Round 1

Reviewer 1 Report

Reviewer #1: applsci-1184994.

Intra- and Inter- Rater reliability of The Much and Little Used IsoMed 2000 Dynamometer for Knee Flexion and Extension Peak Torque in Athletes

This paper assessed the differences in reliability between two different (newer vs. older) IsoMed 2000 dynamometers. The idea was to see if overall usage had a negative effect on reliability of the dynamometer. All participants in the study were considered weight lifting athletes.
The main concerns I have about this paper relate to:

1) Methods

Subjects: What type of resistance training did the athletes perform? Free-weight? Body weight? Stationary machines? Olympic lifting? What was the participants 1-RMs? Were the athletes required to get their strength assessments twice a year at the health center? What sports/events did the people participate in?

Procedures: What was the specific angle for the knee extensors and knee flexors when the movements were performed? 90-180 degrees? The main issues with the present study is how was strength values calculated? Did you use a custom software system to detect the onset of torque? Was there a certain torque threshold that had to be applied to allow for your data output? 7.5 N.m for the knee extensors and 4 N.m for the knee flexors? Without knowing how you calculated the torque signals, the present output of your study is hard to interpret.

Author Response

Reviewer 1:

Methods

Subjects: What type of resistance training did the athletes perform? Free-weight? Body weight? Stationary machines? Olympic lifting? What was the participants 1-RMs? Were the athletes required to get their strength assessments twice a year at the health center? What sports/events did the people participate in?

Response: The athletes included in the study are professional weightlifting athletes. Athletes compete at an international level. Athletes apply standard weightlifting training. Therefore, the information you mentioned about the athletes was not received. Athletes are included in the Olympic athletes’ pool of our country. However, there are no athletes participating in the Olympics.

Procedures: What was the specific angle for the knee extensors and knee flexors when the movements were performed? 90-180 degrees? The main issues with the present study is how was strength values calculated? Did you use a custom software system to detect the onset of torque? Was there a certain torque threshold that had to be applied to allow for your data output? 7.5 N.m for the knee extensors and 4 N.m for the knee flexors? Without knowing how you calculated the torque signals; the present output of your study is hard to interpret.

Response: The angular velocities of 60 and 180 degrees / sec that we used in our study were determined by looking at the literate. Likewise, these angular velocities are frequently used in muscle strength measurement and strength balance calculation. In our study, our hypothesis was to compare PTs between the two devices. For this reason, the angle values at the time that PTs were formed were not analyzed. The other data mentioned above are provided automatically by the device's own software.

Reviewer 2 Report

The comments and suggestions for authors are in the attached file.

Comments for author File: Comments.docx

Author Response

Reviewer 3:

 

Abstract

The smallest detectable change of the old dynamometer was 2.48, limits of agreement were -2.24 and 2.62; and of the new dynamometer were 2.22, limits of agreement were -2.48 and 2.16.

Row 22-23 - in accordance with the calculated results, it is not clear whether these differences significantly support the conclusions of the study.

Introduction

Row 39 – The citation at the end of this sentence [4] deals with shoulder rotators, not the strength of knee flexors. Insert a more appropriate citation.

 „For this reason, the strength of the knee flexor and extensor muscles is frequently evaluated both for the preparation and regulation of training programs at the beginning of the season and in the middle of the season and at the end of the season to determine the amount of training programs to reach the desired goal“ [4]

Response: Thank you for your attention, we put wrong citation, we corrected it.

Row – 44 The citation at the end of this sentence deals [6] only with a held-hand dynamometer, insert a more appropriate citation, or combine it with the following sentence (in row 46)

„Muscle strength can be evaluated by different methods such as a manual muscle test…” [6].

Response: The literature added near to the [6] literature

Row 60 - The authors describe the situation of dynamometers in their clinic: It does not seem appropriate to me to describe the situation in the clinic in the introduction. Rather, there should be a description and literature focused on measured data in research such as ICC and reliability (Intra, Inter) of dynamometers in stated parameters (knee flexion/extension).

Response: The paragraph starting with 'In our clinic' was not removed because one of the referees deemed this section appropriate, but the ICC discussion you suggested was added to the previous paragraph.

Participants

Row 96 - there is stated two times Table 1. in the title of table 1.

Response: It was corrected.

Statistical Analysis

Row 181 – The authors stated effect sizes 0.5 for calculating G*power. Why this value? And why 90% power, when standard is 95 % power. This was not properly explained. Moreover, simple size, for two-tail t-test with a difference between two dependent means are 44 subjects needed.  

 

Row 201 - parentheses are missing from the citation or what does it mean the number 19 at the end of sentence. The original source of the citation is presented in citation [20] not in citation [19].

Response: It was corrected.

Discussion

Row 339 – Citation [20] In this study was not evaluated ICC of the knee flexor/extensor but maximum grip strength and pressure pain threshold.

Response: This citation was removed and the correct one was placed. We are sorry.

 

Row 358-68 - Citations of the studies in the text are different than in other parts of the manuscript (e.g., row 382, Thompson et al (1989) and Keilani et al. (2007)). Choose the correct norm in concomitant the rules of the journal.

Response: The citations were made in line of the rules of the journal.

 

Row - 528 Is it not clear how the study cited at the end of the sentence [33] contributes to the meaning of this statement?

Response: This section was removed.

Overall - I recommend commenting more clearly in the discussion on the specific differences found between dynamometers in connection with a possible error of repeated measurements on one dynamometer (row 377 – 381).

Response: The discussion section was rearranged in line of the comment.

 

References

Row 462 – wrong citation, correct (2019), 28, 1202-1211

Response: It was corrected as recommended.

 

Reviewer 3 Report

The authors aimed to evaluate the intra-rater and inter-rater reliability of two IsoMed-2000 dynamometers for knee flexor and extensor peak torque measurements in a concentric test. The dynamometers had been used to different extents and referred to as “old” and “new”. They evaluated 37 young (aged 15-19 years) weightlifters (12 female, 25 male) twice on each dynamometer. Their principal findings were that the old dynamometer measured higher peak flexion torques but lower peak extension torques than the new dynamometer. The authors also present high intra-class correlation coefficients (ICCs) for repeated testing on the same dynamometer and between dynamometers. However, I expect high ICCs since the participants varied due to sex, age, and body mass, resulting in a relatively small within-individual variation compared to between-individual variation (the authors seem not to have normalized the peak torque values to body mass).

The study appears well-executed, and it is evident that the authors have invested substantial time to execute this study (also including a habituation testing session which is good but rare).

I like that you wanted to investigate the observation of different measurements further (lines 60-68) and publish these results for others' benefit. However, by only including human testing and presenting peak torque values with ICCs, SEMs and limits of agreement, this study is not designed to explain why a well-used dynamometer would provide different measurements compared to a less-used dynamometer. The main recommendation not to compare strength data between dynamometers is, to be frank, relatively accepted. As such, this paper comes off as an exploratory case-control study that compared a well-used and a less-used dynamometer. Further comparisons of other dynamometers with different histories of usage would provide a stronger case in generalizing the observations to dynamometers in general.

Further, on lines 68-70 on page 2 you write: “In the literature, we did not find any studies that reveal the relationship between the duration of use of isokinetic dynamometers and their measurement output”. I wonder whether this also includes information from companies that produce dynamometers. These companies should have a lot of such data and explanations as to why you observed different measurements, and I wonder whether or not such information is available (I have not checked, so this is genuine speculation). Have you contacted the D&R Ferstl GmbH company about this, or do you know of any material concerning calibration related to this? You mention at line 398 that you contacted sellers but were given no information? If such information is available, the main finding of this study is, unfortunately, superfluous. But, if no such information is available, I believe that your results are worthwhile publishing, seeing that this study seems to be the first of its kind and comes with an apparent clinical recommendation. In that scenario, this study's publishing hopefully spurs others in publishing similar studies, which ultimately results in a better knowledge of how strength should be evaluated in dynamometers.  If there is a clear explanation of different peak torque measurements between dynamometers, this knowledge should be out there.

Regarding the manuscript itself, I think there are a few issues that require major revision to improve the paper's clarity and the data presented, but these should be resolved reasonably simple. See below for both major and minor comments, presented in the line numbers' order, thus from start to end.

Title

I find the current title a bit boring and generic. Also, I think the description of “The Much and Little Used” is unclear. I prefer a title that describes your main finding of different peak torques values between a well-used and a less-used dynamometer since the high reliability (ICCs) is given considering the relatively small within-person variation compared to the between-subject variation.

Introduction

Lines 46-50, long sentence, divide into two.

Lines 56 and 57, use active voice instead of “It is stated”.

Line 60, consider a new section here, starting with “In our clinic”. But add another final sentence to appropriately finish the prior section.

Lines 66-68, since your results show different knee flexion and knee extension differences between the dynamometers (one higher, one lower), you should clarify the peak values you mean.

Lines 80-83, you write “In light of this information”, but to what information are you referring? You describe no previous information that motivates any hypotheses apart from your own experience of underestimation. Also, the “metal fatigue” should be described in the Introduction if included. If you do not state any previous information to support your hypothesis, it sounds post hoc and makes the reader (at least me) skeptical. Please extend the Introduction a bit to allow the reader to understand and (in theory) agree with your hypothesis.

Materials and Methods

Line 91, my preference is to write out the exact number of females and males. It is easy to see the ratio of females and males for your total sample.

Line 92, please use the same number of significant figures for age to be accurate.

Table 1, weight should be body mass since in kg, and the data for height should be in m, not cm. Also, use the same number of significant figures across the data for each variable.

Lines 100-101, you write “…to have participated in an international competition for at least 3 years.”, did the 15-year-old weightlifters start international competitions when 12 years old? Did you mean “…competition within the last 3 years.”?

Line 101-102, did you use exclusion criteria of quadriceps- and hamstring-specific resistance training for internationally competitive weightlifters? What did the weightlifters do during training? Please explain since this does not make sense to me.

Line 103, I do not understand “throwing” here; please clarify.

Line 111 and 114, please provide the same information of how old these dynamometers are that you stated in the Introduction on lines 62-63. I think this information adds value here and should be easy to find.

Line 123, when was the familiarization test performed in relation to the actual tests?

Lines 124-127, first, you state that participants were required to refrain from any vigorous activity for at least 8 hours before testing, and in the following sentence 48 hours before testing. Which is it?

Table 2, I like the block randomization protocol.

I think that lines 151-166 and 176-178 should be combined and revised to blend with the information in the current heading 2.3. Procedures, since all of it describing the procedures. The information in lines 166-175 can have a heading on its own, e.g., 2.4. Outcome variables, or similar. These outcome variables could also be presented in a table if room for this.

I lack information that describes data analysis. You do not mention: if you normalized the data (I guess not when looking at the actual data you present in Tables 3-6); if you processed the data with any filter before extracting peak values; if you used the individual average peak value for test-retest analyses. Please add this information.

Lines 180-182, your description of the power analysis does not provide any information that allows the reader to understand what you did and, therefore, do not really add anything here. A more accurate description should include additional information: What type of power analysis? For what statistical test is the power analysis for? What is your dependent variable? What is your independent variable? How did you estimate the effect size?

Lines 197-202, I am not too fond of the SEM calculation where the ICC value is included for a paper such as this one. Since you want to have a proper estimation of the SEM, I think it is better to calculate it without the ICC, thus as the average within-person standard deviation from the trials included in the analyses. This calculation is also described in the study by Weir. The reason for this argument is that your sample (males and females of different ages and body mass) is sure to result in relatively small within-individual variation compared to between-individual variation, therefore very high ICCs. I would not want these high ICCs (basically by study design) to affect the SEM calculations when trying to understand differences in actual data between the two dynamometers. Please consider the basic calculation for SEM as the average within-person standard deviation from the analyses' trials.

Line 202, the references 19 and 20 have used different populations than weightlifters, so the classification of “acceptable” and “excellent” here is redundant. I do not think these classifications adds anything to the paper since the data stands by itself and should be discussed instead. Consider removing the classifications of “acceptable” and “excellent” throughout the paper.

Results

Tables 3, 4, 5, 6, add units to all data that have units. Also, I do not understand what the significant ICCs mean. Significant for what? Further, the discrepancies in 95% LoA and Figures 1-4 warrant discussion since unclear at the moment (95% LoA do not contain the value 0, but the plots tell another story).

Discussion

In general, I think that your actual data need further discussion, as do the analyses. Also, you have no information on study limitations, which there always are that you should acknowledge.

The text on lines 298-310 and 311-323 are essentially repetitions. Please revise and combine this text.

Lines 345-348, well, you mention the issue of high variance between subjects that causes high ICCs on lines 385-386. Please discuss your actual data a bit more with your results.

Author Response

Reviewer 2:

Open Review

The authors aimed to evaluate the intra-rater and inter-rater reliability of two IsoMed-2000 dynamometers for knee flexor and extensor peak torque measurements in a concentric test. The dynamometers had been used to different extents and referred to as “old” and “new”. They evaluated 37 young (aged 15-19 years) weightlifters (12 female, 25 male) twice on each dynamometer. Their principal findings were that the old dynamometer measured higher peak flexion torques but lower peak extension torques than the new dynamometer. The authors also present high intra-class correlation coefficients (ICCs) for repeated testing on the same dynamometer and between dynamometers. However, I expect high ICCs since the participants varied due to sex, age, and body mass, resulting in a relatively small within-individual variation compared to between-individual variation (the authors seem not to have normalized the peak torque values to body mass).

Response: The body weight of the athletes is manually recorded on the Isomed 2000 device before measurement. This device uses this manually entered value to normalize PT to body weight. We use both PT and PT / W values in our studies. However, in this study, the ICC values between the two measurements and the two devices were not shown in the tables because both PT and PT / W has the same ICC and the number of tables was large.

The study appears well-executed, and it is evident that the authors have invested substantial time to execute this study (also including a habituation testing session which is good but rare).

I like that you wanted to investigate the observation of different measurements further (lines 60-68) and publish these results for others' benefit. However, by only including human testing and presenting peak torque values with ICCs, SEMs and limits of agreement, this study is not designed to explain why a well-used dynamometer would provide different measurements compared to a less-used dynamometer. The main recommendation not to compare strength data between dynamometers is, to be frank, relatively accepted. As such, this paper comes off as an exploratory case-control study that compared a well-used and a less-used dynamometer. Further comparisons of other dynamometers with different histories of usage would provide a stronger case in generalizing the observations to dynamometers in general.

Further, on lines 68-70 on page 2 you write: “In the literature, we did not find any studies that reveal the relationship between the duration of use of isokinetic dynamometers and their measurement output”. I wonder whether this also includes information from companies that produce dynamometers. These companies should have a lot of such data and explanations as to why you observed different measurements, and I wonder whether or not such information is available (I have not checked, so this is genuine speculation). Have you contacted the D&R Ferstl GmbH company about this, or do you know of any material concerning calibration related to this? You mention at line 398 that you contacted sellers but were given no information? If such information is available, the main finding of this study is, unfortunately, superfluous. But, if no such information is available, I believe that your results are worthwhile publishing, seeing that this study seems to be the first of its kind and comes with an apparent clinical recommendation. In that scenario, this study's publishing hopefully spurs others in publishing similar studies, which ultimately results in a better knowledge of how strength should be evaluated in dynamometers.  If there is a clear explanation of different peak torque measurements between dynamometers, this knowledge should be out there.

Response: Yes, we talked to D&R Ferstl GmbH company through the distributor firm of the device. He suggested that we test the weight meter of the device using ten kilograms of weight. However, there was no method standardization in this test application; The following ten kilograms were asked to be placed in a bag and placed on the longest dynamometer attachment with a rope. It was reported that if there was no difference between the three measurements, the device would give an accurate measurement. In this measurement, the force arm, the length and oscillation of the tied bag were not taken into account, which made us doubt about the calibration. Later, we formed the hypothesis of this study with the information from the athletes that 'he could not obtain the PT value he obtained in one device, in the other device'.

Also, we have not received information from the company about how the measurement outputs have changed over the years. However, when we received the device, the company suggested regular annual maintenance and repair of the device. We have the annual maintenance of the devices made by the company.

 

Title

I find the current title a bit boring and generic. Also, I think the description of “The Much and Little Used” is unclear. I prefer a title that describes your main finding of different peak torques values between a well-used and a less-used dynamometer since the high reliability (ICCs) is given considering the relatively small within-person variation compared to the between-subject variation.

Introduction

Lines 46-50, long sentence, divide into two.

Response: The sentence divided into two.

Lines 56 and 57, use active voice instead of “It is stated”.

Response: We used it as an active sentence.

Line 60, consider a new section here, starting with “In our clinic”. But add another final sentence to appropriately finish the prior section.

Response: A new section starting with “In our clinic” created. And another final sentence added to the prior section.

Lines 66-68, since your results show different knee flexion and knee extension differences between the dynamometers (one higher, one lower), you should clarify the peak values you mean.

Response: We added to the sentence the ‘knee flexion and knee extension’

Lines 80-83, you write “In light of this information”, but to what information are you referring? You describe no previous information that motivates any hypotheses apart from your own experience of underestimation. Also, the “metal fatigue” should be described in the Introduction if included. If you do not state any previous information to support your hypothesis, it sounds post hoc and makes the reader (at least me) skeptical. Please extend the Introduction a bit to allow the reader to understand and (in theory) agree with your hypothesis.

Response: “In light of this information” section corrected and “metal fatigue” is deleted.

Materials and Methods

Line 91, my preference is to write out the exact number of females and males. It is easy to see the ratio of females and males for your total sample.

 

Line 92, please use the same number of significant figures for age to be accurate.

Response: It was corrected according to your advice.

Table 1, weight should be body mass since in kg, and the data for height should be in m, not cm. Also, use the same number of significant figures across the data for each variable.

Response: It was corrected according to your advice.

Lines 100-101, you write “…to have participated in an international competition for at least 3 years.”, did the 15-year-old weightlifters start international competitions when 12 years old? Did you mean “…competition within the last 3 years.”?

Response: It was corrected as “…competition within the last 3 years.”

Line 101-102, did you use exclusion criteria of quadriceps- and hamstring-specific resistance training for internationally competitive weightlifters? What did the weightlifters do during training? Please explain since this does not make sense to me.

Response: That means the isokinetic exercise training, it was also changed in the text.

Line 103, I do not understand “throwing” here; please clarify.

Response: This sentence was rearranged.

Line 111 and 114, please provide the same information of how old these dynamometers are that you stated in the Introduction on lines 62-63. I think this information adds value here and should be easy to find.

Response: It was added in line with your recommendation.

Line 123, when was the familiarization test performed in relation to the actual tests?

Response: Just before each test.

Lines 124-127, first, you state that participants were required to refrain from any vigorous activity for at least 8 hours before testing, and in the following sentence 48 hours before testing. Which is it?

Response: Since the isokinetic tests were taken in the morning hours, they were asked to rest the night before and not to take any stimulants. In addition, 48 hours before the tests were carried out, routine training was allowed, and heavy training or competition was not allowed.

Table 2, I like the block randomization protocol.

Response: Thank you for your comment.

I think that lines 151-166 and 176-178 should be combined and revised to blend with the information in the current heading 2.3. Procedures, since all of it describing the procedures. The information in lines 166-175 can have a heading on its own, e.g., 2.4. Outcome variables, or similar. These outcome variables could also be presented in a table if room for this.

Response: As you indicated, combining and opening new headers were done.

I lack information that describes data analysis. You do not mention: if you normalized the data (I guess not when looking at the actual data you present in Tables 3-6); if you processed the data with any filter before extracting peak values; if you used the individual average peak value for test-retest analyses. Please add this information.

Response: There were outliers in our initial data and we removed them to get a normal distribution. Because we compared the dynamometers and trials, we applied Shapiro-Wilks Test separately for the trials and dynamometers. The final data had a normal distribution (p > 0.05).

The detail is given in Section 2.5, paragraph 3.

“Before the analysis, the outliers were removed from the data. Each measure was evaluated for normality with the Shapiro-Wilks Test and demonstrated that all data had a normal distribution (p > 0.05).”

 

Lines 180-182, your description of the power analysis does not provide any information that allows the reader to understand what you did and, therefore, do not really add anything here. A more accurate description should include additional information: What type of power analysis? For what statistical test is the power analysis for? What is your dependent variable? What is your independent variable? How did you estimate the effect size?

Response: We changed and did again the power analysis and wrote it below another reviewer comment.

Lines 197-202, I am not too fond of the SEM calculation where the ICC value is included for a paper such as this one. Since you want to have a proper estimation of the SEM, I think it is better to calculate it without the ICC, thus as the average within-person standard deviation from the trials included in the analyses. This calculation is also described in the study by Weir. The reason for this argument is that your sample (males and females of different ages and body mass) is sure to result in relatively small within-individual variation compared to between-individual variation, therefore very high ICCs. I would not want these high ICCs (basically by study design) to affect the SEM calculations when trying to understand differences in actual data between the two dynamometers. Please consider the basic calculation for SEM as the average within-person standard deviation from the analyses' trials.

“For assessment of absolute reliability standard error of measurement (SEM) and smallest detectable change (SDC) were used. SEM was calculated as  and SDC was calculated as SDC95=SEM×1.96√2, where 1.96 is the z score for a 95% CI [18].”

 

Line 202, the references 19 and 20 have used different populations than weightlifters, so the classification of “acceptable” and “excellent” here is redundant. I do not think these classifications adds anything to the paper since the data stands by itself and should be discussed instead. Consider removing the classifications of “acceptable” and “excellent” throughout the paper.

 

Results

Tables 3, 4, 5, 6, add units to all data that have units. Also, I do not understand what the significant ICCs mean. Significant for what? Further, the discrepancies in 95% LoA and Figures 1-4 warrant discussion since unclear at the moment (95% LoA do not contain the value 0, but the plots tell another story).

Response: In statistics, first the significance of a coefficient is investigated. If the coefficient is statistically significant, then it is interpreted. In Tables 3, 4, 5, 6, the significance of the interclass coefficient (ICC) is investigated and the coefficients are found statistically significant (p<0.01).

 

95% LoA of Table 5 and 6 are re-calculated. Also, the following interpretations are added.

The Bland-Altman plots showed no systematic bias for EXDMPT180 and EXNDMPT180 (see Figures 1 and 2). According to Figure 3, it could be said that there is a proportional bias between the FLNDMPT60 and FLNDMPT180 measures of two dynamometers (FLDMPT60 r = 0.333, p = 0.044; FLDMPT180 r = 0.401, p = 0.014). According to Figure4,  it could be said that there is a proportional bias between the FLDMPT60, FLNDMPT60 and FLNDMPT180 measures of two dynamometer (FLDMPT60 r = 0.333, p = 0.046; FLNDMPT60 r=0.407, p=0.012; FLDMPT180 r = 0.438, p = 0.007).

Discussion

In general, I think that your actual data need further discussion, as do the analyses. Also, you have no information on study limitations, which there always are that you should acknowledge.

Response: The discussion section revised and our last paragraph before the conclusions has been arranged as our limitation paragraph.

The text on lines 298-310 and 311-323 are essentially repetitions. Please revise and combine this text.

Response: These two paragraphs were revised and combined.

Lines 345-348, well, you mention the issue of high variance between subjects that causes high ICCs on lines 385-386. Please discuss your actual data a bit more with your results.

Response: The discussion was rearranged in line with your comment.

Round 2

Reviewer 1 Report

The biggest issue with the manuscript is how you are calculating strength values from the isokinetic contractions. When you are assessing strength, it is important to be able to know what the onset of torques are, and what the thresholds are for onset. To get numbers from the machine doesn't allow for a direct understanding on how your data is being calculated (filtered). Strength and force manuscripts typically include references, or state the exact filters that are using when assessing strength/force. 

Author Response

Reviewer 1

Reviewer Comment: The biggest issue with the manuscript is how you are calculating strength values from the isokinetic contractions. When you are assessing strength, it is important to be able to know what the onset of torques are, and what the thresholds are for onset. To get numbers from the machine doesn't allow for a direct understanding on how your data is being calculated (filtered). Strength and force manuscripts typically include references, or state the exact filters that are using when assessing strength/force.

Authors Response: We contacted the company in line with this referee criticism. The firm told us that the power values from isokinetic contractions were stuck with a torque sensor. When evaluating the power, he answered the question of what the onset of torques and what their thresholds mean for starting by specifying the characteristic of the torque sensor (0.25% FS). While taking numbers from the machine, we added how the data is calculated (filtered) to the text as in the following sentences, in line with the company's statements.

‘Raw data was collected with 200 Hz sampling frequency and 6th order 200 Hz low pass Butterworth filter was applied by the device. The company reported that the torque measurement accuracy of the device is 0.25% FS.’

Reviewer 3 Report

Please see attached.

Comments for author File: Comments.pdf

Author Response

Reviewer 2 

My old comment: The authors aimed to evaluate the intra-rater and inter-rater reliability of two IsoMed-2000 dynamometers for knee flexor and extensor peak torque measurements in a concentric test. The dynamometers had been used to different extents and referred to as “old” and “new”. They evaluated 37 young (aged 15-19 years) weightlifters (12 female, 25 male) twice on each dynamometer. Their principal findings were that the old dynamometer measured higher peak flexion torques but lower peak extension torques than the new dynamometer. The authors also present high intra-class correlation coefficients (ICCs) for repeated testing on the same dynamometer and between dynamometers. However, I expect high ICCs since the participants varied due to sex, age, and body mass, resulting in a relatively small within-individual variation compared to between-individual variation (the authors seem not to have normalized the peak torque values to body mass).

Response: The body weight of the athletes is manually recorded on the Isomed 2000 device before measurement. This device uses this manually entered value to normalize PT to body weight. We use both PT and PT / W values in our studies. However, in this study, the ICC values between the two measurements and the two devices were not shown in the tables because both PT and PT / W has the same ICC and the number of tables was large.

My new response: Since body mass varies between individuals, you change the inter-individual differences in data after normalizing to body mass. This results in different ICCs with the absoluteagreement two-way mixed-effect model, which was my original point. As such, it is expected that you will get high ICCs when not normalizing the data, and, therefore, these results are not that interesting. However, I can understand that you want to analyze the non-normalized data since there is no reason to believe that the participants’ strength would change between the testing sessions as a result of training, while body mass might vary and thus introduce noise in the analyses if normalizing. Nevertheless, very high ICCs are still expected, and I would recommend toning down these results and instead focusing on discussing the SEMs and limits of agreement in the Discussion since those results better reflect the story and the questions you construct throughout the Introduction. I think it is interesting to discuss the size of the actual offset in force that an older and more used dynamometer provides since this result has an immediate practical application.

Authors New Response: We have re-done all the analysis with the normalized peak torque data. Thus, our all results changed. After we normalized the data, we had different results and also lower ICC values. Thus, thank you for your valuable comment. We discussed the results on ICCs, SEMs, SDC95s, SDC%, and 95% limit of agreements.

My old comment: The study appears well-executed, and it is evident that the authors have invested substantial time to execute this study (also including a habituation testing session which is good but rare).

My old comment: I like that you wanted to investigate the observation of different measurements further (lines 60-68) and publish these results for others' benefit. However, by only including human testing and presenting peak torque values with ICCs, SEMs and limits of agreement, this study is not designed to explain why a well-used dynamometer would provide different measurements compared to a less-used dynamometer. The main recommendation not to compare strength data between dynamometers is, to be frank, relatively accepted. As such, this paper comes off as an exploratory case-control study that compared a well-used and a less-used dynamometer. Further comparisons of other dynamometers with different histories of usage would provide a stronger case in generalizing the observations to dynamometers in general.

My old comment: Further, on lines 68-70 on page 2 you write: “In the literature, we did not find any studies that reveal the relationship between the duration of use of isokinetic dynamometers and their measurement output”. I wonder whether this also includes information from companies that produce dynamometers. These companies should have a lot of such data and explanations as to why you observed different measurements, and I wonder whether or not such information is available (I have not checked, so this is genuine speculation). Have you contacted the D&R Ferstl GmbH company about this, or do you know of any material concerning calibration related to this? You mention at line 398 that you contacted sellers but were given no information? If such information is available, the main finding of this study is, unfortunately, superfluous. But, if no such information is available, I believe that your results are worthwhile publishing, seeing that this study seems to be the first of its kind and comes with an apparent clinical recommendation. In that scenario, this study's publishing hopefully spurs others in publishing similar studies, which ultimately results in a better knowledge of how strength should be evaluated in dynamometers. If there is a clear explanation of different peak torque measurements between dynamometers, this knowledge should be out there.

Response: Yes, we talked to D&R Ferstl GmbH company through the distributor firm of the device. He suggested that we test the weight meter of the device using ten kilograms of weight. However, there was no method standardization in this test application; The following ten kilograms were asked to be placed in a bag and placed on the longest dynamometer attachment with a rope. It was reported that if there was no difference between the three measurements, the device would give an accurate measurement. In this measurement, the force arm, the length and oscillation of the tied bag were not taken into account, which made us doubt about the calibration. Later, we formed the hypothesis of this study with the information from the athletes that 'he could not obtain the PT value he obtained in one device, in the other device'.

Also, we have not received information from the company about how the measurement outputs have changed over the years. However, when we received the device, the company suggested regular annual maintenance and repair of the device. We have the annual maintenance of the devices made by the company.

My new response: Ok, good that you talked to them.

I can understand the problem with having a standardized method to compare force data between the dynamometers for isokinetic testing. However, you could easily check if you get the same force data between the two dynamometers for weight measurements by locking the lever arm in certain angle positions and comparing the results (e.g., isometric testing). Such data would add additional results that directly answer the reason for the different concentric versus eccentric measurements (that you currently discuss on lines 391-393). 

Regarding your response, I think that one or two short sentences that explain the problems with a standardized method to evaluate if the dynamometers show different values should be added to the paper. This addition would further motivate the design you chose and would very nicely complement the standardized isometric evaluation that I suggested above.

Authors New Response: After doing the statistical analysis again, our results have changed. We've moved away from the results we had predicted as the previous weight measurement error.

Title

My old comment: I find the current title a bit boring and generic. Also, I think the description of “The Much and Little Used” is unclear. I prefer a title that describes your main finding of different peak torques values between a well-used and a less-used dynamometer since the high reliability (ICCs) is given considering the relatively small within-person variation compared to the between-subject variation.

 

My new response: If you want to stick with this title, which I still think is too generic and boring, I
would prefer to change “the well-used and less-used” to “a well-used and a less-used” since more
objective. The same should be applied to the aim on line 83 and in the Discussion on line 323.

 

Authors New Response: The title topic has been considered, but no more suitable title describing the method could be found. For this reason, the current title was edited in line with the referee suggestion.

Introduction

My old comment: Lines 46-50, long sentence, divide into two.

Response from Authors: The sentence divided into two.

My new response: Ok.

 

My old comment: Lines 56 and 57, use active voice instead of “It is stated”. Response from Authors: We used it as an active sentence.

My new response: Ok, good.

My old comment: Line 60, consider a new section here, starting with “In our clinic”. But add another final sentence to appropriately finish the prior section.

Response from Authors: A new section starting with “In our clinic” created. And another final sentence added to the prior section.

My new response: Ok, I think this structure results in an easier reading.

However, in line with my previous comment about “It is stated”, your added final sentence on lines
60-64 also includes “it is stated”.
“In these studies, evaluating the reliability of strength and endurance measurements of knee flexor and
extensor muscles using isokinetic dynamometers, it has been stated that the strength and endurance
tests of the knee flexor and extensor muscles were generally very wide with the results obtained (ICC:
0.42-0.89) [11, 12].“

Shorten this sentence by eliminating repetition and simply write what the studies found, e.g., “These authors found that the between-session reliability of strength and endurance testing of the knee flexor and extensor muscles substantially varied (ICC: 0.42-0.89) [11, 12].” 

(I have not read these studies, so my suggestion is only an example. It may be inaccurate). Also,
connect the sentence better to the previous sentences on lines 55-60 since it is a bit off.

Authors new response: It is corrected as commended.

 

My old comment: Lines 66-68, since your results show different knee flexion and knee extension differences between the dynamometers (one higher, one lower), you should clarify the peak values you mean.

Response from Authors: We added to the sentence the ‘knee flexion and knee extension’

 

My new response: The problem with the sentence, as I read it, is that I interpret it such that the athletes cannot achieve or exceed the knee flexion and extension peak torque value on one dynamometer (e.g., the older) compared to the other. Meaning, one dynamometer showed lower values for both knee flexion and extension peak torque. Since you found different results between the dynamometers for knee extension and flexion peak torque, it may be clearer to write something like:

“As a result of our observations, we saw that our athletes demonstrated higher knee flexion peak
torque (PT) values on one dynamometer but higher knee extension PT values on the other.”
I think this makes more sense in relation to your results, and it should reflect your previous
observations.

Authors new response: It is corrected as commended.

 

My old comment: Lines 80-83, you write “In light of this information”, but to what information are you referring? You describe no previous information that motivates any hypotheses apart from your own experience of underestimation. Also, the “metal fatigue” should be described in the Introduction if included. If you do not state any previous information to support your hypothesis, it sounds post hoc and makes the reader (at least me) skeptical. Please extend the Introduction a bit to allow the reader to understand and (in theory) agree with your hypothesis.

Response from Authors: “In light of this information” section corrected and “metal fatigue” is deleted.

My new response: Ok, good. Regarding your hypothesis on lines 85-87, remove “is” since past tense
and write “We hypothesized that…” or “Our hypothesis was that…“.
Also, revise “… , and will not give the same muscle strength measurement outputs with new
dynamometer.”

to (or similar) 

“… , such that an older and well-used dynamometer will not provide the same measurement output as
a new dynamometer.”

Authors new response: It is corrected as commended.

 

Materials and Methods

My old comment: Line 91, my preference is to write out the exact number of females and males. It is easy to see the ratio of females and males for your total sample.

My old comment: Line 92, please use the same number of significant figures for age to be accurate.

Response from Authors: It was corrected according to your advice.

My new response: Ok, good.

 

My old comment: Table 1, weight should be body mass since in kg, and the data for height should be in m, not cm. Also, use the same number of significant figures across the data for each variable.

Response from Authors: It was corrected according to your advice.

My new response: You only corrected half of it. The height is still reported in cm, not m, and I doubt
that you accurately measured height with 1/10 cm correct. Even though this is a very minor detail, I
would change to m and round up to cm	.

Authors new response: The height values were reported in meter according to your advice.

 

My old comment: Lines 100-101, you write “…to have participated in an international competition for at least 3 years.”, did the 15-year-old weightlifters start international competitions when 12 years old? Did you mean “…competition within the last 3 years.”?

Response from Authors: It was corrected as “…competition within the last 3 years.”

My new response: Ok, good.

 

My old comment: Line 101-102, did you use exclusion criteria of quadriceps- and hamstring-specific resistance training for internationally competitive weightlifters? What did the weightlifters do during training? Please explain since this does not make sense to me.

Response from Authors: That means the isokinetic exercise training, it was also changed in the text.

My new response: Of course, good.

 

My old comment: Line 103, I do not understand “throwing” here; please clarify.

Response from Authors: This sentence was rearranged.

My new response: Ok, good.

 

My old comment: Line 111 and 114, please provide the same information of how old these dynamometers are that you stated in the Introduction on lines 62-63. I think this information adds value here and should be easy to find.

Response from Authors: It was added in line with your recommendation.

My new response: Ok, good, but revise the sentence and simply state the age, e.g.,:

“The less-used dynamometer had six months of usage and the well-used dynamometer had three years
and nine months of usage. Both dynamometers had annual service maintenance.“

Authors new response: It is corrected as commended.

 

My old comment: Line 123, when was the familiarization test performed in relation to the actual tests?

Response from Authors: Just before each test.

 

My new response: Ok, I still think that you need to add additional information about resistance during
the familiarization to allow for study replication.

Authors new response: The following sentence is included to the text ‘Our aim with familiarization is to make the athletes to understand the resistance of the angular velocity to be tested’  

 

My old comment: Lines 124-127, first, you state that participants were required to refrain from any vigorous activity for at least 8 hours before testing, and in the following sentence 48 hours before testing. Which is it?

Response: Since the isokinetic tests were taken in the morning hours, they were asked to rest the night before and not to take any stimulants. In addition, 48 hours before the tests were carried out, routine training was allowed, and heavy training or competition was not allowed.

My new response: Ok, more clear now.

 

My old comment: I think that lines 151-166 and 176-178 should be combined and revised to blend with the information in the current heading 2.3. Procedures, since all of it describing the procedures. The information in lines 166-175 can have a heading on its own, e.g., 2.4. Outcome variables, or similar. These outcome variables could also be presented in a table if room for this. Response from Authors: As you indicated, combining and opening new headers were done.

 

My new response: Basically all text in the section 2.4. Equipment and Materials are procedures, apart
from information about the bike. Since you already have described the dynamometers in detail, I don’t
think you need an Equipment and Materials heading at all. I suggest removing this heading and the
Evaluation of isokinetic muscle strength heading and combine all text in the 2.3. Procedures heading,
with one section for the warm-up and another section for the isokinetic testing (since correct timeline).

I like the 2.5 Outcome variables section, although I would recommend presenting it similar to:

“The outcome variables extracted and used in statistical analyses were: …mention all of them…”

Authors new response: It is revised as commended.

 

My old comment: I lack information that describes data analysis. You do not mention: if you normalized the data (I guess not when looking at the actual data you present in Tables 3-6); if you processed the data with any filter before extracting peak values; if you used the individual average peak value for test-retest analyses. Please add this information.

Response from Authors: There were outliers in our initial data and we removed them to get a normal distribution. Because we compared the dynamometers and trials, we applied Shapiro-Wilks Test separately for the trials and dynamometers. The final data had a normal distribution (p > 0.05).

My old comment: The detail is given in Section 2.5, paragraph 3.

Response from Authors: “Before the analysis, the outliers were removed from the data. Each measure was evaluated for normality with the Shapiro-Wilks Test and demonstrated that all data had a normal distribution (p > 0.05).”

My new response: You need to provide details of outliers, not just state that outliers were removed. How did you classify outliers? Outliers in what type of data; PTs or between-dynamometer differences in PTs or other? How many were removed? Etc.

Authors new response: We would like to thank to the reviewer for his/her valuable comment. After the comments, we have re-done all the analysis with the normalized peak torque data.

We used the Shapiro-Wilks Normality Test for the differences between the trials and dynamometers of each measure and noticed that some of the differences were not normally distributed. We had 42 athletes involved at the beginning of the study but 4 of their measures were incomplete. We applied the box-plot was used to detect the outliers by considering the differences between the trials and dynamometers of each measures. One of the athletes was removed because his calculated differences were out of the Q1-1.5*IQR or Q3+1.5*IQR in most of the measures.

We put the following sentences to text. 

“42 athletes were involved at the beginning of the study. Four of their measures were incomplete. By considering the differences between the trials and dynamometers of each measures, the box-plot was used to detect the outliers and one of the athletes was removed because the calculated differences were out of the Q1-1.5*IQR or Q3+1.5*IQR in most of the measures. Each measure for 37 athletes was evaluated for normality with the Shapiro-Wilks Test and demonstrated that all data had a normal distribution (p > 0.05).”

 

 

My old comment: Lines 180-182, your description of the power analysis does not provide any information that allows the reader to understand what you did and, therefore, do not really add anything here. A more accurate description should include additional information: What type of power analysis? For what statistical test is the power analysis for? What is your dependent variable? What is your independent variable? How did you estimate the effect size?

Response from Authors: We changed and did again the power analysis and wrote it below another reviewer comment.

My new response: Please add “paired” to Student’s t-test such that it reads “two-tailed paired
Student’s t-test”.

The information is better, but you should state where you got the effect size of 0.50 from. 

Authors new response: We wrote as “two-tailed paired Student’s t-test”. We added Cohen (1988) reference for effect size.

 

My old comment: Lines 197-202, I am not too fond of the SEM calculation where the ICC value is included for a paper such as this one. Since you want to have a proper estimation of the SEM, I think it is better to calculate it without the ICC, thus as the average within-person standard deviation from the trials included in the analyses. This calculation is also described in the study by Weir. The reason for this argument is that your sample (males and females of different ages and body mass) is sure to result in relatively small within-individual variation compared to between-individual variation, therefore very high ICCs. I would not want these high ICCs (basically by study design) to affect the SEM calculations when trying to understand differences in actual data between the two dynamometers. Please consider the basic calculation for SEM as the average within-person standard deviation from the analyses' trials.

Response from Authors: “For assessment of absolute reliability standard error of measurement (SEM) and smallest detectable change (SDC) were used. SEM was calculated as and SDC was calculated as SDC95=SEM×1.96√2, where 1.96 is the z score for a 95% CI [18].”

 

My new response: Ok, good!

However, you still need to define what the SDdiff is, so readers don’t have to search in the Weir study
(it is the standard deviation of the difference scores for the two trials).

 

Authors New Response: The definition of SDdiff is added as follows. “For assessment of absolute reliability standard error of measurement (SEM) and smallest detectable change (SDC) were used. SEM was calculated as  where  was the standard deviation of the difference scores for the two trials or two dynamometers.”

 

My old comment: Line 202, the references 19 and 20 have used different populations than weightlifters, so the classification of “acceptable” and “excellent” here is redundant. I do not think these classifications adds anything to the paper since the data stands by itself and should be discussed instead. Consider removing the classifications of “acceptable” and “excellent” throughout the paper.

Results

 

My new response: Ok, I see that you removed the text I referred to. You still have one “excellent”
(line 223) and one “acceptable” (line 237) in the Results section, and I think you can remove these
since they don’t add anything. It is better to discuss the size of these data instead.

Authors New Response: Thank you for your comment. These words are removed.

 

My old comment: Tables 3, 4, 5, 6, add units to all data that have units. Also, I do not understand what the significant ICCs mean. Significant for what? Further, the discrepancies in 95% LoA and Figures 1-4 warrant discussion since unclear at the moment (95% LoA do not contain the value 0, but the plots tell another story).

Response from Authors: In statistics, first the significance of a coefficient is investigated. If the coefficient is statistically significant, then it is interpreted. In Tables 3, 4, 5, 6, the significance of the interclass coefficient (ICC) is investigated and the coefficients are found statistically significant (p<0.01).

95% LoA of Table 5 and 6 are re-calculated. Also, the following interpretations are added.

The Bland-Altman plots showed no systematic bias for EXDMPT180 and EXNDMPT180 (see

Figures 1 and 2). According to Figure 3, it could be said that there is a proportional bias between the

FLNDMPT60 and FLNDMPT180 measures of two dynamometers (FLDMPT60 r = 0.333, p = 0.044; FLDMPT180 r = 0.401, p = 0.014). According to Figure4, it could be said that there is a proportional bias between the FLDMPT60, FLNDMPT60 and FLNDMPT180 measures of two dynamometer (FLDMPT60 r = 0.333, p = 0.046; FLNDMPT60 r=0.407, p=0.012; FLDMPT180 r = 0.438, p = 0.007).

My new response: In statistics, the analyst often tests an assumption with data and defines a null hypothesis and an alternative hypothesis. What did you test with the ICC that warrants a p-value? How do you interpret a significant and non-significant ICC? 

The SPSS output provides p-values for the “F Test with True value 0”, but if you don’t describe anything about this test, then I think you should remove the information about the statistical significance for ICCs. As it is now, you don’t provide any context for this information, and, therefore, it seems redundant. 

Regarding the Bland-Altman plots for systematic bias and the result on lines 288-294, you need to
describe this further. Please clarify in the statistics section how you evaluated systematic bias.

 

Authors New Response: Thank you for your comment. We have removed the “significance” part of the ICC.

We used the paired t-test results to assess systematic bias between trials and dynamometers. 95% LoA also gave the same information. The sentence below is added to Section 2.5, Paragraph 5.

“The paired t-test and 95% LoA were used to assess systematic bias between trials and dynamometers. The regression between the difference of trials or dynamometers and their means were used to detect the proportional bias.”

 

Discussion

My old comment: In general, I think that your actual data need further discussion, as do the analyses. Also, you have no information on study limitations, which there always are that you should acknowledge.

Response from Authors: The discussion section revised and our last paragraph before the conclusions has been arranged as our limitation paragraph.

 

My new response: The Discussion is still relatively poor and needs to properly discuss the data and
possible reasons for them in relation to the aims. It seems like a draft at the moment.

The text about ICCs is difficult to understand, and it seems, in part, that the authors have misunderstood (or fail to communicate) the related literature about ICCs (e.g., lines 378-387). I want further discussion about the actual size in Newtons that differed between the dynamometers and how this difference relates to testing. If you want to have a take-home message for the reader about dynamometers and testing, this is something that really should be adequately discussed.

Authors New Response: As the statistical analysis was reconstructed and interpreted in line with the recommendations of the referees, most of the discussion part was also updated.

My old comment: The text on lines 298-310 and 311-323 are essentially repetitions. Please revise and combine this text.

Response from Authors: These two paragraphs were revised and combined.

My old comment: Lines 345-348, well, you mention the issue of high variance between subjects that causes high ICCs on lines 385-386. Please discuss your actual data a bit more with your results.

Response from Authors: The discussion was rearranged in line with your comment.

Author Response File: Author Response.docx

Round 3

Reviewer 1 Report

It is suggested that you refer to Jenkins, Palmer and Cramer 2013. Isokinetics and Exercise Science, "Comparisons of voluntary and evoked rate of torque development and rate of velocity development during isokinetic muscle actions. The methods section gives a good elaboration on how strength/velocity was calculated/assessed. 

 

The information (PT) that you are presenting in the present study, do not directly tie back to how strength is assessed. The present investigation is reporting on numbers that the dynamometer is creating, when traditionally researchers assess strength output with custom softwares, and filters on how onset was/is determined.  I understand that you are trying to look at the differences in accuracy with 2 different dynamometers across usage time. Do you know if both dynamometers use the exact same method for assessing torque?

Author Response

Reviewer Comment: It is suggested that you refer to Jenkins, Palmer and Cramer 2013. Isokinetics and Exercise Science, "Comparisons of voluntary and evoked rate of torque development and rate of velocity development during isokinetic muscle actions. The methods section gives a good elaboration on how strength/velocity was calculated/assessed. 

Authors Response: Jenkins, Palmer and Cramer 2013. Isokinetics and Exercise Science, "Comparisons of voluntary and evoked rate of torque development and rate of velocity development during isokinetic muscle actions referred. And the method section was cited on how to evaluate the strength suitable for our study.

 

Reviewer Comment: The information (PT) that you are presenting in the present study, do not directly tie back to how strength is assessed. The present investigation is reporting on numbers that the dynamometer is creating, when traditionally researchers assess strength output with custom softwares, and filters on how onset was/is determined.  I understand that you are trying to look at the differences in accuracy with 2 different dynamometers across usage time. Do you know if both dynamometers use the exact same method for assessing torque?

Authors Response: As the reviewer have stated, the information (PT) we provide in this study is not directly related to how power is evaluated. This research traditionally reports the numbers generated by the dynamometer when researchers evaluate the strength output with special software and determine the filters on how / to determine the baseline. We tried to look at the accuracy differences with 2 different dynamometers during the usage period. We know that both dynamometers use the same method to evaluate torque. We talked to D&R Ferstl GmbH to confirm and they answered us 'Yes torque sensor and signal conditioning hardware and analogue Filters are equal also the used software'. We also added this information to the text.

Round 4

Reviewer 1 Report

Please include a limitations paragraph and elaborate in this paragraph, that the present investigation looked at torque output between two dynamometers that are the same brand, but there was no control for torque signals or software creation to detect onset. 

Further elaboration can be given about the present study being applicable to clinical testing (physiotheraphy clinics, PT clinics), but may not successfully cross-over into performance based research. 

 

Author Response

Reviewer Comment: Please include a limitations paragraph and elaborate in this paragraph, that the present investigation looked at torque output between two dynamometers that are the same brand, but there was no control for torque signals or software creation to detect onset.

Authors Respond: We added limitations paragraph and we added ‘Additionally, this study investigated torque output between two dynamometers that are the same brand, but there was no control for torque signals or software creation to detect onset’ to the limitation paragraph.

 

Reviewer Comment: Further elaboration can be given about the present study being applicable to clinical testing (physiotherapy clinics, PT clinics), but may not successfully cross-over into performance-based research.

Authors Respond: The suggested explanation has been added to the conclusion section.