life Support, Advanced Cardiovascular Life Support and Pediatric Advanced Life Support and provider or instructor status for the Japanese version of the cardiopulmonary resuscitation course (Immediate Cardiac Life Support) and basic trauma life support course (Japan Prehospital Trauma Evaluation and Care). The advanced-level ELSTs were also asked about their ETI procedural experience in 2016 and available ETI skill-maintenance programmes, including regular simulation training using a mannequin and re-education in the operating room. Finally, the survey queried respondents’ confidence in airway management skills (including manual bag-mask ventilation, SGA insertion and ETI); their anxiety about lack of clinical ETI experience, ETI skill retention and lack of proper ETI re-education programmes and the perceived importance of ETI in patients with OHCA and of ETI education on the use of a video laryngoscope. Five-point Likert scales were employed to measure confidence in performing airway-management procedures (1=not confident at all, 2=minimally confident, 3=somewhat confident, 4=confident, 5=very confident) and anxiety regarding clinical ETI experience and skill retention (1=not anxious at all, 2=minimally anxious, 3=somewhat anxious, 4=anxious, 5=very anxious). Participants also indicated their degree of agreement with the following two statements: ‘ETI is an important life-saving procedure for OHCA’ and ‘Education on the use of video laryngoscopes should be strengthened’ (1=completely disagree, 2=disagree, 3=neither agree nor disagree, 4=agree, 5=completely agree).
Outcome measures and statistical analysis
Outcomes of interest in this study were: (1) annual ETI procedural experience and (2) confidence and confidence-associated factors in performing ETI. Annual ETI procedural experience was chosen as an outcome measure because the ELST’s experience with prehospital ETI can influence outcomes of patients with OHCA.12 Since greater self-confidence is correlated with better clinical performance,22–26 confidence in performing ETI was also included as an outcome measure. Responders were blinded to our outcome assessment plans.
All survey items were initially evaluated with descriptive statistics. To determine the factors associated with ETI confidence, differences between confident and non-confident ELSTs were compared. Confident ELSTs were defined as those who reported an ETI confidence of 4 or 5 on the Likert scale. Differences in continuous variables were compared with Student’s t-test or the Mann-Whitney U test for normally and non-normally distributed data, respectively, after application of the Shapiro-Wilk test for normality. Differences in categorical variables were compared with a Χ2 test. Univariable and multivariable logistic regression models were fitted to yield a crude and an adjusted OR (AOR) for confident status in performing ETI. In addition to sex, imbalanced characteristics between confident and non-confident ELSTs (variables with p<0.05 in table 1, see the Results section), such as age, years of clinical experience, certification in video laryngoscope use, annual ETI experience, instructor status for cardiopulmonary and trauma resuscitation courses and availability of specific ETI skill retention programmes, were included as independent variables in the logistic regressions.
In the sensitivity analyses, different definitions for confidence status in performing ETI were used. We repeated the multivariable analyses comparing ELSTs who scored 5 versus ≤4 and ≥3 versus ≤2 on the Likert scale for ETI confidence.
In all multivariable analyses, a variance inflation factor was used to detect multicollinearity. The models’ goodness of fit and discrimination ability were confirmed with the Hosmer-Lemeshow test and the c statistic, respectively. Since less than 3% of data points were missing for all analyses, missing observations were excluded, and complete data sets were used for all relevant analyses. The associations between ETI frequency and ELST characteristics, between confidence in performing ETI and confidence in other airway management skills, and between confidence in performing ETI and anxiety about ETI skill retention were assessed with Spearman’s rank-order coefficient (rs).
All statistical analyses were performed with SPSS Statistics for Windows, V.22.0 (IBM Corp). A p value <0.05 was considered statistically significant.
Patient and public involvement
No patients and public were involved in the development of the research question or the outcome measures nor the design of the study.
Facility characteristics of the responding FDs
In the initial survey, 140 of 149 Northern Japanese FDs returned a completed questionnaire (response rate, 94.0%). Online supplementary table S1 shows the facility characteristics of the responding FDs. The median number of annual ambulance dispatches per FD was 2223 (IQR 1229–4182); the median number of annual ambulance dispatches for OHCA was 70 (IQR 40–152). A rigid video laryngoscope was available at 68.6% of the FDs that responded. Of the 5962 ELSTs working at a responding FD, 2821 (47.3%) were advanced-level ELSTs authorised to perform ETI.
Clinical ETI experience among Northern Japanese ELSTs
Of 2821 advanced-level ELSTs, 2620 returned a completed questionnaire in the second-phase survey (response rate, 92.9%). Of those respondents, complete data were available for 2567 (complete response rate, 91.0%); these were included in all relevant analyses. Figure 1 shows the frequency distribution of annual ETI experience. Among the 2567 ELSTs, 1875 (73.0%) did not have any ETI opportunities, and 2457 (95.7%) were exposed to two or fewer ETI opportunities annually. The median number of ETIs performed by ELSTs was 0 (IQR 0–1; range 0–15). Even distribution of all ETI procedures among all ELSTs would result in only 0.5 ETI per capita. There were negligible correlations between ETI frequency and age (rs=−0.101, p<0.001), years of experience after achieving basic-level (rs=−0.106, p<0.001) or advanced-level (rs=−0.062, p=0.002) ELST status and annual OHCA case load per capita (rs=0.055, p=0.005). There were no significant correlations between ETI frequency and other characteristics of the ELSTs, including male sex (rs=−0.022, p=0.262) and working at an urban FD (rs=−0.007, p=0.733). Although ETI was an uncommon event for most ELSTs, more than half perceived ETI as an important life-saving technique (online supplementary figure S1). Nearly half of respondents believed that education on the use of a video laryngoscope should be strengthened (online supplementary figure S1).
Reported ETI confidence and confidence-associated factors among Northern Japanese ELSTs
As shown in figure 2, approximately 50% of respondents reported confidence (defined as 4 or 5 on the Likert scale) in performing ETI; this percentage was relatively low compared to the percentage reporting confidence in other airway management skills. There were moderate positive correlations between confidence levels in ETI and SGA insertion (rs=0.468, p<0.001) and bag-valve-mask ventilation (rs=0.419, p<0.001; online supplementary table S2).
As shown in figure 3, 87.8% of ELSTs had anxiety (defined as 4 or 5 on the Likert scale) about their lack of ETI experience; 63.5% had anxiety about ETI skill retention and 44.3% about the lack of proper ETI re-education programmes. There was a moderate negative correlation between level of ETI confidence and anxiety about ETI skill retention (rs=−0.458, p<0.001; online supplementary table S2). We also observed a weak negative correlation between level of ETI confidence and anxiety about lack of ETI clinical experience (rs=−0.212, p<0.001) and anxiety about lack of proper ETI skill retention programmes (rs=−0.178, p<0.001; online supplementary table S2).
Table 1 compares the demographic characteristics of ELSTs according to their confidence in performing ETI. Annual ETI experience, age and years of experience after achieving basic or advanced status were associated with confidence in performing ETI (p<0.001 for each). The availability of ETI skill retention programmes, including regular simulation training using a mannequin and re-education in the operating room, was significantly greater (p=0.013 and p=0.002, respectively) for confident ELSTs. Confident ELSTs were also more likely than non-confident ELSTs to be instructors of the Japanese versions of cardiopulmonary resuscitation and basic trauma life support courses (p=0.001 and p=0.038, respectively).
The results of univariable and multivariable analyses for ETI confidence are shown in table 2. After adjustment with the multivariable logistic regression model, years of experience after becoming an advanced-level ELST (AOR 1.09; 95% CI 1.05 to 1.13), annual ETI experience (AOR 1.79; 95% CI 1.59 to 2.03), availability of regular simulation training using a mannequin (AOR 1.31; 95% CI 1.02 to 1.68) and availability of regular operating room training (AOR 1.44; 95% CI 1.14 to 1.83) were independently associated with confidence in performing ETI. The Hosmer-Lemeshow test verified the good fit of this model (p=0.314); the c statistic for this logistic model was 0.745 (95% CI 0.726 to 0.764), suggesting acceptable discrimination.
In sensitivity analyses, the adjusted associations between ETI confidence and years of experience as an advanced-level ELST and annual procedural ETI experience persisted with the use of two different definitions of ETI confidence (online supplementary table S3).
This population-based cross-sectional study conducted in Northern Japan revealed that more than 95% of prehospital rescuers were involved in few or no ETI annually. Nearly half of Japanese ELSTs did not have confidence in their ability to perform ETI, and most had anxiety about their lack of clinical ETI experience and skill retention. Confidence in performing ETI was independently associated with years of clinical experience, annual ETI exposure, availability of regular simulation training and availability of regular operating room training. Since the last two factors are modifiable, ETI re-education that incorporates these training modalities should be considered.
Consistent with prior studies from the UK and the USA13 14 and with single-centre experiences in Japan,17 18 ETI opportunities for most Northern Japanese ELSTs in this study were limited. Previous studies did not assess the associations between ETI opportunities and the demographic data of rescuers.13 14 17 18 To address this knowledge gap, we examined these relationships but found almost no correlations between ETI frequency and characteristics of the ELSTs, including age, sex, years of experience and urban versus rural setting. These data suggest that ETI opportunities are equally limited among all ELSTs in Northern Japan. The reasons for the limited ETI experiences among Japanese ELSTs are likely multifactorial; potential explanations include strict ETI protocols,31 rare recommendation by medical controls for ETI, prehospital rescuers’ hesitation in performing interventions in which they lack confidence or some combination of these factors. With limited experience, it is difficult to maintain proficiency in out-of-hospital ETI. Confirming this association, most ELSTs were anxious about their lack of clinical ETI experience and skill retention, and ETI confidence was associated with annual exposure. Many Japanese ELSTs are likely frustrated because more than half regarded ETI as an important life-saving technique.
Inadequate ETI procedural experience, low confidence and high anxiety among Japanese ELSTs might lead to poorer outcomes for patients with OHCA who need advanced airway management. In fact, a previous study indicated that ETI by Japanese ELSTs was independently associated with poorer neurological outcomes among adult patients with OHCA.3 One option to address this problem is to remove ETI from the skill set of ELSTs, as previously advocated by Wang et al.13 However, this approach disregards situations in which ETI is indicated, such as airway obstruction. Since annual ETI exposure is independently associated with ETI confidence, another option is to assure exposure by concentrating ETI in the hands of fewer skilled providers.13 However, in Japan, there is currently no agency responsible for accreditation of ETI performance among ELSTs.18 A previous study found that a rigid video laryngoscope enabled Japanese ELSTs to achieve a high ETI success rate in the operating room, regardless of previous experience with a direct laryngoscope.30 A third option to address the current situation is therefore strengthening education on the use of a video laryngoscope. Nearly half of surveyed ELSTs agreed with this idea. However, in this study, there was insufficient availability both of video laryngoscopes and of ELSTs permitted to use the device. Our findings reveal the current situation to policy-makers in the community and in national organisations and provide the opportunity to rethink the current practical configuration of advanced prehospital airway management for the OHCA population.
In Japan, airway management skill re-education programmes for ELSTs depend on local medical controls and are not standardised.18 We believe that improving and standardising airway intervention re-education is important to provide better prehospital care. In this study, retraining in the operating room was independently associated with ETI confidence, but opportunities for this training were limited. Similarly, limited opportunities for operating room exposure for prehospital rescuers have been reported in the USA.32 Although previous research has documented that simulation training is useful for the development of critical intervention skills, data regarding how this training modality affects competency in performing ETI among prehospital rescuers are scare.44 In this population-based survey, we observed that the availability of regular simulation training was independently correlated with ETI confidence among Japanese ELSTs. Collectively, our data underscore the need for reinforcement of airway re-education methodology (eg, training in a controlled setting and the use of simulators) for inexperienced providers to improve their ETI confidence. Improved confidence will, in turn, improve performance. This increased confidence may also be beneficial for other airway management skills, because the level of ETI confidence was moderately associated with confidence in bag-valve-mask ventilation and SGA insertion.
Our survey also revealed that more than 40% of ELSTs were anxious about the lack of proper ETI skill retention programmes currently available in Japan. Japanese national bureaus, professional organisations, local medical controls, anaesthesiologists and emergency physicians should work together to address this problem. We believe that there is an enhanced opportunity to improve the quality of ETI re-education programmes, if the leadership and guidance of governmental and professional agencies are strengthened.
We believe our study has several implications. For ELSTs, our observations provide a reference point regarding their ETI procedural experience, confidence and available skill maintenance programmes. Additionally, for decision-makers, professional organisations and medical controls, our findings indicate room for improvement and suggest that standardisation and dissemination of appropriate nationwide ETI re-education training are warranted.
Study limitations and advantages
Our study had several limitations. First, as with any cross-sectional study that uses a self-administered questionnaire, self-reporting bias (both social desirability and recall bias) was possible. Since clinicians and FD directors were involved in the survey collection process, there may also have been administration bias. If so, ETI experience and confidence may be even poorer than those reported in this survey. To decrease the effects of social desirability and administration bias, we used anonymous questionnaires in the second-phase survey. Responders were also blinded to our outcome assessment plans. To mitigate recall bias, we asked ELSTs for their most recent 1 year of ETI experience. Given these potential biases, achieving a high response rate is critical to ensure the quality of data in an epidemiological survey; one major advantage of the present study is that it had few non-responders (less than 10% for all relevant analyses).
Second, we did not design this study to measure patient outcomes or the ETI procedure itself. Nevertheless, we speculate that lack of ETI confidence might worsen the outcomes of patients with OHCA, because this study showed that ETI confidence was significantly associated with annual ETI experience, and a previous study12 demonstrated that ETI experience significantly influenced the outcomes of patients with OHCA. Further studies are required to clarify how lack of ETI confidence affects the ETI procedure and outcomes of patients with OHCA.
Third, our population-based study describes the situation in Northern Japan only. A similar study with data from other areas of Japan or other countries could result in different findings. For example, while ETI is a relatively new skill for ELSTs in Japan, paramedics in the USA have performed ETI in clinical practice for over 30 years and may possess greater clinical exposure to and comfort with ETI.45
Despite these limitations, this study also had several strengths. In addition to the above-mentioned high-response rate, this study is the first to investigate the factors associated with ETI confidence among prehospital rescuers. Our survey provides an opportunity to re-evaluate current ETI practice and re-education programmes among Northern Japanese ELSTs. We believe that the quality improvement implications of our results would be beneficial not only for our study population but also for other countries.
This population-based cross-sectional study revealed that most Northern Japanese ELSTs were involved in only a few or no ETI annually and lacked confidence in performing ETI. In addition, many had anxiety about ETI skill retention and felt that proper ETI re-education programmes were lacking. ETI re-education that incorporates regular simulation training and operating room exposure may be beneficial for prehospital rescuers, because the availability of those two training modalities was independently associated with ETI confidence.
We thank all of the participating FDs and ELSTs for their earnest cooperation in this project. We also thank the following persons: Ms Siho Sato (Emergency and Critical Care Medical Center, Fukushima Medical University Hospital, Fukushima) and Ms Kasumi Ouchi (Office for Gender Equality Support, Fukushima Medical University, Fukushima) for their secretarial assistance; Mr Takashi Minowa (an advanced-level ELST, Fukushima FD, Fukushima) for his assistance in generating the questionnaire; Aya Goto, MD, PhD, MPH (Integrated Center for Science and Humanities, Fukushima Medical University, Fukushima) and Nozomi Ono, MD (Department of Psychiatry, Hoshigaoka Hospital, Koriyama, Japan) for their assistance in reviewing the manuscript and Rebecca Tollefson, DVM, from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.
Contributors YO conceived the study design, drafted the initial manuscript and takes primary responsibility for the paper as a whole. KS, KT and KI supervised the undertaking of the survey and data collection. YO and TK managed the data and performed the statistical analysis. All authors contributed to the construction of the questionnaire, interpreted the survey results, participated in related discussions, contributed substantially to its revision and read and approved the final version of the manuscript.
Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Ethics approval This study was approved by the Institutional Review Board at Fukushima Medical University (no 2989) on 23 February 2017.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Extra data can be accessed via the Dryad data repository at http://datadryad.org/ with doi: 10.5061/dryad.f1vn500.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.