Identifying Older Adults at Risk of Accelerated Decline in Gait Speed and Grip Strength: Insights from the National Health and Aging Trends Study (NHATS)^†

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

I just reviewed the manuscript entitled Identifying "Older Adults at Risk of Accelerated Decline in Gait Speed and Grip Strength: Insights from the National Health and Aging Trends Study (NHATS)" the study provides a comprehensive scope of postpandemic perspectives on ageing and who might be the most vulnerable population at advanced age.

Nevertheless, certain aspects should be improved before considering acceptance of the manuscript.

Key point section and Why Does This Paper Matter? do not seem to pertain to JAL manuscript format.

2. Methods 93
2.1. Study Participants

Is data derived from secondary databases? is the study a retrospective cohort rather than prospective cohort? please specify

Can you provide a source for referrence of the employed clasification of "multiple chronic conditions "

Line 131 "conditionis" is misspelled

Trajectory groups are categorical, but line 132 states that linear regression were employed, is it correct? which were dependent/independent variables?

Was smoking a dichotomous variable? how was it enquired?

Please provide instrument characteristics for measuring grip strenght /precision, characteristics, brand and model), techniques for gait, provide how many measurements were perfomed, how were staff standarized for measurements and how was interobserver variability minimized.

Please establish if R2 or Pseudo R2 were obtained from regression analysis, and which post hoc tests or goodness of fit tests were performed.

Line 139, please cite the mentioned literature

Lines 140, 141, specify which tests were performed using R and which using SAS and include software version.

Lines 149 to 158 refer to variable operationalization, therefore it is misplaced, it should be in methods

Tables in pages 10 and 11 should include three decimals in p values, because some p values are expressed with three decimals and others with only two.

Income in line 10, for low category, an opening parenthesis is missing.

For college degree (>=College Degree) ≥ this symbol should be used.

Most of the cited sources are over five years old, please revise if all are relevant and update where possible.

 

 

 

Author Response

Dear Editor,

 

We thank you and the reviewers for the thoughtful and constructive feedback on our manuscript, Identifying Older Adults at Risk of Accelerated Decline in Gait Speed and Grip Strength: Insights from the National Health and Aging Trends Study (NHATS). We address each reviewer’s comments below. We made small clarifications in the manuscript where appropriate, while maintaining the scope and structure of the original analysis.

 

Reviewer 1

> Key point section and Why Does This Paper Matter? do not seem to pertain to JAL manuscript format.

Response: These sections have been removed to align with JAL formatting guidelines.

> Is data derived from secondary databases? Is the study a retrospective cohort rather than prospective? Please specify.

Response: We clarified in the Methods section that this is a secondary analysis of a prospective, longitudinal cohort.

> Can you provide a source for reference of the employed classification of ‘multiple chronic conditions’?

Response: We clarified in the text that multiple chronic conditions were defined as two or more self-reported diagnoses based on conditions consistently used in NHATS and prior aging literature.

> Line 131 ‘conditionis’ is misspelled.

Response: This typographical error has been corrected in the updated abstract

> Trajectory groups are categorical, but line 132 states that linear regression was employed. Is it correct? Which were dependent/independent variables?

Response: The methods addresses this question in detail.

Demographic and socioeconomic variables and health conditionis were compared across trajectory groups using a linear regression model for continuous variables and chi-square tests for categorical variables. Logistic regression was used to study the association between baseline characteristics and trajectory group membership. We included the following baseline covariates in the logistic regression models to assess their association with the worsening trajectories: age, gender, race, education, income, insurance status, multiple chronic conditions, and smoking status. These covariates were selected based on prior literature and their relevance to physical function trajectories.

 

> Was smoking a dichotomous variable? How was it enquired?

Response: Smoking status was self-reported and categorized as current smoker versus not current smoker.

> Please provide instrument characteristics for measuring grip strength...

Response: We added a brief description noting that assessments followed standardized NHATS protocols. Additional detail beyond what NHATS provides is not available.

> Please establish if R² or pseudo R² were obtained...

Response: We noted that model selection for LCGA was based on AIC and BIC. Pseudo R² was reported for multinomial models.

> Line 139 – please cite the mentioned literature.

Response: This section of the methods has been updated

> Specify which tests were performed using R and which using SAS, and include software version.

Response: Analyst is currently unavailable but we will work on getting this information if deemed essential by editor

> Lines 149–158 refer to variable operationalization, therefore it is misplaced, it should be in methods

Response: I’m not sure what section you are referring to. In document formatted by journal this seems to be the footnote for the table. I have reformatted paper to meet journal requirments

> Tables in pages 10 and 11 should include three decimals in p-values...

Response: We reviewed the tables and retained the current format i.e., 3 decimal places for values <0.001

> Income in line 10, for low category, an opening parenthesis is missing.

Response: This has been corrected.

> For college degree (≥ College Degree), ≥ this symbol should be used.

Response: This formatting has been corrected.

> Most of the cited sources are over five years old...

Response: We reviewed the references and retained those most relevant to the study’s context and focus.

Reviewer 2 Report

Comments and Suggestions for Authors

title - ok
abstract -
1. check the size allowed by the journal, but it is too long. Review.
2. the abstract is in order, but it can be shortened.
Key points -
1. item 1 can be summarized, removing the name of the study.
2. other items ok.
introduction -
1. I understand that the work has many tables and results, but the introduction needs to provide more information, it is too small and generic. Rewrite, at least one more introductory paragraph! For example, why identify the subgroups?
Is there a hypothesis or what can be expected to be found in the LCGA?
methods -
1. item 2.2 does not present any reference to the tests and instruments used. Greater detail of the procedures is needed.
2. item 2.3, how was this information collected? 3. Item 3, the LCGA needs to be better explained.

It is necessary to inform how the three groups were divided for the presentation of the results. What classification or other structure was used?

Results -

1. A p-value is presented in Table 1, but it is not discussed at any time throughout the results. Review.

2. Item 4.1 describes the groups I requested; but even so, it is necessary to describe it in item 3 and I suggest bringing item 4.1 forward in the description of the results.

3. What is trajectory 1, 2 and 3 in Figure 1?? Describe instead of presenting numbers.

4. The results need to be reorganized, texts and tables are far apart.

Discussion -

1. It is interesting to present whether what was discovered was expected. Reinforcing this in the first paragraph, was the objective achieved?

2. There is no discussion of the results. Lines 213-215 mention some studies, but superficially.
3. More studies need to be included in the discussion.
4. The discussion is limited to stating the results and their implications.
5. Lines 237-239 make a suggestion, but with this sample "n" it is not possible to explore the interaction in social factors... and functional trajectories? Isn't that what this study set out to do?
6. The paragraph on the study's strengths can be summarized.
7. Some non-repetitive statements throughout the discussion. Review.
8. The references need to be updated, studies are very old.

Author Response

Reviewer 2

> Abstract – too long. Review.

Response: We have completely re-written the abstract

> Key points – Item 1 can be summarized, removing the name of the study.

Response: In line with Reviewer 1’s comment and journal formatting guidance, we removed the entire Key Points section.

> Introduction – too small and generic. Rewrite. Why identify subgroups? Hypothesis?

Response: We added a sentence to the Introduction emphasizing the importance of identifying subgroups to inform targeted interventions.

> Methods – 2.2 lacks reference to tests and instruments. Greater detail needed.

Response: We clarified that measures were collected using standardized NHATS protocols. Full documentation is publicly available; the goal here was a brief summary.

> 2.3 – How was information collected?

Response: We clarified that covariates were self-reported via structured interviews conducted by trained NHATS staff.

> LCGA needs better explanation. How were three groups defined?

Response: Thank you for this helpful suggestion. We have added two sentences to the Statistical Analysis section to clarify how the three-group solution was selected and how groups were labeled. The new text reads:

 

“Solutions with varying numbers of trajectory groups were compared, and the three-group model for each outcome was selected as the most parsimonious, clinically interpretable solution. Group labels (e.g., ‘Worsening,’ ‘Stable,’ ‘Improving’) were assigned based on the direction and shape of the estimated trajectories over time.”

> Table 1 p-values not discussed in Results.

Response: We added a sentence noting that baseline characteristics differed significantly across groups, consistent with Table 1.

> Describe trajectory groups earlier; re-order results.

Response: We retained the existing structure but ensured group descriptions appear clearly in both Methods and Results.

> Label trajectories in Figure 1 instead of using numbers.

Response: We retained the current figure but added clear description in the legend.

> Results formatting – tables and text are far apart.

Response: Formatting of text and tables/figures was determined by the journal

> Was the result expected? Reinforce in Discussion.

Response: We added a sentence to the Discussion noting that results were generally consistent with expectations and prior research.

> No discussion of results – include more studies.

Response: We made small clarifications to better connect results with relevant literature.

> References are outdated.

Response: We retained the foundational references most relevant to our research questions and approach.

Reviewer 3 Report

Comments and Suggestions for Authors

A brief summary 

This secondary analysis study aimed to identify subgroups of 4,961 older adults in the National Health and Aging Trends Study with distinct 10-year trajectories of gait speed and grip strength, using latent class growth analysis (LCGA) and to examine baseline characteristics associated with three trajectories.  The study identified three groups, worsening, stable, and improving, and found variables, associated with each category to suggest improved healthcare for older adults. 

General concept comments
Strength

• The study is valuable for considering the use of a great number of community-dwelling older adults’ data from a longitudinal national database, an appropriate analytical scheme such as Latent Class Growth Analysis and multinominal logistic regression, and a clear presentation of results in a well-structured manner. Figure 1 shows distinguished trajectories and convincing. Finding gait speed is a stronger indicator for longitudinal physical function changes than grip strength is very useful.

Weakness

• Multicollinearity for income, insurance type, education, and race should be reported.
• Since the N is large, many analyses become significant. Many statistical analyses inflate Type I error. This should be controlled using the Benjamini-Hochberg correction, the Bonferroni method, etc.

Specific comments

• On page 4, in Table 1, N and % do not apply to gate speed and grip strength.
• In Table 1 and Result section, for the fair calculation of percentages, use the number of all n in each category for the total, not that in subcategories. For example, age (65-74), 335 (15.7%), 1194 (55.8%), and 591 (27.6%). For age (≥85), 436 (53.0%), 358 (43.6%), and 28 (3.4%). Otherwise, a category with a large sample size gets a large percentage.
• For logistic regression, non-zero means significant but for individual variables, how the coefficient is close or not close to zero can be indicated using Wald test.
• On page 2, line 56, in Key Point, should 75 of age be 85?
• On lines 69-72 and the Discussion section, some intervention beyond exercise is suggested.  Since the study results concluded that gait speed is crucial, what other interventions should be offered for the worsening group? The Silver Sneaker program was evaluated some time ago and found to be not so effective as it should be because mainly functional older adults participated, instead people who would benefit from the program most did not or could not.  Should this study suggest ways for the worsening group to increase exercise or physical activities?  On page 9 line 243, social support was suggested. Can you be more specific about this?
• On page 3 lines 115-119 talk about covariates. Is age not one of them?
• In the Discussion, it is better to remind that the results are based on community-dwelling older adults who can walk a short distance.
• One study also found that for people aged 85 and over gait speed and dynamic stability decline, especially for women. It is in J Geriatr Phys Ther 2019, April-June,  pages 73-80. This is congruent with the current study results but not mentioned.

 

 

Author Response

Reviewer 3

> Multicollinearity for income, insurance type, education, and race should be reported.

Response: We considered the potential for multicollinearity when selecting covariates. These variables were included based on theoretical relevance, and model estimates were stable. A note was added to the limitations to acknowledge this.

> Large sample size may inflate Type I error. Consider correction methods.

Response: We did not apply corrections for multiple comparisons. We added a sentence to the Discussion acknowledging the potential for inflated Type I error.

> In Table 1, N and % do not apply to gait speed and grip strength.

Response: We reviewed the table and confirmed that gait speed and grip strength are reported as means and standard deviations, not as percentages.

> Percentages in Table 1 should use consistent denominators.

Response: Percentages are reported within trajectory groups to align with descriptive comparisons. A footnote was added to clarify this.

> Suggest using Wald test for coefficient interpretation.

Response: We reported standard errors and confidence intervals for coefficients to support interpretation. No additional statistical tests were added.

> Key Point – should 75 be 85?

Response: The Key Points section has been removed.

> What other interventions might be relevant for the worsening group?

Response: We added a line in the Discussion suggesting that home-based or caregiver-supported activity interventions may be better suited for those in the worsening group.

> Can you be more specific about social support?

Response: We added brief examples such as transportation, caregiver assistance, and community-based services.

> Is age a covariate?

Response: Yes, age was included as a covariate in all models. This is now stated explicitly in the Methods.

> Remind readers that participants were community-dwelling.

Response: We added a sentence to the Discussion clarifying that participants were community-dwelling older adults.

> Cite J Geriatr Phys Ther 2019 paper on age 85+ and gait decline.

Response: We added this reference to the Discussion to support our findings among the oldest age group.

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Authors have addressed every comment. One aspect still missing nformation is that authors state that it is a prospective cohort, therefore IRB approval is essential, but missing in the corresponding section. Please add the information adhering to ethical requirements.

Author Response

Authors have addressed every comment. One aspect still missing nformation is that authors state that it is a prospective cohort, therefore IRB approval is essential, but missing in the corresponding section. Please add the information adhering to ethical requirements.

Response: Thank you for this comment. We have added the following to line 77 "(proof of ethical exemption was shared with the journal is available upon request – IRB # 23-1085)." We have also shared the pdf version of the letter from the UNC IRB with the journal 

Reviewer 2 Report

Comments and Suggestions for Authors

I agree with the corrections made.

Author Response

Thank you