The following quote from the article is incorrect and misses the basic definition of power: "Strictly speaking 'power' refers to the number of patients required to avoid a type II error in a comparative study. Sample size estimation is a more encompassing term that looks at more than just the type II error and is applicable to all types of studies. In common parlance the terms are used interchangeably."
"Power" is the probability that the test correctly rejects the null hypothesis H0 when a specific alternative hypothesis H1 is true. It's equal to 1 - type II error probability. "Power" and "sample size" are not the same thing and they are not used interchangeably. It's possible to derive power given sample size, or calculate sample size based on desired power.
Please correct the article as it'll be highly misleading to beginners.
In an observational study where 200 participants were black, 269 asian, and 4330 white, the authors demonstrated an inverse association between blood pressure and pulse oximetry accuracy that was not influenced by ethnicity[1]. In that study no specific mention was made of the degree of pigmentation in individual members of the ethnic subgroups, presumable because self-reported ethnicity was accepted as a surrogate for skin colour. This acceptance is in sharp contrast with the methodology in the study where subjects of African-American descent were further characterised by a description of their degree of pigmentation, using terminology such as "very darkly pigmented".. This was one of the earliest prospective studies conclusively to show that some oximeters overestimate arterial oxygen saturation in hypoxic subjects who are "darkly pigmented" [2].
In retrospective studies such as the ones subsequently undertaken to explore the theme of racial bias in oximetry it was easy to fall into the trap of using ethnicity as a surrogate for skin colour[3],[4], largely because skin colour is not consistently recorded as part of the medical record[3]. Explicit description of skin colour also gets omitted when race and ethnicity are defined using self-reported demographic data[4].
Future studies, however, might seek to ascertain whether or not skin pigmentation compounds the overestimation of oxygen saturation attributable to hypotension....
In an observational study where 200 participants were black, 269 asian, and 4330 white, the authors demonstrated an inverse association between blood pressure and pulse oximetry accuracy that was not influenced by ethnicity[1]. In that study no specific mention was made of the degree of pigmentation in individual members of the ethnic subgroups, presumable because self-reported ethnicity was accepted as a surrogate for skin colour. This acceptance is in sharp contrast with the methodology in the study where subjects of African-American descent were further characterised by a description of their degree of pigmentation, using terminology such as "very darkly pigmented".. This was one of the earliest prospective studies conclusively to show that some oximeters overestimate arterial oxygen saturation in hypoxic subjects who are "darkly pigmented" [2].
In retrospective studies such as the ones subsequently undertaken to explore the theme of racial bias in oximetry it was easy to fall into the trap of using ethnicity as a surrogate for skin colour[3],[4], largely because skin colour is not consistently recorded as part of the medical record[3]. Explicit description of skin colour also gets omitted when race and ethnicity are defined using self-reported demographic data[4].
Future studies, however, might seek to ascertain whether or not skin pigmentation compounds the overestimation of oxygen saturation attributable to hypotension. For those studies to be scientifically robust explicit documentation of the degree of pigmentation will have to be made., using, for example, strategies such as the Melasma Area and Severity Index[5], and the Willis and Earle scale[6], to mention a few.
I have no funding and no conflict of interest
References
[1]Crooks CJ., West J., Morling J et al
Inverse association between blood pressure and pulse oximetry accuracy: an observational study in patients with suspected or confirmed COVID-19 infection
Emergency Medical Journal Article in Press
doi:10.1136/emermed-2022-212443
[2]Bickler PE., Feiner JR., Severinghaus SW
Effects of skin pigmentation on pulse oximeter accuracy at low saturation
Anesthesiology 2005;102:715-719
[3]Valbuena VSM., Seelye S., Sjoding MW et al
Racial bias and reproducibility in pulse oximetry among medical and surgical inpatients in general care in Veterans Administration 2013-19: multicenter retrospective cohort study
BMJ 2022;378:e069775
[4]Wong A-K I., Charpignon M., Kim H et al
Analysis of discrepancies between pulse oximetry and arterisl oxygen saturation measurements by race and ethnicity and association with organ dysfunction and mortality
JAMA Netweork OPEN 2021;4:e2131674
[5]Pandya AG., Hynan LS., Bhore R et al
Reliability assessment and validation of the Melasma Area and Severity Index(MASI) and a new modified MASI scoring method
J Am Acad Dermatol 2011;64:78-83
[6]Roberts W
Skin type classification systems old and new
Dermatology Clinics 2009;27:529-533
We thank Dr. Delaney and colleagues for their valuable research into the concept of midline cervical tenderness. Unlike the NEXUS critiera, the Canadian C-Spine Rule does not use midline tenderness as a positive indication for imaging. Our original study in JAMA 2001 found that assessment of this criterion amongst alert trauma patients at risk of c-spine tenderness had excellent interobserver agreement between ED physicians with a kappa of 0.78. We found that absence of midline tenderness was a good negative predictor of c-spine injury but that presence of of such tenderness was non-specific and not useful. Hence, absence of midline tenderness is considered a low-risk factor. Our NEJM 2003 validation study found that the CCR had both better sensitivity and specificity than NEXUS.
Best regards
1. Stiell IG, Wells GA, Vandemheen K, Clement C, Lesiuk H, De Maio VJ et al. The Canadian Cervical Spine Radiography Rule for alert and stable trauma patients. JAMA 2001; 286(15):1841-1848.
2. Stiell IG, Clement C, McKnight RD, Brison R, Schull MJ, Rowe BH et al. The Canadian C-spine Rule versus the NEXUS low-risk criteria in patients with trauma. N Engl J Med 2003; 349:2510-2518.
A well timed project with the contemporary interest in the subject of drug misuse in Scottish politics, and also with the recent airing of Dopesick on streaming services about the US experience.
I wonder whether the known CYP2D6 polymorphism leading to poor metabolism of codeine has a role in the increase of opioid prescriptions? Mikus and Weiss (2005) state that 5-10% of Caucasian have severely impaired metabolism of codeine and that a further 10-15% show some impairment. This means up to 1 in 4 Caucasians don't get a full analgesic effect from codeine.
Maybe we are seeing a lessening of patients putting up with their lot of suboptimal pain control? And in turn an increase in prescriptions to accommodate that demand.
Dihydrocodeine by contrast doesn't have the same issues with metabolism and ineffective analgesia. While acknowledging the past issues when DF118s were misused, perhaps prescribing dihydrocodeine instead of codeine we'll see better analgesia in our patients and perhaps a reduction in demand for prescriptions?
Mikus G and Weiss J. (2005) 'Influence of CYP2D6 Genetics on Opioid Kinetics, Metabolism and Response', Current Pharmacogenomics, 3, pp43-52
We read with interest your experience of creating care pathways for patients in the out-of-hospital setting during the Covid-19 pandemic, in particular for those with palliative care needs. The benefits of the Physician Response Unit (PRU) being tasked to end of life care related 999 calls, their enhanced level of assessment and management and the resulting reduction in inappropriate hospital admission is to be applauded. There is additionally a need to recognise the wider use of such care pathways, available pan London, that support palliative and end of life care (EoLC) patients accessing emergency care.
In collaboration with hospice and palliative care teams the London Ambulance Service NHS Trust (LAS) has to date, created 19 EoLC appropriate care pathways. These offer our clinicians access to specialist advice, support with complex decision making and provide an alternative to Emergency Department (ED) conveyance. In the last 2.5 years we have undertaken an extensive programme to improve EoLC within the LAS; providing tailored education to augment our clinicians’ knowledge and confidence, creating guidance which includes medications and symptom management at the end of life and increasing clinician use of advance care plans. As a result we have seen a 15% increase in staff confidence and an 18% median reduction in ED conveyance for this patient group; most importantly more patients are being cared for in their place of choice and in line with their wishes....
We read with interest your experience of creating care pathways for patients in the out-of-hospital setting during the Covid-19 pandemic, in particular for those with palliative care needs. The benefits of the Physician Response Unit (PRU) being tasked to end of life care related 999 calls, their enhanced level of assessment and management and the resulting reduction in inappropriate hospital admission is to be applauded. There is additionally a need to recognise the wider use of such care pathways, available pan London, that support palliative and end of life care (EoLC) patients accessing emergency care.
In collaboration with hospice and palliative care teams the London Ambulance Service NHS Trust (LAS) has to date, created 19 EoLC appropriate care pathways. These offer our clinicians access to specialist advice, support with complex decision making and provide an alternative to Emergency Department (ED) conveyance. In the last 2.5 years we have undertaken an extensive programme to improve EoLC within the LAS; providing tailored education to augment our clinicians’ knowledge and confidence, creating guidance which includes medications and symptom management at the end of life and increasing clinician use of advance care plans. As a result we have seen a 15% increase in staff confidence and an 18% median reduction in ED conveyance for this patient group; most importantly more patients are being cared for in their place of choice and in line with their wishes.
We welcome further opportunities for collaboration, learning and multidisciplinary education between the LAS and the PRU as we collectively seek to enhance our emergency response to patients nearing their end of life.
Within their conclusion, Charlton et al recommend further research to understand patient behaviour toward seeking help during the pandemic. In response to this, we would like to highlight the findings of our work which address this. We undertook a mortality review of all deaths in Salford during the peak 7 weeks of the initial pandemic surge (522 deaths), looking at themes which, if addressed could result in reduced mortality in future waves. We reviewed all 111, 999, general practice and hospital contacts for all patients from the 1st March 2020, to ensure all help seeking behaviour and the system response was understood. We have summarised these here.
We noted 60 cases where patients delayed seeking help. We were also concerned that patients who were advised to call 111 by their GP, and were offered advice, would only call back when seriously ill. “COVID phobia” was evident in a small number of cases, at its extreme, including refusal to attend hospital and subsequent death at home.
Fewer than half of NHS 111 calls were answered during the review period. Of those that were answered, 46% resulted in advice to contact their GP. Of these, 5 were subsequently admitted to hospital later the same day. Indeed, despite the national directive to telephone 111 as the first point of contact, only 13% actually did. 81% of patients contacted their GP in the period prior to their death.
However, a lack of early face to face assessments was identi...
Within their conclusion, Charlton et al recommend further research to understand patient behaviour toward seeking help during the pandemic. In response to this, we would like to highlight the findings of our work which address this. We undertook a mortality review of all deaths in Salford during the peak 7 weeks of the initial pandemic surge (522 deaths), looking at themes which, if addressed could result in reduced mortality in future waves. We reviewed all 111, 999, general practice and hospital contacts for all patients from the 1st March 2020, to ensure all help seeking behaviour and the system response was understood. We have summarised these here.
We noted 60 cases where patients delayed seeking help. We were also concerned that patients who were advised to call 111 by their GP, and were offered advice, would only call back when seriously ill. “COVID phobia” was evident in a small number of cases, at its extreme, including refusal to attend hospital and subsequent death at home.
Fewer than half of NHS 111 calls were answered during the review period. Of those that were answered, 46% resulted in advice to contact their GP. Of these, 5 were subsequently admitted to hospital later the same day. Indeed, despite the national directive to telephone 111 as the first point of contact, only 13% actually did. 81% of patients contacted their GP in the period prior to their death.
However, a lack of early face to face assessments was identified as a theme of concern. Further to this, patients could then be faced with delays in ambulance attendance, evidenced by 44% of Category 2 ambulance response taking over two hours in March 2020 (These delays were mitigated by changes to the national ambulance protocol in early April).
Our findings demonstrate that not only was there concerning evidence of patients avoiding healthcare contact, but the Healthcare system response was also inadequate to allow appropriate assessment in many cases. Furthermore, patients vulnerable to health inequality such as older people and those with learning disabilities, appear to have suffered the greatest risk of adverse outcome as a result of these changes in system response.
Our review led us to make several recommendations which informed the subsequent local response. These included:
• Clear messages needed about using NHS services when required;
• “It is vital anyone who thinks they need any kind of medical care comes forward and seeks help “
• Patients must be encouraged to contact NHS 111 or GP if COVID-19 positive or with symptoms
• Ask about the health of others in the house, particularly if not tested or vulnerable
• If the ambulance can not respond quickly, just get to hospital
• Be clinically curious, don’t just assume it is COVID (diagnostic overshadowing)
• Establishing virtual ward with easy ‘onboarding’ referral processes.
• Review ceiling of care decisions in light of new knowledge of the disease
• Care and Residential Homes may need additional support for end of life care
• Ensure extra support and reasonable adjustment should be made for patients with learning disabilities, limited English language skills, or other communication difficulty.
Sincerely,
Dr J Tankel, Dr D Ratcliffe, Prof D Green , Dr M Smith, Dr P Bennett
1. Charlton K et al. Incidence of emergency calls and out-of-hospital cardiac arrest deaths during the COVID-19 pandemic: findings from a cross-sectional study in a UK ambulance service. BMJ 2021. doi: http://dx.doi.org/10.1136/emermed-2020-210291
2. Tankel JW et al. Consequences of the emergency response to COVID-19: a whole health care system review in a single city in the United Kingdom. BMC Emergency Medicine. 2021 doi: https://doi.org/10.21203/rs.3.rs-134729/v1
In their article “Agreement and predictive value of the Rockwood Clinical Frailty Scale (CFS) at emergency department triage”, Shrier et al (1) nicely illustrate that the level of agreement between different health care professionals in obtaining the CFS at different clinical settings is weak. The CFS was rated for 8,568 patients over 65 years by the triage nurse at the emergency department (ED) and by the attending physician on admission on the ward. Both scores were compared using the Cohen’s kappa coefficient, which was 0.21 and therefore is considered weak (2).
We found similar results in our AmsterGEM study. The AmsterGEM study is an observational prospective cohort study that investigates the prognostic accuracy of frequently used frailty screening instruments (3). Patients aged over 70 years old attending the ED were screened with four screening instruments, including the Safety Management System (VMS) (4). The VMS consists of four geriatric domains, which are associated with functional decline: delirium, malnutrition, falls and ADL limitations (5,6). All Dutch hospitals have implemented the VMS screening instrument for hospitalized older patients. In a sub-analysis of our study, the VMS was applied on 173 hospitalized patients over 70 years old, at the ED by a research student and at admission by the attending nurse on the ward. The average age was 81.2 years old, 81 (47%) were male and 111 s...
In their article “Agreement and predictive value of the Rockwood Clinical Frailty Scale (CFS) at emergency department triage”, Shrier et al (1) nicely illustrate that the level of agreement between different health care professionals in obtaining the CFS at different clinical settings is weak. The CFS was rated for 8,568 patients over 65 years by the triage nurse at the emergency department (ED) and by the attending physician on admission on the ward. Both scores were compared using the Cohen’s kappa coefficient, which was 0.21 and therefore is considered weak (2).
We found similar results in our AmsterGEM study. The AmsterGEM study is an observational prospective cohort study that investigates the prognostic accuracy of frequently used frailty screening instruments (3). Patients aged over 70 years old attending the ED were screened with four screening instruments, including the Safety Management System (VMS) (4). The VMS consists of four geriatric domains, which are associated with functional decline: delirium, malnutrition, falls and ADL limitations (5,6). All Dutch hospitals have implemented the VMS screening instrument for hospitalized older patients. In a sub-analysis of our study, the VMS was applied on 173 hospitalized patients over 70 years old, at the ED by a research student and at admission by the attending nurse on the ward. The average age was 81.2 years old, 81 (47%) were male and 111 scored a positive VMS by the nurse and 132 scored a positive VMS by the student. The level of agreement evaluated by Cohen’s kappa coefficient was weak (0.25), as shown in table 1. The domain 'falls' showed the highest agreement, with an acceptable kappa coefficient of 0.61. This could be explained by the fact that questions concerning increased fall risk were asked strictly as written in the VMS. In addition, the percentage of patients with a positive frailty score was higher for students than for nurses (36% vs 24%). Nurses stated they often relied on their own impression of the patients when filling in the questions, where students asked the questions literally. Older patients with memory problems or the clinical setting are other reasons for the low level of agreement. The ED is a busy and hastily place for the patient, caregiver and healthcare professional, which may complicate frailty scoring.
In conclusion, the awareness of the increased risk of negative health outcomes in frail older patients is an important reason why frailty screening should be part of the initial evaluation of older patients. However, the low inter-rater reliability contributes to moderate clinical accuracy of frailty instruments. Hence, training of staffs who apply screening instruments is necessary to minimalize interrater-variability.
Sincerely,
W. van der Heiden, MD
C.S. van Dam, MD
M.J.L. Peters, MD PhD
M.C. Trappenburg, MD PhD
Department of Internal Medicine
Ziekenhuis Amstelland, Amstelveen, the Netherlands
Department of Internal Medicine and Geriatrics
Amsterdam UMC, location VUmc, Amsterdam, the Netherlands
Table 1. Positive VMS scores per domain by student and nurse, with corresponding kappa.
Student
Nurse
Kappa
Delirium
81
43
0.23
Falls
73
53
0.61
Malnutrition
89
77
0.18
ADL limitations
73
95
0.21
Total VMS
132
111
0.25
References
1. Shrier W, Dewar C, Parrella P, Hunt D, Hodgson LE. Agreement and predictive value of the Rockwood Clinical Frailty Scale at emergency department triage. Emerg Med J. 2020 Nov;
2. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977 Mar;33(1):159–74.
3. C.S. van Dam, M.C. Trappenburg, M.M. ter Wee, E.O. Hoogendijk, H.C.W. de Vet, Y.M. Smulders, P.W.B. Nanayakkara, M. Muller MJLP. The accuracy of four frequently used frailty instruments for the prediction of adverse health outcomes among older adults at two Dutch Emergency Departments: findings of the AmsterGEM study. Manuscript submitted for publication.
4. van Noord I, Zwijnenberg N, Wagner C. Patiëntveiligheidscultuur in Nederlandse Ziekenhuizen: Een stap in de goede richting. 2013. 44 p.
5. Heim N, van Fenema EM, Weverling-Rijnsburger AWE, Tuijl JP, Jue P, Oleksik AM, et al. Optimal screening for increased risk for adverse outcomes in hospitalised older adults. Age Ageing. 2015 Mar;44(2):239–44.
6. Heim N. Kwetsbare patiënten opsporen met VMS. Med Contact (Bussum). 2013;68(2):94–6.
Having recently updated our Emergency Department guidelines for suspected PE in pregnancy, we read the secondary analysis of the DiPEP study with great interest.1 However, we were quite surprised at the poor overall D-dimer sensitivity. Only 66% (8/12) of PEs would have been identified based on the recommended positivity threshold of 400ng/ml. This is considerably lower than the pooled estimate of 97% (95% CI 96-98%) found by a recent meta-analysis evaluating D-dimer for PE, and largely explains the poor performance of the YEARS and Geneva algorithms in the DiPEP cohort.2
This result does not seem to fit with the known physiology of pregnancy. We know that D-dimer levels increase throughout pregnancy, which should improve sensitivity and worsen specificity.3 To our knowledge there are no other studies demonstrating impaired sensitivity of D-dimer in pregnant vs. non-pregnant populations.
The DiPEP authors note that most of the study participants had received anticoagulation before blood samples were taken, which can decrease D-dimer levels by up to 25% in the first 24 hours.1 They also note however, that this would be insufficient to explain all their false negative D-dimer results. Aside from random error, we wondered if anything else could explain the poor sensitivity.
One feature of the DiPEP study that stood out to us was the D-dimer assay used. As a microplate ELISA assay, the Zymutest D-dimer should be very sensitive but we could not find any st...
Having recently updated our Emergency Department guidelines for suspected PE in pregnancy, we read the secondary analysis of the DiPEP study with great interest.1 However, we were quite surprised at the poor overall D-dimer sensitivity. Only 66% (8/12) of PEs would have been identified based on the recommended positivity threshold of 400ng/ml. This is considerably lower than the pooled estimate of 97% (95% CI 96-98%) found by a recent meta-analysis evaluating D-dimer for PE, and largely explains the poor performance of the YEARS and Geneva algorithms in the DiPEP cohort.2
This result does not seem to fit with the known physiology of pregnancy. We know that D-dimer levels increase throughout pregnancy, which should improve sensitivity and worsen specificity.3 To our knowledge there are no other studies demonstrating impaired sensitivity of D-dimer in pregnant vs. non-pregnant populations.
The DiPEP authors note that most of the study participants had received anticoagulation before blood samples were taken, which can decrease D-dimer levels by up to 25% in the first 24 hours.1 They also note however, that this would be insufficient to explain all their false negative D-dimer results. Aside from random error, we wondered if anything else could explain the poor sensitivity.
One feature of the DiPEP study that stood out to us was the D-dimer assay used. As a microplate ELISA assay, the Zymutest D-dimer should be very sensitive but we could not find any studies demonstrating its test characteristics. Product information available online contained no references and states it is “for research use only and should not be used for patient diagnosis or treatment”.4 This is perhaps a concern when considering whether we can generalise these results to clinical practice. In contrast, the Artemis5 and CT-PE-Pregnancy6 studies used assays that are well validated, highly sensitive, and widely used in clinical practice:7
- Quantitative latex agglutination assays: TinaQuant (Roche); Liatest (Stago); Innovance (Siemens); HemosIL (Instrumentation Laboratory).
- Rapid modified ELISA assays (ELFA): VIDAS (Biomerieux).
A second feature of the DiPEP study that stood out was the method in which samples were stored and analysed. After collection and initial processing, samples were stored in freezers for “the duration of the study until all samples were transported for analysis to Guy’s St Thomas Trust”. This presumably could be up to 18 months as recruitment occurred between February 2015 and August 2016. The Zymutest D-dimer product information however, states that samples should only be frozen for up to 6 months.4 A 2015 review by Kline and Kabrhel noted that false negative D-dimer results can be caused by “proteolysis that can occur with prolonged time from sample draw to analysis”.3
Could this extended storage of blood samples possibly explain the low D-dimer sensitivity and therefore poor performance of the revised YEARS and Geneva algorithms in this cohort? Given these limitations, the iatrogenic risks of imaging, and the imperfect sensitivity and specificity of CTPA (94% vs. 98%) and V/Q scans (96% vs. 95%),2 it is not clear to us that a “scan-all strategy” for PE in pregnancy appropriately balances the competing risks of under-diagnosis and over-treatment.
References
1. Goodacre S, Nelson-Piercy C, Hunt B, & Fuller G. Accuracy of PE rule-out strategies in pregnancy: secondary analysis of the DiPEP study prospective cohort. Emerg Med J. 2020;37:423-28.
2. Patel P, Patel P, Bhatt M, Braun C, Begum H, Wiercioch W et al. Systematic review and meta-analysis of test accuracy for the diagnosis of suspected pulmonary embolism. Blood Adv. 2020;4(18):4296-311.
3. Kline J, & Kabrhel C. Emergency evaluation for pulmonary embolism, part 2: Diagnostic approach. J Emerg Med. 2015;49(1):104-17.
4. https://www.aniara.com/mm5/PDFs/IFU/IFU_ARK023A.pdf
5. Van der Pol L, Tromeur C, Bistervels I, ni Ainle F, van Bemmel T, Bertoletti L et al. Pregnancy-adapted YEARS algorithm for diagnosis of suspected pulmonary embolism. N Engl J Med. 2019;380:1139-49.
6. Righini M, Robert-Ebadi H, Elias A, Sanchez O, Le Moigne E, Schmidt J et al. Diagnosis of pulmonary embolism during pregnancy. Ann Intern Med. 2018;169(11)766-773.
7. Righini M, Perrier A, de Moerloose P, & Bounameaux H. D-dimer for venous thromboembolism diagnosis: 20 years later. J Thromb Haemost. 2008;6:1059-71.
To the Editor
We have read with great interest the recent article of Miller et al1 determining the technical factors associated with first-pass success (FPS) during endotracheal intubation with C-MAC video laryngoscope (VL) in children. They showed that placement of the blade tip into the epiglottic vallecula regardless of blade types, adequate glottic view and locating the glottic opening within second quintile of video displayer were significantly associated with FPS. Given that paediatric airway management is a great challenge to emergency physicians and the benefits of videolaryngoscopy are often significant in airway management of emergency paediatric patients,2 their findings have potentially clinical implications. Other than limitations described by authors in discussion, however, we noted several methodological issues in their article on which we invited authors to comment.
First, primary outcome of this study was FPS, which was defined as passage of C-MAC VL into the mouth with the intention of intubation that terminated with successful intubation at first attempt. As described by authors in introduction, however, C-MAC VL is an intubating device with ability to perform both direct and video laryngoscopy using same device. That is, the larynx can be seen either under direct vision or on a monitor when using C-MAC VL.3 This advantage of C-MAC VL makes it exceptionally useful for emergency intubation. For example, in the event of a failed video laryngosc...
To the Editor
We have read with great interest the recent article of Miller et al1 determining the technical factors associated with first-pass success (FPS) during endotracheal intubation with C-MAC video laryngoscope (VL) in children. They showed that placement of the blade tip into the epiglottic vallecula regardless of blade types, adequate glottic view and locating the glottic opening within second quintile of video displayer were significantly associated with FPS. Given that paediatric airway management is a great challenge to emergency physicians and the benefits of videolaryngoscopy are often significant in airway management of emergency paediatric patients,2 their findings have potentially clinical implications. Other than limitations described by authors in discussion, however, we noted several methodological issues in their article on which we invited authors to comment.
First, primary outcome of this study was FPS, which was defined as passage of C-MAC VL into the mouth with the intention of intubation that terminated with successful intubation at first attempt. As described by authors in introduction, however, C-MAC VL is an intubating device with ability to perform both direct and video laryngoscopy using same device. That is, the larynx can be seen either under direct vision or on a monitor when using C-MAC VL.3 This advantage of C-MAC VL makes it exceptionally useful for emergency intubation. For example, in the event of a failed video laryngoscopy attempt, intubators can immediately switch direct laryngoscopy to continuously perform intubation procedure without the need to remove device from the mouth for a second attempt, and vice versa.4 In fact, available evidence indicates that common reasons for failed first attempt with C-MAC VL are the fogging of the lens, bloods and secretions on the camera, poor glottic view and video dysfunction (screen failure or inability to start the video function), rather than technical factors.3 As the authors did not provide the causes of failed first attempt with C-MAC VL, it was unclear whether all of the failed first attempts should be attributable to technical factors.
Second, types of initial laryngoscopists, such as resident, fellow and attending physicians, were used for assessment on the competence level of intubators with C-MAC VL in this study. These may not be good proxies for the level of competence with an intubating device, as the competence with an intubating tool is mainly dependent on number of successful use.5 The recent work of Amalric et al6 in showed that the operators' expertise, which is assessed by number of previous video laryngoscopies performed, is an independent determinant of FPS with McGrath MAC video laryngoscope in critically ill patients.
Third, the method of building multivariable model to determine clinical variables associated with FPS was questioned. In fact, the data provided by authors in table 2 are the results of initial comparisons on demographics and intubation characteristics between patients with and without FPS, rather than the results of univariate analysis for demographics and intubation data, which include P values and unadjusted odds ratios with their 95% confidence intervals. According to general principles of establishing a multivariable model, only the variables with statistical significance (P<0.05) in the initial comparison can be incorporated into the univariable model to examine multicollinearity among candidate variables. Furthermore, only the variables with large P values (P<0.2) in the univariate analysis can be entered into the multivariable model for statistical adjustment. By such adjustment of patients’ baseline characteristics and controlling of selection biases, the independent factors associated with FPS can be identified. Furthermore, independent contribution of each significant factor to FPF can also be assessed by its adjusted odds ratios and 95% CI.7 Because of this statistical question, we are concerned that the inference of multivariable analysis to determine clinical variables associated with FPS would have been distorted.
Finally, the appropriateness of multivariable model to determine clinical variables associated with FPS was not assessed by the Hosmer-Lemeshow goodness-of-fit test. Furthermore, the c-index of multivariable model was not calculated. Thus, it cannot determine whether the multivariable model established in this study has a good fit and an adequate discriminative performance.7
Contributors
TT suggested comments and drafted the manuscript. BH revised the comments and the manuscript. FSX critically revised the comments and wrote the manuscript. All authors have read and approved the final manuscript.
Funding
The authors have not declared a specific grant for this work from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests
None declared.
Patient and public involvement
Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this work.
Patient consent for publication
Not required.
REFERENCES
1. Miller KA, Monuteaux MC, Nagler J. Technical factors associated with first-pass success during endotracheal intubation in children: analysis of videolaryngoscopy recordings. Emerg Med J. 2020 Nov 10: doi: 10.1136/emermed-2020-209700.
2. Miller KA, Nagler J. Advances in Emergent Airway Management in Pediatrics. Emerg Med Clin North Am. 2019; 37:473-91.
3. Xue FS, Li HX, Liu YY, et al. Current evidence for the use of C-MAC videolaryngoscope in adult airway management: a review of the literature. Ther Clin Risk Manag 2017; 13:831-41.
4. Sakles JC, Mosier JM, Patanwala AE, et al. The Utility of the C-MAC as a Direct Laryngoscope for Intubation in the Emergency Department. J Emerg Med 2016; 51:349-57.
5. Crewdson K, Lockey D, Voelckel W, et al. Best practice advice on pre-hospital emergency anaesthesia & advanced airway management. Scand J Trauma Resusc Emerg Med 2019; 27:6.
6. Amalric M, Larcher R, Brunot V, et al. Impact of Videolaryngoscopy Expertise on First-Attempt Intubation Success in Critically Ill Patients. Crit Care Med. 2020; 48(10):e889-e96.
7. Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol 2007; 165:710-8.
We read with great interest the study by Goodacre et al. [1], where post-exertion oxygen saturation adds modest prognostic information to clinical assessment of suspected COVID-19 in the ED.
This is partially in contrast with the findings of our study [2], performed within the Fenice Network (Italian group for clinical research in Emergency Medicine). In consecutive ED patients with SpO2 ≥ 95%, the most promising version of a standardised quick walk test (QWT) yielded a sensitivity of 83.3% (95% confidence interval [CI] = 35.9% to 99.6%), and specificity of 93.4% (95% CI = 91.5% to 95.0%). The positive and negative predictive values (PPVs and NPVs) were 8.6% (95% CI = 2.9% to 19.0%) and 99.9% (95% CI = 99.3% to 100.0%).
This difference is probably due to 3 factors:
1. The test standardisation. In our case series, the test was standardised for all centres and consists of a 30-40 metres walk at the maximun possible speed for each patient. Conversely, in the study by Goodacre et al., exertion SpO2 was either recorded after different, not standardised, intentional tests or could have made opportunistically after a spontaneous patient’s effort. The latter condition is particularly worrisome for the purpose of a thorough assessment of the prognostic value of the post-exertion oxygen saturation.
2. The outcome definition. Goodacre et al. considered patients who died or required respiratory, cardiovascular or renal support within 30 days after initial p...
We read with great interest the study by Goodacre et al. [1], where post-exertion oxygen saturation adds modest prognostic information to clinical assessment of suspected COVID-19 in the ED.
This is partially in contrast with the findings of our study [2], performed within the Fenice Network (Italian group for clinical research in Emergency Medicine). In consecutive ED patients with SpO2 ≥ 95%, the most promising version of a standardised quick walk test (QWT) yielded a sensitivity of 83.3% (95% confidence interval [CI] = 35.9% to 99.6%), and specificity of 93.4% (95% CI = 91.5% to 95.0%). The positive and negative predictive values (PPVs and NPVs) were 8.6% (95% CI = 2.9% to 19.0%) and 99.9% (95% CI = 99.3% to 100.0%).
This difference is probably due to 3 factors:
1. The test standardisation. In our case series, the test was standardised for all centres and consists of a 30-40 metres walk at the maximun possible speed for each patient. Conversely, in the study by Goodacre et al., exertion SpO2 was either recorded after different, not standardised, intentional tests or could have made opportunistically after a spontaneous patient’s effort. The latter condition is particularly worrisome for the purpose of a thorough assessment of the prognostic value of the post-exertion oxygen saturation.
2. The outcome definition. Goodacre et al. considered patients who died or required respiratory, cardiovascular or renal support within 30 days after initial presentation as having an adverse outcome. We think that such a definition has two problems. On the one hand, the post-exertion oxygen saturation depends just on pulmonary dysfunction, which is not necessarily related to the dysfunction of other organs or to death. Using a combined outcome may introduce a bias. On the other hand, waiting as many as 30 days to assess the occurrence of the outcome may lead to consider events that are not related to the condition of the lung at the time of initial presentation. For these reasons, we considered the need for invasive mechanical ventilation within 15 days from the initial presentation as the outcome to be predicted by the QWT.
3. The population studied. Goodacre et al. studied a more heterogeneous and probably sicker than ours, since we enrolled only patients with a rest SpO2 ≥ 95% on room air. This is indeed an important point, as the main target of the QWT is to identify patients with a minor pulmonary dysfunction that yields a normal saturation at rest but a decompensation after mild exertion. Of note, in the study of Goodacre et al., about 50% of the patients who experienced the adverse outcome had a baseline saturation <95% and would have been excluded from our study. Indeed, when the authors performed a secondary analysis limited to more appropriate cases, improved sensitivity, specificity and discriminant value were found.
We believe that, albeit further prospective studies are essential to confirm our findings [3], a standardised QWT is promising, because it can be performed rapidly, without specialized equipment and by nonmedical staff, and may have the potential to reliably identify patients who can be safely discharged home or hospitalized in low‐intensive regimens, after the ED visit.
References
1. Goodacre, S., Thomas, B., Lee, E., Sutton, L., Loban, A., Waterhouse, S., Simmonds, R., Biggs, K., Marincowitz, C., Schutter, J., Connelly, S., Sheldon, E., Hall, J., Young, E., Bentley, A., Challen, K., Fitzsimmons, C., Harris, T., Lecky, F., Lee, A., Maconochie, I., & Walter, D. (2020). Post-exertion oxygen saturation as a prognostic factor for adverse outcome in patients attending the emergency department with suspected COVID-19: a substudy of the PRIEST observational cohort study. Emergency Medicine Journal, Dec 3: emermed-2020-210528. doi: 10.1136/emermed-2020-210528.
2. Paglia, S., Nattino, G., Occhipinti, F., Sala, L., Targetti, E., Cortellaro, F., Cosentini, R., Costantino, G., Fichtner, F., Mancarella, M., Marinaro, C., Sorlini, C., Bertolini, G., & Fenice Network (Italian group for clinical research in Emergency Medicine) (2020). The Quick Walk Test: A Noninvasive Test to Assess the Risk of Mechanical Ventilation During COVID-19 Outbreaks. Academic Emergency Medicine, Advance online publication: 10.1111/acem.14180. doi: https://doi.org/10.1111/acem.14180
3. Kalin, A., Javid, B., Knight, M., Inada-Kim, M., & Greenhalgh, T. What is the Efficacy and Safety of Rapid Exercise Tests for Exertional Desaturation in Covid-19: A Rapid Systematic Review, 17 November 2020, PREPRINT (Version 1) available at Research Square [https://doi.org/10.21203/rs.3.rs-105883/v1].
The following quote from the article is incorrect and misses the basic definition of power: "Strictly speaking 'power' refers to the number of patients required to avoid a type II error in a comparative study. Sample size estimation is a more encompassing term that looks at more than just the type II error and is applicable to all types of studies. In common parlance the terms are used interchangeably."
"Power" is the probability that the test correctly rejects the null hypothesis H0 when a specific alternative hypothesis H1 is true. It's equal to 1 - type II error probability. "Power" and "sample size" are not the same thing and they are not used interchangeably. It's possible to derive power given sample size, or calculate sample size based on desired power.
Please correct the article as it'll be highly misleading to beginners.
In an observational study where 200 participants were black, 269 asian, and 4330 white, the authors demonstrated an inverse association between blood pressure and pulse oximetry accuracy that was not influenced by ethnicity[1]. In that study no specific mention was made of the degree of pigmentation in individual members of the ethnic subgroups, presumable because self-reported ethnicity was accepted as a surrogate for skin colour. This acceptance is in sharp contrast with the methodology in the study where subjects of African-American descent were further characterised by a description of their degree of pigmentation, using terminology such as "very darkly pigmented".. This was one of the earliest prospective studies conclusively to show that some oximeters overestimate arterial oxygen saturation in hypoxic subjects who are "darkly pigmented" [2].
Show MoreIn retrospective studies such as the ones subsequently undertaken to explore the theme of racial bias in oximetry it was easy to fall into the trap of using ethnicity as a surrogate for skin colour[3],[4], largely because skin colour is not consistently recorded as part of the medical record[3]. Explicit description of skin colour also gets omitted when race and ethnicity are defined using self-reported demographic data[4].
Future studies, however, might seek to ascertain whether or not skin pigmentation compounds the overestimation of oxygen saturation attributable to hypotension....
We thank Dr. Delaney and colleagues for their valuable research into the concept of midline cervical tenderness. Unlike the NEXUS critiera, the Canadian C-Spine Rule does not use midline tenderness as a positive indication for imaging. Our original study in JAMA 2001 found that assessment of this criterion amongst alert trauma patients at risk of c-spine tenderness had excellent interobserver agreement between ED physicians with a kappa of 0.78. We found that absence of midline tenderness was a good negative predictor of c-spine injury but that presence of of such tenderness was non-specific and not useful. Hence, absence of midline tenderness is considered a low-risk factor. Our NEJM 2003 validation study found that the CCR had both better sensitivity and specificity than NEXUS.
Best regards
1. Stiell IG, Wells GA, Vandemheen K, Clement C, Lesiuk H, De Maio VJ et al. The Canadian Cervical Spine Radiography Rule for alert and stable trauma patients. JAMA 2001; 286(15):1841-1848.
2. Stiell IG, Clement C, McKnight RD, Brison R, Schull MJ, Rowe BH et al. The Canadian C-spine Rule versus the NEXUS low-risk criteria in patients with trauma. N Engl J Med 2003; 349:2510-2518.
A well timed project with the contemporary interest in the subject of drug misuse in Scottish politics, and also with the recent airing of Dopesick on streaming services about the US experience.
I wonder whether the known CYP2D6 polymorphism leading to poor metabolism of codeine has a role in the increase of opioid prescriptions? Mikus and Weiss (2005) state that 5-10% of Caucasian have severely impaired metabolism of codeine and that a further 10-15% show some impairment. This means up to 1 in 4 Caucasians don't get a full analgesic effect from codeine.
Maybe we are seeing a lessening of patients putting up with their lot of suboptimal pain control? And in turn an increase in prescriptions to accommodate that demand.
Dihydrocodeine by contrast doesn't have the same issues with metabolism and ineffective analgesia. While acknowledging the past issues when DF118s were misused, perhaps prescribing dihydrocodeine instead of codeine we'll see better analgesia in our patients and perhaps a reduction in demand for prescriptions?
Mikus G and Weiss J. (2005) 'Influence of CYP2D6 Genetics on Opioid Kinetics, Metabolism and Response', Current Pharmacogenomics, 3, pp43-52
We read with interest your experience of creating care pathways for patients in the out-of-hospital setting during the Covid-19 pandemic, in particular for those with palliative care needs. The benefits of the Physician Response Unit (PRU) being tasked to end of life care related 999 calls, their enhanced level of assessment and management and the resulting reduction in inappropriate hospital admission is to be applauded. There is additionally a need to recognise the wider use of such care pathways, available pan London, that support palliative and end of life care (EoLC) patients accessing emergency care.
In collaboration with hospice and palliative care teams the London Ambulance Service NHS Trust (LAS) has to date, created 19 EoLC appropriate care pathways. These offer our clinicians access to specialist advice, support with complex decision making and provide an alternative to Emergency Department (ED) conveyance. In the last 2.5 years we have undertaken an extensive programme to improve EoLC within the LAS; providing tailored education to augment our clinicians’ knowledge and confidence, creating guidance which includes medications and symptom management at the end of life and increasing clinician use of advance care plans. As a result we have seen a 15% increase in staff confidence and an 18% median reduction in ED conveyance for this patient group; most importantly more patients are being cared for in their place of choice and in line with their wishes....
Show MoreDear Sir,
Within their conclusion, Charlton et al recommend further research to understand patient behaviour toward seeking help during the pandemic. In response to this, we would like to highlight the findings of our work which address this. We undertook a mortality review of all deaths in Salford during the peak 7 weeks of the initial pandemic surge (522 deaths), looking at themes which, if addressed could result in reduced mortality in future waves. We reviewed all 111, 999, general practice and hospital contacts for all patients from the 1st March 2020, to ensure all help seeking behaviour and the system response was understood. We have summarised these here.
Show MoreWe noted 60 cases where patients delayed seeking help. We were also concerned that patients who were advised to call 111 by their GP, and were offered advice, would only call back when seriously ill. “COVID phobia” was evident in a small number of cases, at its extreme, including refusal to attend hospital and subsequent death at home.
Fewer than half of NHS 111 calls were answered during the review period. Of those that were answered, 46% resulted in advice to contact their GP. Of these, 5 were subsequently admitted to hospital later the same day. Indeed, despite the national directive to telephone 111 as the first point of contact, only 13% actually did. 81% of patients contacted their GP in the period prior to their death.
However, a lack of early face to face assessments was identi...
To the editor,
In their article “Agreement and predictive value of the Rockwood Clinical Frailty Scale (CFS) at emergency department triage”, Shrier et al (1) nicely illustrate that the level of agreement between different health care professionals in obtaining the CFS at different clinical settings is weak. The CFS was rated for 8,568 patients over 65 years by the triage nurse at the emergency department (ED) and by the attending physician on admission on the ward. Both scores were compared using the Cohen’s kappa coefficient, which was 0.21 and therefore is considered weak (2).
Show MoreWe found similar results in our AmsterGEM study. The AmsterGEM study is an observational prospective cohort study that investigates the prognostic accuracy of frequently used frailty screening instruments (3). Patients aged over 70 years old attending the ED were screened with four screening instruments, including the Safety Management System (VMS) (4). The VMS consists of four geriatric domains, which are associated with functional decline: delirium, malnutrition, falls and ADL limitations (5,6). All Dutch hospitals have implemented the VMS screening instrument for hospitalized older patients. In a sub-analysis of our study, the VMS was applied on 173 hospitalized patients over 70 years old, at the ED by a research student and at admission by the attending nurse on the ward. The average age was 81.2 years old, 81 (47%) were male and 111 s...
Having recently updated our Emergency Department guidelines for suspected PE in pregnancy, we read the secondary analysis of the DiPEP study with great interest.1 However, we were quite surprised at the poor overall D-dimer sensitivity. Only 66% (8/12) of PEs would have been identified based on the recommended positivity threshold of 400ng/ml. This is considerably lower than the pooled estimate of 97% (95% CI 96-98%) found by a recent meta-analysis evaluating D-dimer for PE, and largely explains the poor performance of the YEARS and Geneva algorithms in the DiPEP cohort.2
This result does not seem to fit with the known physiology of pregnancy. We know that D-dimer levels increase throughout pregnancy, which should improve sensitivity and worsen specificity.3 To our knowledge there are no other studies demonstrating impaired sensitivity of D-dimer in pregnant vs. non-pregnant populations.
The DiPEP authors note that most of the study participants had received anticoagulation before blood samples were taken, which can decrease D-dimer levels by up to 25% in the first 24 hours.1 They also note however, that this would be insufficient to explain all their false negative D-dimer results. Aside from random error, we wondered if anything else could explain the poor sensitivity.
One feature of the DiPEP study that stood out to us was the D-dimer assay used. As a microplate ELISA assay, the Zymutest D-dimer should be very sensitive but we could not find any st...
Show MoreTo the Editor
Show MoreWe have read with great interest the recent article of Miller et al1 determining the technical factors associated with first-pass success (FPS) during endotracheal intubation with C-MAC video laryngoscope (VL) in children. They showed that placement of the blade tip into the epiglottic vallecula regardless of blade types, adequate glottic view and locating the glottic opening within second quintile of video displayer were significantly associated with FPS. Given that paediatric airway management is a great challenge to emergency physicians and the benefits of videolaryngoscopy are often significant in airway management of emergency paediatric patients,2 their findings have potentially clinical implications. Other than limitations described by authors in discussion, however, we noted several methodological issues in their article on which we invited authors to comment.
First, primary outcome of this study was FPS, which was defined as passage of C-MAC VL into the mouth with the intention of intubation that terminated with successful intubation at first attempt. As described by authors in introduction, however, C-MAC VL is an intubating device with ability to perform both direct and video laryngoscopy using same device. That is, the larynx can be seen either under direct vision or on a monitor when using C-MAC VL.3 This advantage of C-MAC VL makes it exceptionally useful for emergency intubation. For example, in the event of a failed video laryngosc...
We read with great interest the study by Goodacre et al. [1], where post-exertion oxygen saturation adds modest prognostic information to clinical assessment of suspected COVID-19 in the ED.
This is partially in contrast with the findings of our study [2], performed within the Fenice Network (Italian group for clinical research in Emergency Medicine). In consecutive ED patients with SpO2 ≥ 95%, the most promising version of a standardised quick walk test (QWT) yielded a sensitivity of 83.3% (95% confidence interval [CI] = 35.9% to 99.6%), and specificity of 93.4% (95% CI = 91.5% to 95.0%). The positive and negative predictive values (PPVs and NPVs) were 8.6% (95% CI = 2.9% to 19.0%) and 99.9% (95% CI = 99.3% to 100.0%).
This difference is probably due to 3 factors:
1. The test standardisation. In our case series, the test was standardised for all centres and consists of a 30-40 metres walk at the maximun possible speed for each patient. Conversely, in the study by Goodacre et al., exertion SpO2 was either recorded after different, not standardised, intentional tests or could have made opportunistically after a spontaneous patient’s effort. The latter condition is particularly worrisome for the purpose of a thorough assessment of the prognostic value of the post-exertion oxygen saturation.
2. The outcome definition. Goodacre et al. considered patients who died or required respiratory, cardiovascular or renal support within 30 days after initial p...
Show MorePages