Review: Neglecting Connection:
Associations Between Parental Emotional Support and Youth Mental Health in Four
Latin American Countries {under peer review}
Reviewer: Hyojin Im
Completed: 04-10-2026 15:54
Recommendation: Resubmit for Review
|
|
Yes |
No |
N/A |
|
Is the research question clearly defined? |
+ |
||
|
Are the methods appropriate and sufficiently detailed? |
+ |
||
|
Is the data analysis robust and replicable? |
|
+ |
|
|
Are the conclusions supported by the results? |
|
+ |
|
|
Is the manuscript well organised and clearly written? |
|
+ |
|
|
Are tables, figures, and supplementary material informative and necessary? |
+ |
|
|
|
Is the abstract an accurate summary of the study? |
+ |
|
|
|
Does the manuscript contribute meaningfully to the field? |
+ |
|
|
|
Is it relevant to the field of mental health or related disciplines that are connected to the scope of the Journal? |
+ |
|
|
|
Are ethical approvals and participant consents adequately described? |
+ |
|
|
|
Have competing interests, funding, and data availability been transparently declared? |
+ |
Comments for the authors:
This
manuscript examines an important question about parental relationships and
mental health among young people in Latin America and the Caribbean. Its
attention to both maternal and paternal relationships is useful, and the
national household surveys offer an opportunity to extend the regional
evidence. However, several outcome definitions are inconsistent, and some of
the conclusions do not match the reported results. Essential details about the
analyses are also missing. Please see suggestions as below:
- The abstract
states that alcohol associations were consistent across all four countries, but
Table 8 shows no statistically significant association in Haiti and
communication was not examined there.
- The
Introduction can establish the importance of youth mental health in the region
without ranking countries on this basis.
The opening prevalence comparison gives the impression that the cited
figures are comparable national estimates of mental disorders, but they come
from studies of different populations and outcomes.
- Family
Systems Theory offers a reasonable framework for considering these
relationships, but the current analyses do not directly assess family systems
or the cultural mechanisms. Familism and patriarchal norms were not measured,
either, so these ideas can be discussed as possible explanations informed by
prior research, with clear acknowledgment that this study does not test them.
Attention to variation within countries would make the interpretation more
convincing than treating each country as a uniform cultural context. Ex. broad
claims about Latino family life also need care when applied to Haiti.
- The Methods
should explain how the VACS sampling design was incorporated into the analyses.
There is no clear account of the survey weights, strata, or primary sampling
units used for the descriptive estimates and regression models.
- The sample
counts need correction. Ex. Colombia is reported as n = 2,433 in the Methods
and n = 2,705 in Table 1. Model-specific sample sizes are also missing. Table 1
should identify unweighted counts, weighted percentages where applicable, and
the denominators for each measure. Adding basic demographic characteristics
would make the analytic populations easier to understand.
- The
distinction between nonsuicidal self-injury (NSSI) and suicidal ideation needs
to be resolved before the findings can be interpreted. The Methods assign
thoughts about suicide to NSSI, and Table 3 is titled “NSSI (suicidal
thought).” Thinking about suicide is not a measure of NSSI.
- Please trace
each outcome back to the source questionnaire and specify which variable was
used in Tables 3-5. It would be helpful to provide the original item wording
alongside the analytic label, including the recall period and treatment of
questionnaire skips.
- The six
distress items appear to correspond to the Kessler K6, but the scale is not
named or referenced. The manuscript lists anxiousness, whereas the El Salvador
questionnaire includes hopelessness. I'm not sure whether this is a wording
error in the manuscript or reflects a departure from the original measure.
- Kessler’s
scoring guidance recognizes both the 0-24 and 6-30 ranges, so it would be
helpful to state the scoring procedure.
- Please
explain why a Poisson mean model is used for this bounded symptom score and
clarify the quantity being estimated. Under the reported model, an
exponentiated coefficient of 1.44 represents a ratio of expected distress
scores, conditional on the included covariates. It does not represent a 44%
higher probability of experiencing psychological distress. The language of
higher chances, risks, and incidence rates is therefore misleading for this
outcome. If the model is retained, report the estimates as mean-score ratios
and revise the interpretation accordingly. Adjusted mean scores or mean
differences would help judge the magnitude in the original units of the scale.
- The alcohol
measure also needs verification. The El Salvador VACS male and female
questionnaires use thresholds of five and four drinks, respectively, whereas
the manuscript describes four drinks for all respondents. Please report the
actual thresholds by country and sex and explain how the measures were made
comparable.
- The meaning
of the binary relationship variable is also unclear. The Methods describe
whether respondents had a relationship with a parent, but the Results sometimes
interpret this as having a stronger relationship. Similarly, the treatment of
parental absence also needs explanation (whether they were deceased, separate,
or unknown, for example). It's also unclear whether each parental indicator was
modeled separately or whether multiple indicators were included together.
- For the
ordered logistic analyses, please report whether the proportional-odds
assumption was evaluated and how any departures were handled. The
interpretation should make clear that these models concern the ordered
frequency categories of binge drinking. The tables should also identify whether
p values are overall tests for an exposure or tests of individual contrasts.
- There are
sevearal discrepancies in findings. Ex. the Honduras substance-use paragraph
begins with alcohol but then reports a 449% increase corresponding to the
drug-use odds ratio of 5.49 in Table 6. It also identifies maternal
communication as the only parental indicator associated with alcohol use, but
Table 7 shows other associations. The paternal-relationship in El Salvador,
Haiti distress results, etc. also don't match with Tables. Please check all narrative results against
the tables.
- The table
layout currently makes some estimates difficult to follow. Numerical values
should remain intact, and country and model headings should repeat when tables
continue onto another page. Table 8 needs explicit labels for the effect
measures and confidence intervals.
- Throughout
the Results, please report the point estimate with its confidence interval and
identify the actual reference category. Odds ratios should be described in
terms of odds, while the distress estimates need the score-based interpretation
discussed above.