Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

General comments

This study addresses an important and practical topic by evaluating knowledge acquisition and short-term retention following a Helping Babies Breathe course in a low-resource setting. The complete assessment of 38 participants at three timepoints and the item-level reporting are strengths. However, substantial concerns remain regarding the item-level statistical analysis, the interpretation of knowledge retention, the baseline–change correlation, and the reporting of ethical approval. These issues should be addressed before the findings can be interpreted reliably.

Major concerns

1. Potentially invalid McNemar test results

   The p-values in Supplementary Table 2 appear inconsistent with the reported marginal counts. Most notably, item 18 increased from 32/38 correct responses before the course to 37/38 after the course. This necessarily implies discordant pairs, and an exact McNemar test should therefore be calculable. The statement “Could not be calculated” is not statistically plausible.

   Several other results, including items 10, 14, and 15, also require verification. Please repeat all item-level analyses using the participant-level paired data and report whether the conventional or exact McNemar test was used, the numbers of discordant pairs, the statistical software and function, and any missing observations. The revised results should be cross-checked across the text, Figure 5, and Supplementary Table 2.

2. Multiplicity and interpretation of the item-level findings

   At least 36 item-level comparisons were conducted. The isolated p-value of 0.009 for item 16 may therefore represent a chance finding. For example, a simple Bonferroni adjustment would yield an adjusted p-value of approximately 0.324.

   If the authors elect not to adjust for multiple comparisons because these analyses are exploratory, the item-level findings should be explicitly presented as descriptive and hypothesis-generating. The delayed improvement in item 16 should not be attributed to the course without acknowledging alternative explanations, including repeated testing, subsequent clinical experience, and other educational exposure.

3. Knowledge retention cannot be inferred from a nonsignificant difference

   The median score decreased from 15 immediately after the course to 14 at three months, with p=0.46. Failure to detect a statistically significant decline does not establish retention or equivalence. Please report the paired changes, their distribution, an effect size, and preferably a confidence interval.

   Unless a retention criterion or equivalence margin was prespecified, statements regarding “partial knowledge retention” should be substantially tempered. A descriptive statement about the observed scores would be more appropriate.

4. Baseline score–change correlation

   The reported correlation between baseline score and subsequent change is susceptible to mathematical coupling, regression to the mean, and the ceiling imposed by the maximum score of 18. Consequently, the conclusion that the course was mainly beneficial for participants with lower baseline knowledge is not sufficiently supported.

   This analysis should either be removed or explicitly described as exploratory and interpreted with considerable caution. An analysis of the post-course score adjusted for the baseline score may be considered, although the small sample size remains an important limitation.

5. Ethics review

   The Institutional Review Board Statement reports that the broader project was requested by the Ministry of Health. However, administrative authorization for a health project is not necessarily equivalent to ethics committee approval for research involving human participants.

   Please provide the name of the approving ethics committee, approval number and date, or clearly explain the basis for an ethics-review exemption. The manuscript should also distinguish consent to attend the course from consent to participate in the research assessments.

Author Response

Reviewer #1
1. The p-values in Supplementary Table 2 appear inconsistent with the reported marginal counts. Most
notably, item 18 increased from 32/38 correct responses before the course to 37/38 after the course. This
necessarily implies discordant pairs, and an exact McNemar test should therefore be calculable. The
statement “Could not be calculated” is not statistically plausible. Several other results, including items 10, 14,
and 15, also require verification. Please repeat all item-level analyses using the participant-level paired data
and report whether the conventional or exact McNemar test was used, the numbers of discordant pairs, the
statistical software and function, and any missing observations. The revised results should be cross-checked
across the text, Figure 5, and Supplementary Table 2.
Re: We thank the Reviewer for noticing such inconsistency that was likely due to misprints during one of the
steps of manuscript preparation. In the revised version, we repeated all analyses, updated Supplementary
Table 2 and revised the results in manuscript text and figures accordingly. The numbers of concordant and
discordant pairs are displayed in Supplementary Tables 3-4 to improve clarity for the reader.
2. At least 36 item-level comparisons were conducted. The isolated p-value of 0.009 for item 16 may therefore
represent a chance finding. For example, a simple Bonferroni adjustment would yield an adjusted p-value of
approximately 0.324. If the authors elect not to adjust for multiple comparisons because these analyses are
exploratory, the item-level findings should be explicitly presented as descriptive and hypothesis-generating.
The delayed improvement in item 16 should not be attributed to the course without acknowledging
alternative explanations, including repeated testing, subsequent clinical experience, and other educational
exposure.
Re: We agree with the Reviewer about the descriptive purpose of the item-level comparisons, which were
included and presented as such in the original manuscript (only descriptive purpose, no adjustment for
multiple comparisons, suggested caution in interpretation of item-level comparisons). In light of Reviewer’s
comment, we think that such purpose might not have been sufficiently clarified in the original manuscript,
hence we revised the text to improve clarity on this aspect: “Adjustment for multiple testing was not applied
given the exploratory (rather than confirmatory) nature of the study, and item-level comparisons were
presented only as descriptive and hypothesis-generating.” (Methods, page 5). This was also highlighted in
Results section (“Item-level comparisons were presented only as descriptive and hypothesis-generating.”,
page 6) and in the figure legend (“Item-level comparisons were presented only as descriptive and
hypothesis-generating.”). The Discussion did not include any interpretation of the item-level comparisons,
which were left only in Results section with descriptive purpose and to improve clarity for the reader.
3. The median score decreased from 15 immediately after the course to 14 at three months, with p=0.46.
Failure to detect a statistically significant decline does not establish retention or equivalence. Please report
the paired changes, their distribution, an effect size, and preferably a confidence interval. Unless a retention
criterion or equivalence margin was prespecified, statements regarding “partial knowledge retention”
should be substantially tempered. A descriptive statement about the observed scores would be more
appropriate.
Re: We agree with the Reviewer and we removed the statements regarding “partial knowledge retention”
throughout the manuscript. In addition, we changed the statements about the observed scores in a more
descriptive way: “Overall, the median total score was 14 points (IQR 12-15) before the course, 15 points (IQR
13-16) after the course, and 14 points (IQR 13-15) three months later (Figure 1).” (Results, page 4). The paired
changes were displayed as waterfall plots Figure 4. As suggested, we also reported the median difference of
the paired changes with 99% confidence interval: “Median difference of score change from before to after the
course was 1.0 points (99% bootstrap confidence interval -0.4 to 1.9 points). Median difference of score
change from after the course to three months later was 0.0 points (99% bootstrap confidence interval -0.8 to
1.1 points).” (Results, page 4). The Statistical Analysis section was updated accordingly.

4. The reported correlation between baseline score and subsequent change is susceptible to mathematical
coupling, regression to the mean, and the ceiling imposed by the maximum score of 18. Consequently, the
conclusion that the course was mainly beneficial for participants with lower baseline knowledge is not
sufficiently supported. This analysis should either be removed or explicitly described as exploratory and
interpreted with considerable caution. An analysis of the post-course score adjusted for the baseline score
may be considered, although the small sample size remains an important limitation.
Re: We agree with the Reviewer and we removed those correlation analyses from the revised manuscript.
5. The Institutional Review Board Statement reports that the broader project was requested by the Ministry
of Health. However, administrative authorization for a health project is not necessarily equivalent to ethics
committee approval for research involving human participants. Please provide the name of the approving
ethics committee, approval number and date, or clearly explain the basis for an ethics-review exemption.
The manuscript should also distinguish consent to attend the course from consent to participate in the
research assessments.
Re: In the revised manuscript, we explained that formal ethics committee approval was not required because
it reported the anonymized results of educational activities that were performed within a health project
framework approved by the local Ministry of Health: “Formal ethics committee approval was not required
because this report summarized the results of educational activities that were performed within a health
project framework approved by the Ministry of Health of Côte d'Ivoire (protocol number
06346/MSHPCMU/CAB. 2024, September 6). Participant privacy was ensured (analysis of anonymized
dataset, no identifying information presented in the manuscript) and the course and the assessment did not
expose the participants to any harms.” (page 9). In addition, we specified that: “Informed consent to attend
the course and participate in the research assessments was obtained from all participants.” (page 9).

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors,

This is a well-conducted study that contributes valuable knowledge regarding the effects of neonatal resuscitation training. Nevertheless, I encourage the authors to more explicitly acknowledge an important limitation of the study: retention was assessed only three months after the intervention.

Previous neonatal resuscitation research has consistently demonstrated that both knowledge and, in particular, practical skills start decreasing after three months following training. Consequently, the selected follow-up time point may represent the onset of skill and knowledge decline rather than evidence of sustained retention. The current findings therefore provide insight into short-term retention only and should not be interpreted as demonstrating long-term retention. I recommend that this limitation be clearly addressed in the discussion section (limitations) and reflected in the conclusions. Future research should include longer follow-up periods (e.g., 6-12 months) to better evaluate the durability of training effects and to identify the optimal timing for refresher training.

 

Author Response

Reviewer #2
1. This is a well-conducted study that contributes valuable knowledge regarding the effects of neonatal
resuscitation training. Nevertheless, I encourage the authors to more explicitly acknowledge an important
limitation of the study: retention was assessed only three months after the intervention.
Re: We thank the Reviewer for the comment, and we explicitly acknowledged that retention was assessed
only three months after the course: “Third, the follow-up assessment was restricted to three months, hence
the data provide only limited information on onset of knowledge decline. Long-term assessments are
required to fully understand the durability of training effects and to identify the optimal timing for refresher
training.“ (Discussion, page 8).
2. Previous neonatal resuscitation research has consistently demonstrated that both knowledge and, in
particular, practical skills start decreasing after three months following training. Consequently, the selected
follow-up time point may represent the onset of skill and knowledge decline rather than evidence of
sustained retention. The current findings therefore provide insight into short-term retention only and should
not be interpreted as demonstrating long-term retention. I recommend that this limitation be clearly
addressed in the discussion section (limitations) and reflected in the conclusions. Future research should
include longer follow-up periods (e.g., 6-12 months) to better evaluate the durability of training effects and
to identify the optimal timing for refresher training.
Re: We thank the Reviewer for the comment, and we changed the concept “knowledge retention” to
“knowledge decline” when discussing the results of the 3-month assessment in the revised manuscript: “In
healthcare staff of a referral hospital in Côte d'Ivoire, neonatal resuscitation knowledge was increased after a

HBB course but partially decreased three months later.” (Abstract, page 1); “Our report showed an overall
knowledge improvement on neonatal resuscitation after a HBB course among healthcare providers in a
referral hospital in Côte d'Ivoire, although the knowledge partially declined three months later.”
(Discussion, page 7); “In a referral hospital in Côte d'Ivoire, neonatal resuscitation knowledge of healthcare
staff was increased after a HBB course but partially declined three months later.” (Conclusions, page 8). We
also updated the limitations paragraph: “Third, the follow-up assessment was restricted to three months,
hence the data provide only limited information on onset of knowledge decline. Long-term assessments are
required to fully understand the durability of training effects and to identify the optimal timing for refresher
training.” (Discussion, page 8). Moreover, we changed the final statements in the Conclusions: “Further
research should include longer follow-up periods (such as 6-12 months) to better understand the durability
of training effects and to identify the optimal timing for refresher training. Moreover, future educational
interventions may target other areas of improvement such as first minute priority, recognition of healthy
newborn, consequences of unnecessary suctioning, bag-mask ventilation, newborn's heart rate and postresuscitation
care.” (Discussion, page 8).

Reviewer 3 Report

Comments and Suggestions for Authors

The objective of this prospective study was to evaluate the knowledge improvement after a Helping Babies Breathe course and the knowledge retainment three months later among healthcare providers in a referral hospital. Among the the 38 participants, 32 were midwives and 6 were clinicians with a varied length of professional experience.

Although based on a small number of participants in a single center, as rightly pointed out by the authors, it is an important pilot study.

It seems the score after 3 months of the course dropped back to baseline (Before the course).

It would be important to mention in the discussion a plausible explanation of this finding and ways to mitigate it. e.g. perhaps there should be a refresher HBB course every 3 months.

Author Response

Reviewer #3
1. The objective of this prospective study was to evaluate the knowledge improvement after a Helping
Babies Breathe course and the knowledge retainment three months later among healthcare providers in a
referral hospital. Among the 38 participants, 32 were midwives and 6 were clinicians with a varied length of
professional experience. Although based on a small number of participants in a single center, as rightly
pointed out by the authors, it is an important pilot study.
Re: We thank the Reviewer for the kind comment that acknowledged the message of this pilot study within
its clear limitations.
2. It seems the score after 3 months of the course dropped back to baseline (Before the course). It would be
important to mention in the discussion a plausible explanation of this finding and ways to mitigate it. e.g.
perhaps there should be a refresher HBB course every 3 months.
Re: We discussed this aspect in the manuscript: “In low-resource settings, newborn resuscitation training
programs may result in common decline in performance over time, with skills deterioration that may occur
as early as 12 weeks following the initial training. [24,25] Previous studies recognized knowledge
deterioration as one of the bottlenecks for long-term educational efficacy, [23,26,27] and low-dose highfrequency
training has been suggested as a potential approach for knowledge retention over time.
[25,26,28,29] In these settings, the number of healthcare providers is often limited and resuscitation
opportunities may be infrequent, hence routine practice of these skills is not always feasible. In our context,
we believe that the involvement of local trainers and facilitators may have contributed to participant
engagement and course acceptability, [19,30] but low-dose high-frequency training may be implemented to
improve the long-term educational efficacy.” (Discussion, pages 7-8). We also highlighted this aspect at the
end of the manuscript: “Further research should include longer follow-up periods (such as 6-12 months) to
better understand the durability of training effects and to identify the optimal timing for refresher training.
Moreover, future educational interventions may target other areas of improvement such as first minute
priority, recognition of healthy newborn, consequences of unnecessary suctioning, bag-mask ventilation,
newborn's heart rate and post-resuscitation care.” (Conclusions, page 8).

Reviewer 4 Report

Comments and Suggestions for Authors

The study focuses on a topic of considerable clinical and global health importance: the evaluation of neonatal resuscitation training in a resource-limited setting. The provision of real-world data from Côte d’Ivoire, the assessment of the same participants at three different time points, and the inclusion of local healthcare workers in the training are strengths of the study.

However, in its current form, the manuscript’s main conclusion that “knowledge was retained for three months” is not adequately supported by the data. The median scores at baseline, after training, and at three months were 14, 15, and 14, respectively. Thus, a limited, short-term increase in scores was observed; however, sustained knowledge acquisition was not clearly demonstrated. Furthermore, the details of the educational intervention, the administration of the assessment instrument, and the characteristics of the participants are not described with sufficient clarity. The manuscript is not suitable for publication until these issues have been addressed. My decision is major revision.

1. The fact that the difference between the post-training score and the score at three months was not statistically significant does not prove that knowledge was retained. “No significant difference” does not indicate that the two measurements are equivalent or that no loss occurred. Moreover, the median score at three months returned to the baseline level: 14 → 15 → 14. I could not fully understand this section and would expect the authors to provide a clearer explanation. In addition, the concept of “retention” should be defined a priori. The authors should demonstrate, at the individual level, how much of the post-training gain was retained at three months. They should either provide strong justification for the phrase “partial knowledge retention” or remove it from the title, abstract, highlights, discussion, and conclusion.

2. The phrase “HBB-based course” is insufficient. The training appears to have consisted of three four-hour sessions; however, I could not fully understand certain aspects. I would expect the following points to be clarified:

• Whether the course fully adhered to the standard HBB program,

• The time allocated to theoretical instruction, skills stations, and simulation,

• The manikins and equipment used,

• The number and content of the scenarios,

• Whether the training was delivered as planned,

• Whether any additional training, practice, or refresher activities took place between the training and the three-month assessment.

3. HBB is primarily a hands-on resuscitation program; therefore, assessing theoretical knowledge alone is an important limitation. Although the course included manikin-based practice and simulation, bag-mask ventilation skills, OSCE performance, teamwork, clinical behavior, or neonatal outcomes could have been assessed more effectively.

4. The authors classify certain topics as “unsatisfactory,” but they do not provide a prespecified threshold for this classification. Furthermore, the fact that only item 16 was found to be statistically significant among the numerous item-level and time-point comparisons is an exploratory finding.

5. The limitations regarding causality and generalizability could have been addressed more thoroughly in the Discussion. The study’s original contribution to the literature should also be stated more clearly.

6. English-language revision is required. “Sperman” should be corrected to “Spearman.” In addition, the spelling of “Côte d’Ivoire” should be consistent throughout the manuscript.

Author Response

1. The fact that the difference between the post-training score and the score at three months was not
   statistically significant does not prove that knowledge was retained. “No significant difference” does not
   indicate that the two measurements are equivalent or that no loss occurred. Moreover, the median score at
   three months returned to the baseline level: 14 → 15 → 14. I could not fully understand this section and
   would expect the authors to provide a clearer explanation. In addition, the concept of “retention” should be
   defined a priori. The authors should demonstrate, at the individual level, how much of the post-training gain
   was retained at three months. They should either provide strong justification for the phrase “partial
   knowledge retention” or remove it from the title, abstract, highlights, discussion, and conclusion.
   Re: In the original manuscript, we used the term “knowledge retention” to generally refer to the concept of
   somehow maintaining the knowledge over time. Nonetheless, we understand that a strict definition should
   defined a priori to avoid misunderstanding in the interpretation of the findings, hence we removed the
   statements regarding “knowledge retention” when discussing the results throughout the manuscript. In
   addition, we changed the statements about the observed scores in a more descriptive way: “Overall, the
   median total score was 14 points (IQR 12-15) before the course, 15 points (IQR 13-16) after the course, and 14
   points (IQR 13-15) three months later (Figure 1).” (Results, page 4). The paired changes were displayed as
   waterfall plots Figure 3 (scores) and reported in Supplementary Tables 3-4 (single items). The trends of the
   passing criterion of ≥80% achieved by the participants in the three timepoints is displayed in Figure 4 using a
   Sankey diagram. Finally, we removed “knowledge retention” from the title, abstract, highlights, discussion,
   and conclusion.
   2. The phrase “HBB-based course” is insufficient. The training appears to have consisted of three four-hour
   sessions; however, I could not fully understand certain aspects. I would expect the following points to be
   clarified: whether the course fully adhered to the standard HBB program, the time allocated to theoretical
   instruction, skills stations, and simulation, the manikins and equipment used, the number and content of the
   scenarios, whether the training was delivered as planned, whether any additional training, practice, or
   refresher activities took place between the training and the three-month assessment.
   Re: In the revised manuscript, we added such information: “In September 2025, an HBB-based course was
   held in French by two neonatologists and two certificated pediatricians of the Ministry of Health, with the
   support of three facilitators. The course fully adhered to the standard HBB program. The course focused on
   neonatal resuscitation and was arranged in three 4-hour in-person educational sessions over two days. The
   course included 1.5 hours of theoretical lessons (physiology and pathophysiology of birth, initial steps,
   ventilation, practical demonstration on a mannequin, essential newborn care, group discussion), 1.5 hours of
   hands-on skill stations (mannequin-based exercises) and 1 hour of simulations (a scenario of a mild
   depressed newborn, a scenario of an severely asphyxiated newborn and a scenario of a preterm infant; all
   requiring resuscitation maneuvers and newborn stabilization). The practical sessions used four full-term
   neonatal manikins (two NewBorn Anne and two NeoNatalie; Laerdal, Stavanger, Norway), blankets, caps,
   suction pumps, and self-inflating bags. The training was delivered as planned. Additional training, practice,
   or refresher activities did not take place between the course and the three-month assessment.” (Methods,
   page 3).
   3. HBB is primarily a hands-on resuscitation program; therefore, assessing theoretical knowledge alone is an
   important limitation. Although the course included manikin-based practice and simulation, bag-mask
   ventilation skills, OSCE performance, teamwork, clinical behavior, or neonatal outcomes could have been
   assessed more effectively.
   Re: We agree with the Reviewer and we discussed this aspect among the limitations of the study: “Finally,
   only theoretical knowledge was assessed due to time constrains, without evaluating changes in practical
   skills or quality of care provided at birth in clinical practice. The HBB program includes both theoretical and
   practical education, which should be assessed after the training. [7] Of note, theoretical knowledge and
   practical skills may follow different temporal trajectories, with practical skills declining earlier than
   theoretical knowledge. [18,24]” (Discussion, page 8).
2. 4. The authors classify certain topics as “unsatisfactory,” but they do not provide a prespecified threshold for
   this classification. Furthermore, the fact that only item 16 was found to be statistically significant among the
   numerous item-level and time-point comparisons is an exploratory finding.
   Re: We agree with the Reviewer. In the revised manuscript, we removed the referring to ““unsatisfactory”
   and stated that “The number of correct answers regarding some topics such as first minute priority,
   recognition of healthy newborn, consequences of unnecessary suctioning, bag-mask ventilation, newborn's
   heart rate and post-resuscitation care suggested that birth pathophysiology and bag-mask ventilation should
   require more attention during the training.” (Discussion, page 8). In the original manuscript, the item-level
   comparisons had only descriptive purpose and were presented as such (only descriptive purpose, no
   adjustment for multiple comparisons, suggested caution in interpretation of item-level comparisons). In light
   of Reviewer’s comment, we think that such purpose might not have been sufficiently clarified in the original
   manuscript, hence we revised the text to improve clarity on this aspect: “Adjustment for multiple testing
   was not applied given the exploratory (rather than confirmatory) nature of the study, and item-level
   comparisons were presented only as descriptive and hypothesis-generating.” (Methods, page 4). This was
   also highlighted in Results section (“Item-level comparisons were presented only as descriptive and
   hypothesis-generating.”, page 6) and in the figure legend (“Item-level comparisons were presented only as
   descriptive and hypothesis-generating.”). The Discussion did not include any interpretation of the item-level
   comparisons, which were left only in Results section with descriptive purpose and to improve clarity for the
   reader.
   5. The limitations regarding causality and generalizability could have been addressed more thoroughly in
   the Discussion. The study’s original contribution to the literature should also be stated more clearly.
   Re: We added some considerations in the revised manuscript: “Second, the single-center design limited the
   generalizability of the findings to similar settings, hence different knowledge improvement and retention
   may be found in other healthcare systems with different professional background and resource availability.
   For example, participants with low baseline knowledge may show the largest magnitude of improvement
   due to the ceiling imposed by the maximum score of 18, whereas participants with high baseline knowledge
   may improve the overall achievement of the passing criterion of ≥80%.” (Discussion, page 8). In addition, we
   specified that “Our findings contributed to the evidence about knowledge benefits after a HBB course, and
   confirmed that knowledge decline might occur even at short-term. [8,21-25] Moreover, they suggested that
   the minimal duration of the course (4 hours) might provide limited benefits when baseline knowledge is
   low. [7,16] In low-resource settings, newborn resuscitation training programs may result in common decline
   in performance over time, with skills deterioration that may occur as early as 12 weeks following the initial
   training. [24,25] Previous studies recognized knowledge deterioration as one of the bottlenecks for long-term
   educational efficacy, [23,26,27] and low-dose high-frequency training has been suggested as a potential
   approach for knowledge retention over time. [25,26,28,29] In these settings, the number of healthcare
   providers is often limited and resuscitation opportunities may be infrequent, hence routine practice of these
   skills is not always feasible. In our study, we believe that the involvement of local trainers and facilitators
   may have contributed to participant engagement and course acceptability, [19,30] but low-dose highfrequency
   training may be implemented to improve the long-term educational efficacy.” (Discussion, pages
   7-8).
   6. English-language revision is required. “Sperman” should be corrected to “Spearman.” In addition, the
   spelling of “Côte d’Ivoire” should be consistent throughout the manuscript.
   Re: We performed an English-language revision of the manuscript. We also corrected “Sperman” to
   “Spearman.” and ensured consistent spelling of “Côte d’Ivoire” throughout the manuscript.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for the careful revision. The manuscript has improved substantially, and most of my previous concerns have been satisfactorily addressed. I have only a few remaining comments.

1. Please clarify the implementation of the McNemar analyses.
   The revised manuscript states that McNemar tests were performed using the mcnemar.test function, and the addition of concordant and discordant pairs in the Supplementary Tables is helpful. However, please explicitly state whether the conventional asymptotic McNemar test or an exact McNemar test was used, particularly for items with a small number of discordant pairs. Please also state whether there were any missing paired observations in the item-level analyses.
2. Please further temper the wording regarding the three-month change.
   The removal of statements regarding “knowledge retention” is appropriate. However, the remaining statement that knowledge “partially declined” at three months may still be stronger than supported by the data, given that the median paired change from immediately after the course to three months was 0.0 points with a 99% bootstrap confidence interval of −0.8 to 1.1 points. A more descriptive formulation, such as “scores were numerically lower at three months” or “performance was lower at the three-month assessment than immediately after the course,” would be preferable.
3. Please clarify the basis for the ethics-review exemption.
   The revised statement explains that the broader health project was approved by the Ministry of Health of Côte d’Ivoire. However, project or administrative approval is not necessarily equivalent to a formal determination that research ethics review was not required. Please clarify whether a competent ethics committee or institutional authority formally determined that ethics review was not required, or provide the applicable institutional or regulatory basis for this exemption. If no formal exemption determination was obtained, the wording should accurately reflect this distinction.

With these clarifications, I would have no further major concerns.

Author Response

REVIEWER 1

Thank you for the careful revision. The manuscript has improved substantially, and most of my previous concerns have been satisfactorily addressed. I have only a few remaining comments.

 

1. Please clarify the implementation of the McNemar analyses. The revised manuscript states that McNemar tests were performed using the mcnemar.test function, and the addition of concordant and discordant pairs in the Supplementary Tables is helpful. However, please explicitly state whether the conventional asymptotic McNemar test or an exact McNemar test was used, particularly for items with a small number of discordant pairs. Please also state whether there were any missing paired observations in the item-level analyses.

 

Re: We thank the Reviewer for the suggestion and we redid the analysis using the exact McNemar test, which yielded the same results (with some minor changes in the decimals, of course). In the revised manuscript, we specified the use of “the McNemar exact test (function mcnemar.exact in package exact2x2).” (Methods, page 4). We also specified that “There were no missing paired observations in the item-level analyses.” (Results, page 6).

 

 

2. Please further temper the wording regarding the three-month change. The removal of statements regarding “knowledge retention” is appropriate. However, the remaining statement that knowledge “partially declined” at three months may still be stronger than supported by the data, given that the median paired change from immediately after the course to three months was 0.0 points with a 99% bootstrap confidence interval of −0.8 to 1.1 points. A more descriptive formulation, such as “scores were numerically lower at three months” or “performance was lower at the three-month assessment than immediately after the course,” would be preferable.

 

Re: In the revised manuscript, we tempered the wording regarding the three-month change as suggested by stating “the performance was lower at the three-month assessment than immediately after the course” (pages 1,7,9).

 

3. Please clarify the basis for the ethics-review exemption. The revised statement explains that the broader health project was approved by the Ministry of Health of Côte d’Ivoire. However, project or administrative approval is not necessarily equivalent to a formal determination that research ethics review was not required. Please clarify whether a competent ethics committee or institutional authority formally determined that ethics review was not required, or provide the applicable institutional or regulatory basis for this exemption. If no formal exemption determination was obtained, the wording should accurately reflect this distinction. With these clarifications, I would have no further major concerns.

 

Re: In the revised manuscript, we rephrased the section about IRB Statement to improve clarity for the reader: “Formal ethics committee approval was not sought because the activities described in this report were carried out as educational activities rather than as research involving human subjects. This report summarized the outcomes arising from educational activities that were performed within a health project framework approved by the Ministry of Health of Côte d'Ivoire (protocol number 06346/MSHPCMU/CAB. 2024, September 6). On this basis, the work was considered to fall outside the scope of formal ethics committee review. Participant privacy was ensured (analysis of anonymized dataset, no identifying information presented in the manuscript) and the course and the assessment did not expose the participants to any harms.” (page 9).

 

4. Figures and tables must be improved.

 

Re:  We improved the quality of the figures in the manuscript and the clarity of the tables in the Supplementary Material.

Reviewer 4 Report

Comments and Suggestions for Authors

The manuscript is methodologically well-structured, consistent with the current literature, and presents clinically valuable findings. The authors have appropriately addressed the comments and suggestions raised during the previous review process, resulting in a significant improvement in the scientific quality of the study.

In its current form, I consider the manuscript to be scientifically adequate and capable of making a meaningful contribution to the readership of the journal. Therefore, I recommend acceptance of the manuscript in its present form.

Sincerely,

Author Response

1. The manuscript is methodologically well-structured, consistent with the current literature, and presents clinically valuable findings. The authors have appropriately addressed the comments and suggestions raised during the previous review process, resulting in a significant improvement in the scientific quality of the study. In its current form, I consider the manuscript to be scientifically adequate and capable of making a meaningful contribution to the readership of the journal. Therefore, I recommend acceptance of the manuscript in its present form.

 

Re: We thank the Reviewer for the comment.

 

2. The English could be improved to more clearly express the research.

Re: We revised the text of the updated manuscript to improve clarity and readability for the reader. Minor syntax changes were not highlighted with Track Change to improve readability for the Reviewers.