https://orcid.org/0000-0002-8408-8358
Peer Review History
| Original SubmissionJune 27, 2020 |
|---|
|
PONE-D-20-17256 Detection dogs as a help in the detection of COVID-19 Can the dog alert on COVID-19 positive persons by sniffing axillary sweat samples ? Proof-of concept study PLOS ONE Dear Dr. grandjean, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. As you will read the 3 treviewers raised relatively similar methodological options and my opinion is converging in this direction. To summarize as suggested by one of the reviewer you have two options: - Additional data is added, providing the results of a double-blind test using samples from new people that have not previously been screened by the dogs, or - The current abstract, discussion and conclusions are revised to provide a more critical appraisal of the results, covering possible limitations of the study and whether there are any other possible explanations for the dogs' detection performance and possible sources of unintentional bias. It seems also clear that your manuscript would benefit from being corrected by a native english speaker. Please submit your revised manuscript by Aug 27 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript:
If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols We look forward to receiving your revised manuscript. Kind regards, Nadine Ravel Academic Editor PLOS ONE Journal Requirements: When submitting your revision, we need you to address these additional requirements. 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and 2. Thank you for stating the following in the Financial Disclosure section: "The authors received no specific funding for this work" We note that one or more of the authors are employed by a commercial company: "DiagNose, Cynopro Detection Dogs and Biodesive SAS" a) Please provide an amended Funding Statement declaring this commercial affiliation, as well as a statement regarding the Role of Funders in your study. If the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript and only provided financial support in the form of authors' salaries and/or research materials, please review your statements relating to the author contributions, and ensure you have specifically and accurately indicated the role(s) that these authors had in your study. You can update author roles in the Author Contributions section of the online submission form. Please also include the following statement within your amended Funding Statement. “The funder provided support in the form of salaries for authors [insert relevant initials], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.” If your commercial affiliation did play a role in your study, please state and explain this role within your updated Funding Statement. b) Please also provide an updated Competing Interests Statement declaring this commercial affiliation along with any other relevant declarations relating to employment, consultancy, patents, products in development, or marketed products, etc. Within your Competing Interests Statement, please confirm that this commercial affiliation does not alter your adherence to all PLOS ONE policies on sharing data and materials by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests) . If this adherence statement is not accurate and there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared. Please include both an updated Funding Statement and Competing Interests Statement in your cover letter. We will change the online submission form on your behalf. Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Partly Reviewer #2: Partly Reviewer #3: Partly ********** 2. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: No Reviewer #2: No Reviewer #3: I Don't Know ********** 3. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: No Reviewer #2: No Reviewer #3: No ********** 4. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: No Reviewer #2: Yes Reviewer #3: No ********** 5. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: This article investigates whether dogs can differentiate the axillary odor of individuals infected by SARS-CoV-2 from individuals that are not infected. The study is very well justified and potential applications for the management of the current pandemic (especially in case of a feared second wave) are highly relevant. I however have some major issues with this study, as detailed below. 1) Design. With the current information provided in the method, the reader cannot be convinced that the authors controlled for a number of confounding factors. Namely, it should be made clear that the negative and positive samples cannot be different from each other based on other factors than infection to SARS-CoV-2. It is known that body odor changes according to medication intake, gender, age, health status, hormonal status etc. With this in mind, a number of questions can be raised here. Who are the COVID-negative patients: do they have other pathologies? If yes, which ones? Do they all have the same pathology, or different ones? Are they on medication? Are the COVID+ on medication and if yes, is it the same one for everybody? How are the negative samples chosen with regards to gender and age, according to the target (COVID+ sample) characteristics? Two odor sampling methods are mentioned (gauze swabs or filters and polymer tubes): Are they both used for negative and positive samples, and is the same sampling support used in the whole line during a trial? In sum, more information is expected in the method section on these aspects; the limits related to the possible confounding factors must be discussed in more details than it currently is; and the conclusion will probably need to be more cautious (as a consequence). 2) Statistical analyses. The authors mention accuracy of the dog’s responses in introduction and conclusion, in terms of sensitivity and specificity. It should be clearly defined, especially with regards of the tests performed in the present article. The signal detection theory is evoked only in discussion, and false positives are discussed in that same section. It is therefore expected to see the current data analyzed using these criteria, which is not the case. 3) Paragraph about human pheromones in the discussion (lines 359-382). I suggest to simply remove this paragraph, which is 1) erroneous (no pheromones have ever been identified in humans, and the androgenous compounds you are referring to are male – not female – compounds supposed to attract females), 2) very speculative (how did you estimate the fertile phase of the donor menstrual cycle?), 3) disproportionally long compared to other points to discuss that are more crucial and more directly related to the main question. Minor points. 1) Please describe what the mocks are, and how many are there in the line. 2) Please better explain how the training was conducted with regards to sample type: are there negatives at this stage? Mocks? 3) Please explain body odor sample management. From what I understand, they are used fresh (minimum 24 hrs between odor collection and training). Are they thrown away after 3 trials? Are they frozen at any time in the process? Refrozen? Finally, on a more formal aspect, the article should be proofread for English. Also, improvement of the writing would be welcome. Some sentences are difficult to understand (i.e., lines 237-240 / 395-398) and some others do not integrate well in the chain of arguments (i.e. lines 62-64 / 343-345). Most importantly, in its current form the article is very fractionated (“bullet points” feeling in the whole document): It would very much gain in fluidity and articulation of the ideas if the authors work on this aspect. Reviewer #2: Review of the manuscript PONE-D-20-17256 The idea of using trained dogs for COVID-19 screening was boosted by mass media in many countries e.g. by the BBC in UK, and by announcing by many researchers and dog trainers, that they intend to train dogs to detect COVID-19. It is obvious that first published experimental works in this field, will have a great potential for being broadly cited or comercialized. However, it could be presumed that such experiments and manuscripts are probably accomplished under a strong pressure of time, to be published as the first. Therefore the papers like the present one, which, to my knowledge is the first paper on this topic, should be reviewed and evaluated particularly causciously and critically. It is important not to disseminate facts that are not based on sound scientific experiments. This is particularly important since the reviewed paper is already available online at https://www.biorxiv.org/content/10.1101/2020.06.03.132134v1.full.pdf as not certified by peer review. For justification of their study the authors cited publications on cancer detecting dogs and dogs alerting hypoglycaemia and seisure in human patients. However, it should be mentioned that despite of about 20 publications within the last 17 years, the medical detection dogs have not been applied for practical clinical screening so far. This was evident on the latest international conference on cancer detection by dogs, organized by the Curie-Institute in Paris in October 2019. Also, despite of numerous chemical studies within more than 25 years, using gas chromatography / mass spectrometry, which found some differences in volatile organic compounds (VOCs) between cancer positive and cancer negative patients, no single VOC or a simple combination of VOCs were identified, that could be reliable cancer marker(s), mature for practical cancer screening. The reason for problematic application of methods based on identivication of VOCs for disease screening, either by chemical analyses or by trained animals, is a great variability of emited VOCs due to day-to-day differences in diet, medication and numerous other confounding factors. The aim of the reviewed study was to assess if dogs could be trained to discriminate body odor of patients infected with SARS-Cov-2, from body odor of healthy controls. Such aim is justified by the current and probably future situation with COVID-19, untill an effective and commonly accessible COVID-19 vaccine is available. The experiments were conducted at 3 different locations, with a total of 18 dogs which were previously trained either for explosives detection or for search and rescue or for cancer detection. The Authors collected altogether 101 sweat samples from patients showing clinical symptoms of COVID-19, being also COVID-19 positive on RT-PCR or PCR test for SARS-CoV-2. Negative samples (n=97) were collected the same way, by the same trained staffs, from patients who have no clinical symptoms related to COVID-19, and were negative on SARS-CoV-2 PCR test. The main flaw of collecting sweat samples for this study was, that the samples were not controlled quantitatively for the amount of the sweat. It is known that one of the typical COVID symptoms is a high fever. Humans with fever sweat more intensively than healthy people, so the amount of the sweat on the sample may be a clear cue for dogs. As an example, our early studies in the eighties, on dogs detecting so called „silent estrus” in cows, could be mentioned. We used cotton swabs for collection vaginal mucus from cows in estrus vs diestrus. Our dogs perfectly indicated estrus samples. Only later we noticed that cows in estrus usually secreted more mucus and this was a cue for the dogs. When the dogs became more „experienced” they perfectly indicated cotton swabs moistured even with water. This made me causcious and critical about dog’s reliability. For dogs, odors like cow’s in estrus, human cancer, drugs or explosives, or SARS-CoV-2, play no biological role and such odors are indicated by dogs only because an operant conditioned response has been trained by producing association between odors and a reward in form a treat or favorite object to be retrieved and play with. Therefore, the trained dogs, especially when they are more „experienced”, will seek for a cue or a simple solution, to earn a reward. Such strategy of dogs may produce frequent false alerts. At identification of human scents by dogs it should be remembered that samples of human scent contain both individual component and other components as those related to diet, disease(s), medication, hygiene, cosmetics, background odors of the location as well as own odors of the sampling materials (odor carriers). In lines 346-349 the authors report on two samples that were negative according to inclusion criteria for negative samples, but were indicated by two dogs as positive and the PCR re-test showed that in fact they were positive. This, however, is not a convincing evidence that the dogs sniffed out specific molecules induced by the SARS-CoV-2 virus presence. These two persons were not controlled for the amount of the sweat or body temperature measured precisely at the time of sample collection. Without knowing if the odor samples differ quantitatively in the amount of the sweat, it is not clear whether the dogs indicated a hypothetic, characteristic odor emited by COVID positive patients, or rather indicated samples from people with fever, who sweat more. The Authors mentioned that the odor samples were collected „ at the same location”. This is methodologically sound but it should be precised what does it means „the same location” in this context ? Was it the same room or the same buiding or the same district ? The authors must be more precise here, because it is known that the dogs might be conditioned on the characteristic „hospital odors”, as shown by Walczak et al.(2012) in cancer detecting dogs, when the cancer positive samples were collected mostly in hospitals. Another problem with COVID-positive sweat samples is that positive patients are staying either in infectious hospitals or in isolation wards that are probably intensively desinfected. In addition, the access of healthy COVID-negative healthy donors (controls) to the same room where sick patients are staying, is hazardeous and strictly limited. This may be a serious constraint for collecting samples for further studies by other researchers. The description of the training methods only in 3 lines is very scarce. Were the dogs rewarded in every training and testing trial and who and how gave a signal that the dog’s indication was correct ? Lines 219-220 - It is true that search and rescue dogs (SRD), trained to perform disaster and area search, mainly work through the scent of the sweat. However, it must be taken into consideration that SRD are trained for detection of a generalized human odor, without discriminating individual persons and persons’ specific odor e.g. odor(s) related to a disease. Such dogs are relatively easy to be re-trained to work in a scent lineup to discriminate odours of particular person, but a re-training, to be reliable, takes longer than 1-4 hours and 4-10 trials. In lines 262-264 the Authors stated: „…For each involved dog, the acquisition of the specific odour of COVIS-19+ sweat samples required from one to four hours, with an amount of positive samples sniffing ranging from four to ten….” I am skeptical about the ability of each dog to learn precisely the task within such a short time. For example for cancer detecting dogs, the training period according to 13 papers published in the period 2003-2014, ranged from several weeks (Pickel et al. 2004) to 16 months (Cornu et al.2010), and on average took 8 months. Even if the Authors could observe that their dogs correctly indicated the target samples after one hour, performing only four trials from being confronted for the first time with COVID sweat samples, it seems inredible that dogs could acquire the ability to detect COVID odor so quickly. It can not be excluded that the dogs indicated samples in the lineup, that are merely different from the others e,g, by containing more sweat. In lines 265-267 or in lines 294-296 the Authors did not precise what was a criterion (percentage of correct indications, false alerts, misses, regularity of sniffing the samples in the lineup ?) for considering the dog to be fully trained and ready for testing. From the lines 294-296 it could be concluded that separate samples were used for the training and testing, which is methodically correct. However, there is no information how many positive and negative or mock samples, out of the totals given in Table 1, were used for the trainig and real testing. The authors did not conduct so called „zero trial” with no target samples in the lineup. This is another flaw which could have a consequence by making the overall conclusion in lines 407-408 premature. Although negative and mock samples were used in the lineup (lines 269-270), but there is no explanation what kinds of mocks were used. Were they samples with more sweat ? or samples from donors with fever but not related to COVID-19 (for example from people with seasonal influenza) ? „Zero trials” usually provoke more false alerts, especially on samples that are simple different from the others. Minor remarks: The manuscript shoud be better structured. For example the major portion of the Conclusion is appropriate rather for the Discussion section, whereas the real conclusion is given as the last sentence or the Discussion (in lines 407-408) Line 64 More original papers should be cited instead of a handbook. Lines 76-78 „…Scent detection by animals has been published in a large number of diagnostic studies, which all suggest similar or even superior accuracy compared with standard diagnostic methods…” – nevertheless, the scent detection by animals is not used for medical diagnoses or screenings on a large scale (probably except for tuberculosis detecting Gambian pouched rats in Africa). The medical diagnosis of a disease should not be confused with screening. Scent detection by animals will never be a diagnosis. Line 83 „…The first clinical investigation of cancer was published by Willis on bladder cancer in 2010, after having published a proof of principle study in 2004…” – this information is only partly true: McCulloch et al. published their study on canine detection of lung and breast cancer in 2006. Lines 88-89 „….The dog’s nose is also now currently utilised in order to prevent the consequences of crisis (or even prevent them) for diabetic [30, 31, 32, 33] and epileptic people [70]….” - The accuracy and reliability of hypoglycaemia and seizure alerting dogs, when considering on a daily basis, is still questionable. The alerting dogs rather contribute to feeling more safe and thus to improving the quality of life of the patients, than to precise alerting the hypoglycaemia or seizure episodes. Lines 115-117 „….In a second time……” – this assumption is not quite clearly formulated in English. Lines 154-158 Why two types of sampling supports are used ?: (1) - Sterile gauze swabs utilised by the hospitals or sterile gauze filters utilised for explosive detection, and (2)- Inert polymers tubes utilised for explosives, drugs or criminology detection) Line 289 Why acetone was used for disinfection of sample carriers ? Acetone is an irritant, causing mild skin irritation and moderate to severe eye irritation which may be repelling for dogs. Lines 292-297 „…A failure was defined if the dog alerts on any of the other boxes. Trials were registered from the moment dogs recognized positive sample among other samples (mock or negative ones) without having sniffed any of them previously to the test. If the dog marks first a negative sample, the trial is considered as failed, even if he marks the positive afterwards….” - the authors did not precise whether the dogs have to sniff ALL samples prior to the indicatication. Have the dogs to sniff the remaining samples after the correct indication? If so, would the trial be declared as a failure, if the dog made a false alert AFTER prior correct indication in the same trial? . This makes that the calculation of correct indication by chance should take into account how many samples a dog actually has sniffed. Therefore, when the dog indicates correctly for example on the first sample in the lineup and does not sniff the remaining samples, the probability of correct indication by chance at 7 samples in a lineup is not 14% but 50%. This would mean that the probability of correct indication by chance is varying from trial to trial ? Line 350 „….dogs have no reason to lie….” – this is too colloquially expressed. With regards to odor detecting dogs one should not speak about "lying" or "telling truth". From the dog's perspective, detecting trained odors as for example the hypothetical COVID-19 odor, that have no biological relevance, is a rather a game to earn a reward. Dogs, even if tey are well trained, when they are getting more routined, will try to get a reward at a simple way, which means that they will try to find any additional cues to indicate the target odor in the lineup. Therefore the dogs' performance in the scent lineup has to be permamently controlled throughout the whole deployment period, because it may vary. This seems to be one of the reason why so called medical detection dogs have not been used in medical screening on a regular basis. Table 3. The terms detection sensitivity and specificity, which derive from the point 1-4 in lines 385-388, should be given, or percentage of false alerts should be given instead of the column with the number of correct indications, which is redundant in this table The English of the manuscript is for me as a non-native speaker understandable but suffers from many language and wording imperfections. It should be edited by a native English speaker, who also has some knowledge in the specific study field. Line 48 „..State of play..” ? – rather „State-of-the-art.” Line 51 What does it means „..for those of interest in dog’s olfaction…” in this context? Line 148. „…In order to avoid any dog creancement (?) on the olfactive “background noise In several places of the manuscript there are statements that are difficult to follow In lines 115-117 the authors write: „…In a second time, we attempt to validate the method with the dogs checking samples one by one instead of comparatively, in order to fit with what could be an operational utilisation of such trained dogs...” – what does it means exactly Line 372 „ In Sheakspeare’s day…” ? - rather is „ in Sheakspeare’s era” or „in Sheakspeare’s time” Lines 372-374 the passage with the anecdote on peeled apple under woman’s arm should be deleted because it is too colloquial and not closely related to the reviewed study. Line 404 „ …cynotechnicians…” ? or dog handlers ? Line 455 „…by small…” ? – should be „ by smell” In conclusion, this first report on dogs trained to detect COVID-19 is interesting but a flaw in the methodology of collecting scent samples (no control for the amount of the sweat) makes that the conclusion that „…there is a high evidence that dogs can detect a person infected by the virus responsible for COVID-19 disease …” seems to be premature. In addition, some language imperfections make me to evaluate the manuscript as not acceptable for publication in PloS ONE in the present form and needing a major revision before resubmission. First of all, in the major revision the issue of the amount of sweat in odor samples should be addressed, including replication of the trials. Reviewer #3: General comments Thank you for conducting and reporting this very interesting study. In my opinion, this study is a good first step towards testing whether dogs can detect odours associated with SARS-CoV-2 infection. You have used a reasonable sample size of dogs and SARS-CoV-2 samples and the methods are generally appropriate. A main difficulty with any study of this type, where it is not known whether the odour the dogs are being trained to detect exists (in this case SARS-CoV-2 related), is demonstrating beyond reasonable doubt that dogs are detecting the odour rather than incidental contamination, and demonstrating robustly that the research team is not unintentionally influencing/cuing the dogs. Sources of odour contamination may include, for example, medication used in SARS-CoV-2 patients, hospital odours, differences in the odour of the storage containers of positive and negative samples and different odours arising from how positive and negative samples are handled. You have taken some steps to try to control for these issues. However, in my opinion, it is not possible to conclusively say that the sweat of SARS-CoV-2 contaminated persons can be detected by trained dogs based on the methods and results you present so far. In places, the methodology and controls you used just need further clarification in order to rule out possible sources of bias and alternative explanations that the dogs may be detecting contaminating odours associated with the SARS-CoV-2 samples. Most importantly, to support the conclusion that the sweat of SARS-CoV-2 contaminated persons can be detected by trained dogs, a more rigourous, double-blind test is required, in which the research team is provided with new samples from an independent source which they are asked to assess using the dogs. The dog handlers, sample handlers and data recorders (everyone involved in assessing the samples) should be blind to the sample identities (positive or negative) until the dogs have conducted the test. Methods for how to do this have been presented in a number of previous medical detection dog studies. I would suggest two options for revising the paper: - Additional data is added, providing the results of a double-blind test using samples from new people that have not previously been screened by the dogs, or - The current abstract, discussion and conclusions are revised to provide a more critical appraisal of the results, covering possible limitations of the study and whether there are any other possible explanations for the dogs' detection performance and possible sources of unintentional bias. Potential sources of bias/limitations in medical detection dogs studies have been summarised previously (e.g. Elliker et al 2014. Key considerations for the experimental training and evaluation of cancer odour detection dogs; Edwards et al 2017 Animal olfactory detection of human diseases: Guidelines and systematic review; Johnen D, 2017. An approach to identify bias in scent detection dog testing). I recommend you review these guidelines and report where your study meets these criteria, and/or discuss limitations where your study does not meet these criteria. The entire manuscript, including references, needs proof reading and correction for typographical errors, spelling, grammar, and consistency of tense and language. Detailed comments Abstract - For the acquisition phase it would be useful to clarify how many “runs” dogs required during the one to four hours of training. Was it just a single presentation of each sample or were they presented multiple times during? - Please clarify the wording around how many dogs were used for data collection in the study (it currently confusing for the reader whether 8 or 18 dogs were used for data collection). - As per my general comments, I feel the conclusion that the sweat of SARS-CoV-2 contaminated persons can be detected by trained dogs is too strong at present, as the study does not appear to provide a robust double blind assessment and possible sources of bias/limitations have not been discussed. The study appears to be a good first step, but there are other possible explanations for the results based on the data presented. I suggest the conclusion is caveated that this study is a promising first step, and that a double blind trials is needed to more robustly confirm the results (unless this data is available and can be presented). Medical detection dogs: state of play - To avoid bias towards only discussing positive studies and to provide a more balanced discussion, I suggest mentioning some of the difficulties of conducting these studies where it is unknown whether the target odour exists, and that not all studies to date have shown disease odours to be reliable detectable by dogs. Issues are summarised in papers such as Elliker et al 2014, Edwards et al 2017, Johnen D, 2017. Material and Methods Samples - Please clarify whether these 198 samples were each from individual people (i.e. 198 different people), or whether multiple samples were taken from the same person. Please could you clarify which samples were used for training and which were used for the trial stage (if training samples were reused for the trial stage this is a limitation that needs mentioning). Practical realisation and storage of samples - Please clarify if there is any risk that positive and negative samples could have been contaminated with different odours due to the way they were collected. For example, you state positive samples were collected with the person wearing full COVID-19 safety equipment – could this have transferred odour to the positive samples and not the negatives? Were negative samples also disinfected - if not, it is possible that dogs were training on the odour of disinfectant on the positives? Canine Ressources (note spelling mistake in title) The dogs - The photos of dogs are nice but are not required in a scientific paper – suggest delete. Training and first testing of the dogs - Lines 262 to 263. The statement that the acquisition of the specific odour of COVIS-19+ sweat samples required from one to four hours ideally needs further supporting details on the training methods used and statistical evidence that the discrimination had been learnt (e.g. what was the learning criterion, were the handlers blind, were blank/negative samples also placed out, how many presentations did each dog require?). Please clarify if the learning criterion was based on the subjective opinion of the handlers rather than a statistical criterion. - This seems to be extremely rapid acquisition (some medical detection dog studies have required weeks or months of training) so a comment on this would be useful in the discussion section. - Were the training samples different to those used for actual tests? If not, please include as a study limitation as the dogs could have learnt the specific odours of the training samples rather than SARS-CoV-2 generally. Testing protocol - Lines 282-283. Repeated use of the same sample could be considered "pseudoreplication". Please clarify if this was taken into account during the data analysis (i.e. the 3 samples should not be considered as independent trials if the dog had 3 attempts to detect the same sample - it’s 3 attempts at a single "diagnosis"). Ideally each sample, positive or negative, should be used only once in the test. If they were used more than once this need to be discussed as an important study limitation, as the dogs may have learnt individual samples rather than generalising based on a SARS-CoV-2 odour. - Lines 290-291. Please comment on whether there is any risk that the person putting out the samples could have unintentionally influenced the results by handling or storing the positive and negative samples differently. The dogs may learn minor differences in packaging odour, storage location or handling (e.g. positives and negatives stored in different fridges may carry different environmental odours that the dogs learn). - The ideal experimental design would be to run the trial double blind (the person putting out the samples does not know if they are positive or negative during the trial). If it was not double blind this needs to be mentioned as a study limitation. - Lines 292-293. Please comment on whether the individuals collecting data were blind to the sample allocations and whether they could have biased the results (ideally they should be blind too). Lines 295-296. This sentence suggests that dogs had not previously been exposed to the samples used in the trials – please clarify if this is correct. Does this mean new samples were from new individuals not previously assessed by the dogs? Results and statistical analysis - Line 313. As above, if samples have been reused several times for the same dog this may be considered psuedoreplication of data. Please clarify if this was taken into account during the statistical analysis. Reuse of samples/individuals may result in detection based on learning of individual odours rather than COVID odour generalisation (it is not impossible that dogs could potentially learn 33 different individual samples) Table 3: Results obtained on the dogs that finished the training period (n=8) - Please state how many different positive samples (from different people) each dog was presented. It is not clear whether samples collected from the same individuals were reused more than once per dog. If they were reused, this needs to be included as a study limitation. - Ideally it would be best to provide your underlying data (breakdown of trials for each dog showing positive and negative samples put out and response of the dog) as a supplementary file/download. Discussion - As per my general comments above, I recommend that the discussion is revised to highlight that this study a promising first step, but you include a more critical discussion of the possible limitations and possible alternative explanations for the dogs’ performance. I think the current conclusion (lines 331-332) is too strong based on the methods and data presented, without a more robust double-blind and independently validated trial taking place. - Lines 359-382. Although interesting, I feel that the discussion of female hormone detection is moving too far away from the focus of the study subject and would be better deleted from the discussion (the evidence for the dogs indicating these samples due to female pheromones is also weak/anecdotal). Conclusion - As per my general comments above, I think the current conclusion (lines 417-419) is too strong based on the methods and data presented, without a more robust double-blind and independently validated trial taking place – suggest revision of wording in line with my general comments above. Figures - Figures 3. 4 and 11 are not readable due to poor resolution. - Suggest delete figures 5 and 6. ********** 6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No Reviewer #2: No Reviewer #3: No [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. |
| Revision 1 |
|
PONE-D-20-17256R1 Can the detection dog alert on COVID-19 positive persons by sniffing axillary sweat samples? A proof-of concept study PLOS ONE Dear Dr. grandjean, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Two reviewers ar still requiring some clarification or some revision of your conclusions and importantly recognize some indications of the limits of your study. Please submit your revised manuscript by Nov 13 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript:
If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols We look forward to receiving your revised manuscript. Kind regards, Nadine Ravel Academic Editor PLOS ONE [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation. Reviewer #1: (No Response) Reviewer #2: All comments have been addressed Reviewer #3: (No Response) ********** 2. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Partly Reviewer #2: Yes Reviewer #3: Partly ********** 3. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: I Don't Know Reviewer #2: Yes Reviewer #3: No ********** 4. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: No Reviewer #2: Yes Reviewer #3: Yes ********** 5. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: No Reviewer #2: Yes Reviewer #3: Yes ********** 6. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: The authors have considerably improved the manuscript, by providing significant information that was lacking in the previous version, and by improving the structure of the manuscript. For this kind of study, and especially due to the potential impact its results may have on the current management of the Covid-19 disease, it remains essential that the methodology is carefully discussed, so that the validity of the conclusions can be properly evaluated. In the current version, the additional details provided have elicited further questions about the validity of the conclusions: 1) The pairing between COVID-19 positive and COVID-19 negative odor donors represents a strong limitation in this study. Whereas the Lebanese groups are comparable, age differences between both groups in Paris is considerable: as body odor differs with age (to human noses as well as in terms of chemical composition: Mitro et al 2012 “The smell of age: perception and discrimination of body odors of different ages”, Chen & Haviland-Jones 2000 “Human olfactory communication of emotion”, Haze et al 2001 “2-Nonenal newly found in human body odor tends to increase with aging”), this may interfere with the dogs’ discrimination. The results suggest that it may introduce noise that alters the dog’s performance, since French dogs had lower success rates than Lebanese dogs. In future studies, targets and controls should at least be paired for sex and age. Another even more important limit is the fact that COVID-19 negative donors were partly people hospitalized for another reason than COVID-19. Meaning that they probably are on medication, and on medication that differs from the one taken by COVID-19 positive patients. How it may impair dogs’ performance or bias there responses towards the COVID-19 positive sample is currently difficult to predict in the absence of further information regarding patients’ treatment. 2) Mocks. The authors now explain that there were 3 or 4 cones in the lineups, including one COVID-19 positive sample and zero, one or two mocks. First, I assume this was zero or one for the 3-cones lineups, and zero to two mocks for the 4-cones lineups, i.e. there was always at least one COVID-19 negative sample in the lineup. Please confirm and write it explicitly in the text. Second, what was the rationale for using mocks? This is potentially a problem, since the test is much easier for the dog when they have 2 mocks plus 2 body odors (i.e., one positive and only one negative sample): they actually are discriminating between 2 and not 4 samples (the level of successfully answering by chance is much higher), since they were trained before to discriminate sweat samples from mocks (step 3 of the training). 3) Where was the dogs’ testing performed? Why did one dog have only 3 cones, and not 4 as the others? 4) Repetition of sample presentations during testing may also be an issue, as raised by the other reviewers. The authors response to this limit was not convincing: rather, they could test whether the performances increased on the second or third repetition compared to the first presentation. In fact, the results section is very short: the authors could very well test some of the possible bias arising from methodological flaws and present them here. For example, in addition to testing the effect of sample repetition, I suggest that they i) test whether lineups including only COVID-19 positive and healthy COVID-19 negative (no sick COVID-19 negative) elicit higher/lower performance, ot ii) test whether one sampling method provided better performances than the other. On the latter point, the fact that “two sampling supports were used because at the time to the beginning of the proof-of-concept study, we did not know which one of the two sampling supports was the most effective” should appear in the manuscript, otherwise no one will understand why you did so. This study being a proof-of-concept study, this is fine to mention that. And it would be nice to have some info regarding effectiveness of both methods. 5) Last paragraph of introduction: you mention the recommendations on how to proceed to rigorously test your research question but we expect to read how you study is positioned within these recommendations. This is currently not the case. 6) There are erroneous references to the bibliography. First, page 4 “sweat does not appear to be a route of excretion for SARS-CoV-2 virus [44, 45]”. In these papers, sweat is not mentioned. Instead, it appears that in the absence of scientific study on that topic, we don’t know whether sweat is able to convey the virus or not, see the following paper: Propper, R.E., 2020. Is sweat a possible route of transmission of SARS-CoV-2? Second, page 6: “SARS-CoV-2 virus does not survive more than a few hours on cotton and disposable gauze [52]. A more recent study concludes that absorbent materials like cotton and gauze are safer than no-absorptive materials for protection from SARS-CoV-2 infection [53].” References 52 and 53 are about SARS-CoV and not SARS-CoV-2, and about the comparison between virus survival on cotton gown versus disposable (polypropylene) gown. Reformulation is needed. Overall, to make this study publishable, I would recommend to formulate the discussion and abstract even more cautiously. The conclusions are still too affirmative according to the numerous limitations. It should be explicitly stated that more research is need in the future, addressing a number of limitations present in this study – and then all limitations should be carefully listed. Also, the manuscript has clearly not been proofread by an English native speaker, as many language and wording imperfections remain. Reviewer #2: Review of the Revision R1 The Authors have thoroughly addressed my remarks and comments, both in their responses to my review and in the revision. The plenty of changes to the manuscript makes the tracking of them hardly possible, thus the revision could be considered almost as a new manuscript. Obviously the revision substantially benefits from such considerable changes. The Authors are aware that in order to avoid confounding factors, control samples must be in all aspects comparable to positive samples, except for the disease status. This was now evident from the lines 131-132, lines 142-145 and 350-365 of the revision. In lines 371-373 of the revision the Authors properly address the problem of differences in the amount of sweat collected from COVID positive and negative donors, due to possible fever, which is a typical COVID-19 symptom. This issue was my main concern. In the revision the authors inform that the amount of sweat was not recorded and it can not be excluded that it could be a confounding factor. This statement is correct, however, further discussion in lines 374-378 including citation of Concha et al. (2019), is not quite pertinent to the problem I would like to point out. In the study of Concha et al. (2019) on detection threshold of amyl acetate, the dogs had to discriminate between amyl acetate diluted in mineral oil at different concentrations and pure mineral oil as control. There were defined qualitative chemical differences between target odour and control odour up to a dilution of amyl acetate which was for dogs not distinguishable from mineral oil. On the contrary, in the present study it is still not known which volatile organic compounds in the sweat of COVID-positive patiens make that the dogs are able to discriminate the sweat odour of COVID positive from negative humans. Also, the saturation of olfactory receptors by a great amout of the odorant (revision lines 377-378), is not pertinent to the problem of the possible differences in the amount of sweat in the samples. I wanted only to point out that potential differences in the amount of sweat could be a confounding factor and/or could be a cue for dogs. I suggest to delete the passage in revision lines 374-378. An important information was given only in the response to my review, namely that the samples contained approximately 75 mg sweat and no differences were found between positive and negative samples e.g. due to more sweating in fever. In order to avoid doubts of the readers, this information should be given not only in responses to the review, but first of all in the revision, in the paragraph dedicated to sample collection,. In the revision the Authors have resigned from the 4 Ajaccio dogs and 21 sweat samples collected in Ajaccio. This was convincingly justified in Authors’ response to the Reviewers #1 and #3. As to the number of dogs used, the Authors in their explanations to the Reviewers #3 inform that now they used six dogs (for the real tests) and they did not mention anymore the number of 14 dogs used for the training. The Authors now explain that „…it added a non relevant information…”. However, in my opinion it is just an importat information, because it shows that not all dogs are able to meet the training or qualification criteria for this task. An information on how many dogs out of all candidates that started the training, were qualified for the test, would be useful for other researchers or dog trainers who will intend to continue or to repeat such experiments. As to the qualification criterion, the Authors’ information that „…The dog was considered as trained and ready for the testing session by the subjective opinion of the handler, based on the behavior of the dog….” – is acceptable, though for the sake of scientific accuracy in future experiments more precise quantitative criteria would be recommended. It is not clear to me why the Authors think the calculation of the sensitivity and specificity was not possinle ? (revision lines 276-277). It depends on how the outcome of a trial in the lineup is assessed. The authors apparently considered their lineup as a simultaneous lineup, where the dog makes a choice out of 3 or 4 samples. This approach seems to be more „convincing” because the theoretical probability of the correct indication by chance is relatively low (1/4 or 1/3). However, the actual probability of correct choices by chance depends on how many samples the dog had really sniffed. The not sniffed samples do not count for calculation of correct hits by chance. Therefore the Authors are not correct, claiming in their explanations that this probability is still 25% if the dog hits correctly on the first cone, without sniffing the remaining cones. Thus, the probabilities given in the last right-hand column of the Table 2 are not constant but variable from trial to trial. The other approach is a sequential lineup where the dogs are considered to make separate and independent yes/no decisions to each sniffed sample, at the constant 50% probability of correct hits by chance. The sequential lineup is believed to be statistically more sound. If the Authors would consider their lineup as the sequential one, all sniffed but not indicated cones with negative samples are „true negative”, all correct hits on positive samples can be regarded as „true positive”, and the indications labelled in Authors’ protocols as “failures” - are „false positive”. Since in the present protocol „…the dog was previously trained to detect one and only one COVID-19 positive cone, the dog was left free to sniff the cones before marking one cone….” and „….Once a cone was correctly or incorrectly marked, the trial was stopped..(revision lines 258-260) – it means that in the protocol like this, there were no „false negatives” (f.n.=0). Assuming that f.n. =0, the calculated sensitivity would be 100% [Ʃ true positive / Ʃ (true positive + false negative). The specificity for some dogs would be <100% [= Ʃ true negative / Ʃ (false positive + true negative)]. However, from the statement in revision lines 258-260, one can suspect that there was a bias toward reducing „false negative” responses (a miss to indicate the target sample), since the handler knew, and the dogs were trained, that in each trial one of the samples must be the target to be indicated. A problem would be when it comes to testing samples of unknown health status in real scenario. The way to avoid this problem would be either to introduce from time to time a „zero” or „blank” trial (with all negative samples in the lineup), or to set a limit for how many times the dog is allowed to sniff positive samples without indicating, before it eventually hit them. For example if the dogs has sniffed the positive sample 2 or 3 times without indicating, the outcome can be classified as the „false negative”. Since the aim of the study was only to confirm that dogs are able to discriminate COVID positive and negative sweat samples, the term „Success rate” used by the authors, is acceptable, but for the sake of scientific accurracy in future studies the Authors should be able to calculate the sensitivity and specificity. It is worth underlining that the revision is better structured than the first submission. The concluding remarks are more cautious and the references are updated. Minor remarks The terms „odour detection” and „odour discrimination” are frequently used interchangeably, however, in the context of this work I suggest to use the wording „odour discrimination” rather than „odour detection”. Odour detection refers rather to searching for unknown location of odour source in a field setting. In the odour lineup like this, the dogs are trained to visit the odour locations (cones) which are visible, and their task is to compare and to discriminate between the target and control odours. In lines 344 and 352 the year of the publication should be completed for more clarity. This is particularly important when two or more papers of the same first author are cited. Linie 373 „COVID-19 negative” should be inserted for more clarity. To sum up, only few amendments (minor revision), as listed above, should be made to the text to be fully acceptable for publication in the PLoS ONE. Reviewer #3: Thank you to the authors for taking into account the previous comments from the three reviewers, providing very detailed responses to the points raised. The revised manuscript is much improved, provides a more balanced appraisal of the study and is much clearer to read. Some of the additional information that has been added has raised further questions and comments, detailed below, but I think the manuscript would be appropriate for publication if these can be addressed. General comments - The general use of English language and grammar has greatly improved in this revision. However, there are still a few typographical errors and language issues in places that need addressing with a thorough proof reading. - Throughout the manuscript (title, abstract, discussion etc.) suggest you clarify that you were testing a discrimination of *symptomatic* SARS-CoV-2 positive vs asymptomatic SARS-CoV-2 negative. This is an important distinction, as symptomatic individuals may produce confounding odours for a wide variety of reasons. Abstract - Line 49 “The aim … was to evaluate if a sample of sweat produced by COVID-19 50 positive individuals (SARS-CoV-2 positive) can be detected by trained detection dogs among 51 sweat samples produced by COVID-19 negative individuals.” This wording suggests that dogs were evaluated using just a single (one) SARS-CoV-2 PCR positive sample. The line needs clarifying; I think you mean to say something like: “evaluate if dogs could detect individual samples of sweat produced by COVID-19 50 positive individuals (SARS-CoV-2 PCR positive) presented in a line up amongst sweat samples produced by COVID-19 negative individuals”? - I think you need to state in the abstract that there are limitations to the study (e.g. processes for collecting positive and negative samples were not completely identical, the trial stage was not fully double blind, some samples were re-used several times during the trial stage) and that a more rigorous validation study is ideally needed to confirm the results. Introduction - For balance, I think you should mention that not all studies testing canine detection of disease have shown positive results. There is currently very little evidence relating to canine detection of virus odours, particularly using live subjects (you cite one example using cell cultures). Materials and Methods Samples - Line 157. I think you need to briefly discuss in your discussion section/conclusions the possible limitation that SARS-CoV-2 positive samples were collected using full PPE whereas negatives were not. It could have produced a consistent difference in odour that confounded the results. Were the same brands of gloves used for collecting positive and negative samples? If different brands/types were consistently used for positives and negatives, it’s a major limitation as the dogs may have trained on different glove odours. - Line 177. Were the two coolers identical (same brand, from the same source etc.?). Worth stating if this was the case. If not, it’s a limitation as dogs may have trained on the odour of two different coolers. Safety of the dogs regarding SARS-CoV-2 virus infection - Line 216. Should read “non-absorptive” Testing protocol - Line 248. Please clarify the minimum number of SARS-CoV-2 negative samples that were placed out each time. If zero negatives samples were placed out, that would not be valid discrimination data and should be excluded from the analysis/results. If 1, 2 or 3 samples were placed out, please clarify how this was accounted for in the statistical analysis. Statistical analysis - Line 275. As per the abstract, this wording suggests that dogs were evaluated using just a single (one) SARS-CoV-2 positive sample. The line needs clarifying. - As mentioned in the testing protocol comment, some additional detail is needed on how the statistical analysis accounted for the different number of negative samples placed out. If some line-ups included zero negatives, these trials should be excluded from the analysis. It is vital that the analysis is based on the number of negatives placed out, not the number of cones. Clearly, dogs would have a greater chance of detecting the positive samples with lower numbers of negatives present. If not all the cones contained negatives the statistical analysis needs to be revised to account for this. Results - Line 292. The ages are the wrong way around? Table 2, as stated above, the random choice proportion should be based on the number of negatives placed out, not the number of cones (if not all cones (expected the positive) contained SARS-CoV-2 negative samples). Discussion - Line 335. This line needs revising to reflect that only positive and negative samples (not “mocks”) should have been included in the statistical analysis. - line 389 to 396. Recommend you clarify that the lack of double blinding remains a limitation that should ideally be addressed in a future validation study. Previous studies have used various techniques to reward dogs whilst remaining double blind (e.g. use a live mobile phone conversation with an independent referee who has a list of sample “identities” for each line up). The ideal validation study would be to receive samples from an independent source (e.g. from another organisation or even country) which have not been processed by your research team before being presented to the dogs. - 411 to 416. Recommend key improvements for the validation trial should include double-blinding and samples prepared and delivered by an independent organisation/source. - Line 412. Should read ‘sensitivity’ not ‘sensibility’. ********** 7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No Reviewer #2: No Reviewer #3: No [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. |
| Revision 2 |
|
Can the detection dog alert on COVID-19 positive persons by sniffing axillary sweat samples? A proof-of concept study PONE-D-20-17256R2 Dear Dr. grandjean, We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org. If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. Kind regards, Nadine Ravel Academic Editor PLOS ONE Additional Editor Comments (optional): Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation. Reviewer #1: All comments have been addressed Reviewer #3: (No Response) ********** 2. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes Reviewer #3: Yes ********** 3. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes Reviewer #3: Yes ********** 4. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes Reviewer #3: Yes ********** 5. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes Reviewer #3: Yes ********** 6. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: In this second revision of their manuscript, the authors have answered all my comments in a satisfactory manner. To me, the paper is now acceptable providing some minor changes as detailed below: 1) Ethical statements (page 5) should be formulated as proper sentences (or as such but in a separate section of the manuscript before the references, according to the editor’s preference). 2) Figures 1 and 4 are not required within the manuscript, they may appear either as Appendices at the end of the manuscript or as supplementary materials, again according to the editor’s preference. 3) As early as in the method section, it should be stated explicitly that, within a trial in a line of cones, samples are from the same hospital (are they?) and are made of the same material (gauze swabs or gauze filters). Line 190: it may be useful to explicitly say that polymer tubes were not used in the present study but collected for chemical analysis not presented here. 4) There seem to be a contradiction between Line 193 stating an average quantity of sweat secreted and Line 441 stating that amount of sweat was not recorded. Please address this issue. Reviewer #3: Thank you to the authors for revising the manuscript and taking into account the comments of the three referees. Overall, I feel the manuscript has been improved again and now provides a clear and balanced discussion of the study and the possible limitations. I have made a few minor comments below, but otherwise would be content for it to be accepted/published – many thanks. Minor comments: Introduction Line 82: The wording “was good compared with standard diagnostic methods” should be clarified (e.g. to “was comparable with” or “was better than” or “was more accurate than” – whichever term explains the comparison more clearly). Line 126: Suggest change to VOCs “_may_ be detectable by the dogs” (we cannot say for certain that they “are”). Methods Line 191: You say two sampling methods were used but three are discussed above (gauze swabs, gauze filters, polymer tubes)? Discussion Lines 459-461: “Although the data recorder was not blinded to the sample location, he remained at the back of the room and could not be seen by either the dog handler or the dog during the session. Therefore, he could not have influenced the dog’s behaviour.” Because the data recorder put out the samples there is always a risk he/she unintentionally influenced the dogs’ choices. There are many theoretical ways this could happen (e.g. they accidentally touched the COVID positive samples more than the negatives so they smell more strongly of the data recorder; they always put the positive samples out first and they had more time to release odour than the negatives etc. etc.). Fundamentally, if the team knows the sample locations there always the risk they may influence the dogs in some way they do not realise and cannot possibly predict – this is the basis of the Clever Hans Effect. That is why double-blind testing is so important for validation of these studies. I suggest you change the text to clarify that standing at the back of the room reduced the chance of the data recorder directly influencing the dog’s behaviour, but the risk of them having some unintentional influence on the choice of samples cannot be completely ruled out (without the double-blind validation study). ********** 7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No Reviewer #3: No |
| Formally Accepted |
|
PONE-D-20-17256R2 Can the detection dog alert on COVID-19 positive persons by sniffing axillary sweat samples? A proof-of-concept study Dear Dr. Grandjean: I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department. If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org. If we can help with anything else, please email us at plosone@plos.org. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Dr. Nadine Ravel Academic Editor PLOS ONE |
Open letter on the publication of peer review reports
PLOS recognizes the benefits of transparency in the peer review process. Therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. Reviewers remain anonymous, unless they choose to reveal their names.
We encourage other journals to join us in this initiative. We hope that our action inspires the community, including researchers, research funders, and research institutions, to recognize the benefits of published peer review reports for all parts of the research system.
Learn more at ASAPbio .