The Confusion Assessment Method Could Be More Accurate than the Memorial Delirium Assessment Scale for Diagnosing Delirium in Older Cancer Patients: An Exploratory Study
Round 1
Reviewer 1 Report
In the context of the type of study presented by the authors in the "Materials and Methods" section, the use of the notion of "incidence" is inappropriate. I believe the authors are referring in this study to prevalence. They evaluate a sample from a reference population, at a certain moment, aspect specific to transversal studies. In these studies, incidence is not assessed, but prevalence.
In section 2.2. the authors calculate the minimum sample size.
The minimum sample size calculated by the authors was n=44.
Here I noticed an obvious error.
When we know from various sources (specialist literature) the proportion of an event in a population we can use this information to calculate the minimum sample size.
The formula is:
Thus, following the calculations, it follows:
So the calculated correct value is n ≥ 269 cases for 5% error and 6720 cases for 1% error.
Another aspect I would like to highlight is related to the purpose of the research.
In lines 68 and 72 the author’s mention:
"No tools have been validated exclusively in older people, although existing tools show adequate accuracy figures according to the literature [12]. Our aim is to estimate the cumulative incidence of delirium in older patients with cancer and to assess the accuracy of the CAM and MDAS scales against the gold standard medical diagnosis based on DSM-5 criteria."
1. The sensitivity and specificity of MDAS and CAM relative to DSM was calculated. But, wasn't it more relevant to compare the results obtained based on MDAS and CAM with those obtained through DMS (I am referring here to the evaluation based on the concordance coefficient).
2. In figure 1B the authors present the AUC which is very small (AUC=0.59 (95%CI: 0.45 – 0.72)). This contradicts the calculated Se and Sp values. I believe that the results of the study are greatly affected by its design.
3. The study's conclusions are not based on proper analysis. The research conducted does not assess incidence, and the comparison of methods contains obvious methodological errors.
Please see the updated comments in the attachment.
Comments for author File: 
Comments.pdf
Author Response
We would like to thank you the reviewer for the attentive evaluation of our work and your insightful comments, which have helped us to substantially improve the manuscript.
Please find attached the point by point letter
Author Response File: 
Author Response.docx
Reviewer 2 Report
The subject of the ms. is very interesting. However, there are some issues that should be handled by the authors. The main one has to do with the sample of the study. It seems that this is not a sample of older adults with cancer but a sample of seriously demented people with cancer! What kind of results of this study could be representative for the population of older adults with cancer? Another issue is the selection of the specific tools to be examined. Why didn't the authors choose to examine thw DOS instrument as well? In fact, they mention it in the introduction and the discussion but they do not include it in their study. The description of the results should follow the clasic ways of describibg the analyses. As regards the discussion, this appears to don't focus on the findings for the "better instrument" but has to do with cognitive impairment. This is confusing given the aim of the study but it comes as a result of the special population selected to bre studied. Generally speaking, I wonder whether it is possible to discuss about the 'reasons' of delirium when you have to do with people suffering from both cancer and dementia. In this case, extended discussion is needed about the relationships between dementia and delirium.
Author Response
We would like to thank you the reviewer for the attentive evaluation of our work and your insightful comments, which have helped us to substantially improve the manuscript.
Please find attached the point by point letter
Author Response File: Author Response.docx
Reviewer 3 Report
The authors present an interesting article. They found that Confusion assessment Method is more accurate than the Memorial Delirium Assessment for diagnosing delirium in older cancer patients with cognitive impairment.
The topic of the study is highly relevant in everyday clinical practice.
Introduction
- Introduction is detailed and informative to the reader.
- In line 39 „as-sociated should be corrected in associated“
- In line 47 „in-creased should be corrected in increased“
Materials and Methods
- I suggest a clear description of inclusion and exclusion criteria. The title of the study implicates that only people with cognitive impairment were included, but in section „study design, setting and participants“ there is no information about that point.
- I absolutely appreciate authors‘ efforts performing sample size calculation. Please give a more detailed explanation which statistical test was used.
Results
- In table 2 scores for MMSE with a mean of 7.9 are very low. Is this really the case? These scores might be comparable with far progressed dementia. I am in doubt that patients with that low scores were able to follow the informed consent discussion prior study inclusion? Were legal representatives integrated?
- Do the authors have information about previosly diagnosed dementia in the investigated cohort. It might be useful to adapt table 1 with that information. Table 1 contains just a line with number of patients with cognitive impairment.
Discussion
- Authors found a delirium incidence of 62.7% in their study. In introduction (line 51, 52) cumalative incidence of delirium in older oncological patients was stated with around 25%. What are the reasons that the incidence in the presented study was more than twice as high. Please discuss that point.
- Seventy-five patients were included in the study. Sample size calculation revealed a number of 44 patients, so the number of included patients is 1.7 fold as high. Please discuss in limitations section the high number of dropouts.
- Discussion should be adapted with suggestions for future research.
Author Response
We would like to thank you the reviewer for the attentive evaluation of our work and your insightful comments, which have helped us to substantially improve the manuscript.
Please find attached the point by point letter
Author Response File: Author Response.docx
Reviewer 4 Report
The manuscript assesses different instruments for detection of delirium among hospitalized older oncology patients.
Methods: line 97 - who assessed for the gold standard DSM-5 evaluation? Was the same assessor used throughout the study?
Results: - Include in Table 1: # medical and # surgical patients, major reason for admission, length of stay
Clarifications and language modifications
line 191 - consider: prevalence of elderly oncology patients with different geriatric syndromes...
line 202 - consider: vulnerable, and goals of care and decision....
line 208 - consider changing "prevention" to "management"
consider deleting lines 2240228 as duplicative. Do consider addressing: Does the presence of dementia or delirium have implications among older oncology patients for goals of care discussions and treatment decisions?
Conclusons - line 234: consider rephrasing: ...delirium in hospitalized older people...
line 238: consider adding: and the implications of these assessments among older oncology patients for goals of care discussions and treatment decisions?
Author Response
We would like to thank you the reviewer for the attentive evaluation of our work and your insightful comments, which have helped us to substantially improve the manuscript.
Please find attached the point by point letter
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
The authors made some wording changes in the manuscript. They did not significantly change the quality of the manuscript.
If the authors wrote in the manuscript according to my directions:
"Assuming a prevalence of delirium in the older population of 22.6%, a sample size of 269 participants was calculated for an error of 5% and a confidence level of 95% [13]." (lines 88-89)
I believe that it was necessary to make a flow chart that would clearly show the number of excluded patients and for what reasons, of which only 75 remained.
In my opinion, the conclusions of this study are not sufficiently valuable.
Both the design of the study and the interpretation of the results present deficiencies.
Author Response
REVIEWER 1
We would like to thank the reviewer for the critical review and valuable comments, which we have taken into account in this revised manuscript. Itemized responses are listed below. All the modifications have been marked in red throughout the manuscript to facilitate review.
The authors made some wording changes in the manuscript. They did not significantly change the quality of the manuscript.
If the authors wrote in the manuscript according to my directions:
"Assuming a prevalence of delirium in the older population of 22.6%, a sample size of 269 participants was calculated for an error of 5% and a confidence level of 95% [13]." (lines 88-89)
I believe that it was necessary to make a flow chart that would clearly show the number of excluded patients and for what reasons, of which only 75 remained.
In my opinion, the conclusions of this study are not sufficiently valuable.
Both the design of the study and the interpretation of the results present deficiencies
Author’s answer: Thank you for your comment. The authors have added the sample size calculation following the reviewer's directions, After adjusting the sample size calculation, the authors have changed the title and design of the study as an exploratory study.
We regret the misunderstanding regarding the recruitment of the sample, as all patients who agreed to participate in the study period were included, so we cannot add the flowchart suggested by the reviewer.
Reviewer 3 Report
The authors of the manuscript "The Confusion Assessment Method is more accurate than the Memorial Delirium Assessment Scale for diagnosing delirium in older cancer patients " addressed some points of criticism.
I have further comments to the revised version of the manuscript:
- Authors performed a new sample size calculation that yielded 269 participants, but finally only 75 patients were included. This point impacts the results and conclusions of the entire study. It is not provided which reasons lead to the discrepance between the calculated sample size and the much lower number of finally included patients.
- In results section it is mentioned that younger patients and men are more likely to have delirium (lines 142,143). This point is very interesting, but unfortunately not sufficiently discussed.
- Authors state in discussion that delirium prevalence in their study (62.7%) is in line with previous investigations (22-57%) (lines 178, 179, 187, 188). Prevalence of 62.7% is very high and different to the results of the cited studies. In my opinion this statement cannot be maintained.
Author Response
Reviewer 3
We would like to thank the reviewer for the critical review and valuable comments, which we have taken into account in this revised manuscript. Itemized responses are listed below. All the modifications have been marked in red throughout the manuscript to facilitate review.
The authors of the manuscript "The Confusion Assessment Method is more accurate than the Memorial Delirium Assessment Scale for diagnosing delirium in older cancer patients " addressed some points of criticism.
I have further comments to the revised version of the manuscript:
- Authors performed a new sample size calculation that yielded 269 participants, but finally only 75 patients were included. This point impacts the results and conclusions of the entire study. It is not provided which reasons lead to the discrepance between the calculated sample size and the much lower number of finally included patients.
Author’s answer: Thank you for your comment. All patients who agreed to participate in the study period were included. After adjusting the sample size calculation, the authors have changed the title and design of the study as an exploratory study.
- In results section it is mentioned that younger patients and men are more likely to have delirium (lines 142,143). This point is very interesting, but unfortunately not sufficiently discussed.
Author’s answer: Thank you for your comment. The authors have added the following in the Methods section:
“The delirium group was slightly younger than the non-delirium group. The older the age, the greater the risk of suffering delirium according to the literature. This risk factor is considered a low-intensity risk factor in the same way as renal failure, dehydration and drowsiness. In our study, the presence of cognitive impairment could interfere with the analysis so studies with a larger sample and groups without cognitive impairment are needed. There was also a higher proportion of men in the delirium group than in the non-delirium group. Male gender is more related to delirium than female gender and it is considered medium-intensity risk factors. It is important to continue research in this population with delirium in order to implement prevention strategies [24].”
- Authors state in discussion that delirium prevalence in their study (62.7%) is in line with previous investigations (22-57%) (lines 178, 179, 187, 188). Prevalence of 62.7% is very high and different to the results of the cited studies. In my opinion this statement cannot be maintained
Author’s answer: Thank you for your comment. We agree with the reviewer. The authors have reworded the sentence:
“Our results had a higher prevalence than other studies with the prevalence of delirium in medical oncology units and palliative care units, with rates from 22% to 57% [5]. This high prevalence is due to the inclusion of older people only, as well as the large presence of cognitive impairment in the study sample”.
Reviewer 4 Report
The authors have made a credible response to reviewer comments.
Additional detailed comments included in the attached document. Additional comparisons need to be included in the results for this study.
Comments for author File: Comments.pdf
Author Response
We would like to thank the reviewer for the critical review and valuable comments, which we have taken into account in this revised manuscript. Itemized responses are listed below. All the modifications have been marked in red throughout the manuscript to facilitate review.
- … diagnostic complexity, and often subtle presentation...
- by two medical oncologists
- preexisting cognitive impairment
- accuracy metrics
- obtained on admission.
- The oncologists had X% agreement with individual assessments.
Author’s answer: Thank you for the suggestion. The authors have reworded the sentences according to the reviewer’s suggestion.
- Consecutive or convenience sample?
Author’s answer: Thank you for the suggestion. The authors have added the following in the Methods section:
“During the 19-month recruitment period, a consecutive sample of 75 participants was obtained. Due to slow recruitment, an exploratory study was carried out to obtain preliminary results”.
- delete lines 166-178 and tables 2,3. Instead calculate sens/spec for CAM vs MDAS (compared to DSM-5) for the patients identified as demented by MMSE, also if possible, for patients identified as not-demented by MMSE.
Author’s answer: Thank you for the suggestion. The authors calculated the scores in participants with cognitive impairment (Table 4). We could not perform the analysis in persons identified as not dements by MMSE as there were no participants with such MMSE scores in the study sample.
- do not use gold standard for CAM, only DSM-5
- Accuracy of MDAS and CAM compared to DSM-5 delete 3rd row comparing MDAS to CAM
Author’s answer: Thank you for the suggestion. The authors have deleted the 3rd row.
- Figure 1 shoe AUC for MDAS and another for CAM. Don't show sens/spec, as these numbers are reported in the text.
- delete lines 190-192
- … impairment in patients who eventually developed delirium as well as those who did not.
- delete: and not only for diagnosing dementia
...cognitive assessment as this may impact treatment goals.....
- Delete lines 235-232
- Rephrase…in the older population should be informed by a comprehensive understanding of each individual’s Goals of care.
- Delete lines 249-253, 261-262
- delete lines 264-265, rephrase: Raising the MDAS cut point for delirium to 15 still only gives a sensitivity of .50 and specificity of .63.
Author’s answer: Thank you for the suggestion. The authors have reworded the sentences according to the reviewer’s suggestion.
-delete lines 267-270, instead discuss sens/spec of MDAS and CAM for patients with dementia.
Author’s answer: Thank you for the suggestion. The authors have added lines 267-270 (now 264-267) following the suggestion of another reviewer. The authors have added the lines 272 to 277 in the discussion section regarding the sensitivity and specificity of delirium assessment tools in people with dementia.
- delete lines 287-292, Rephrase: Limitations include utilization of two oncologists rather than a psychiatrist to administer the gold standard DMS-5 for delirium. However the oncologists were skilled clinicians comfortable with the care of older adults. It was not possible to analyze the severity, duration or outcomes of delirium in our subjects.
Author’s answer: Thank you for the suggestion. The authors have reworded the sentences according to the reviewer’s suggestion.
Round 3
Reviewer 3 Report
The authors of the manuscript "The Confusion Assessment Method could be more accurate than the Memorial Delirium Assessment Scale for diagnosing delirium in older cancer patients. An exploratory study." implemented some suggestions, but the main point of criticism remains: the discrepance between the calculated sample size and the much lower number of finally included patients. I still miss a sufficient explanation in the manuscript.
Author Response
We would like to thank the reviewer for the critical review and valuable comments, which we have taken into account in this revised manuscript. Itemized responses are listed below. All the modifications have been marked in red throughout the manuscript to facilitate review.
The authors of the manuscript "The Confusion Assessment Method could be more accurate than the Memorial Delirium Assessment Scale for diagnosing delirium in older cancer patients. An exploratory study." implemented some suggestions, but the main point of criticism remains: the discrepance between the calculated sample size and the much lower number of finally included patients. I still miss a sufficient explanation in the manuscript
Author’s answer: Thank you for the suggestion. The authors have reworded the following in the Methods section:
“During the 19-month recruitment period, a consecutive sample of 75 participants was obtained. Due to slow recruitment, an exploratory study was carried out to obtain preliminary results”.
