RR:C19 Evidence Scale rating by reviewer:
Potentially informative. The main claims made are not strongly justified by the methods and data, but may yield some insight. The results and conclusions of the study may resemble those from the hypothetical ideal study, but there is substantial room for doubt. Decision-makers should consider this evidence only with a thorough understanding of its weaknesses, alongside other evidence and theory. Decision-makers should not consider this actionable, unless the weaknesses are clearly understood and there is other theory and evidence to further support it.
***************************************
Review: The study addresses a critical public health issue in a malaria-endemic region, focusing on the determinants of timely malaria treatment among under-five children. The use of the Levesque et al. Conceptual Framework for Patient-Centered Access to Health is appropriate and provides a comprehensive view of the factors influencing healthcare access. The paper is generally well-organized, but several areas require attention to improve clarity, accuracy, and overall quality.
Specific Comments:
Title:
Abstract:
The abstract could also benefit from a more precise language even though it is generally well-written. E.g., consider rephrasing "This study aimed to assess the determinants of timely malaria treatment..." as „This study aimed to identify the factors influencing timely malaria treatment among under-five children...‟
Introduction:
The authors gave a good overview of the problem in the introduction, but could be improved by briefly discussing the potential impact of delayed malaria treatment on morbidity and mortality among children.
Long sentence: please revise the sentence “Despite all these efforts, the proportion of under-five children in Kisumu East Sub-county receiving timely malaria treatment remains low as evidenced in Kenya Malaria Indicator Survey (KMIS) of 2020 which reported that only 36% of children with recent fever received timely treatment on the same or next day following fever onset” into two interconnected units.
Please a citation for the statement "…while another study done in southwest Ethiopia showed that malaria knowledge was not a significant determinant of timely malaria treatment among under-five ".
Methodology:
Conceptual framework:
Provide a citation for the sentence “This study employed Levesque et al Conceptual Framework for Patient-centred access to health to define access to timely malaria treatment.”
Provide a figure for “Levesque et al Conceptual Framework for Patient-centred access to health” and call it out in the text, other than the theoretical clarification.
Study Design:
The authors used an appropriate study design, but the choice of a cross-sectional design inherently limits the ability to draw causal inferences. This is a limitation that should be acknowledged in the discussion to guide users‟ interpretation.
Please revise the following sentence to correct the prepositional typo: “for malaria treatment in from public health facilities in Kisumu East sub-county, Kenya.”
Study Variables:
Data quality control:
The authors pre-tested the questionnaire among 14 participants in two health facilities for adjustments. Even though they stated that these participants were not part of the study, it is unclear if the two health facilities were or not. Also, in which communities or locations were these health facilities, and how characteristically similar were they to those in the actual study? This information is important for clarity of the process and to ensure easy replication/validation of the study.
Sample size determination:
The authors demonstrated how the sample size was obtained. However, it is strange that they used indicators from two different reports of the KMIS (2021, and 2020) to calculate the sample size. Why were the 5% non-response rate and 1.2 design effect obtained from KMIS 2020 version the 2021 version was available? It is a common practice in research to rely on the most recent available data, especially if no newer data is available for certain parameters. Ideally, all parameters for the sample size calculation should come from the most recent and relevant data source to ensure consistency and accuracy. Therefore, the authors need to provide a brief justification in the methods section explaining why the 2020 KMIS data was used for non-response and design effect, while the 2021 KMIS was used for the proportion. This transparency would strengthen the methodological rigor of the study. Otherwise, it should be stated clearly in the limitation section of the study.
Besides, the in-text citation of "(KMIS, 2021)" is inconsistent and should be revised to match the rest of the in-text citations.
Statistical Analysis:
The use of a 20% significance level in the bivariate analysis is unusually high, although the statistical methods are appropriate. The authors should consider justifying this choice or using a more conventional level (e.g., 5%). The reason that “Higher level of significance was chosen to allow more explanatory variables to be included at multivariable analysis stage” is vague and insufficient. Did the 20% allow the inclusion of how many more covariates into the model, and what is the justification for this? It suggests that, researchers could basically choose whatever significance level base on subjective preference e.g. could they have chosen 15 or 35% significance level (and why not if not so)?
It is well-justified to use robust standard errors due to clustering. - What were the results of using the Akaike Information Criteria (AIC) and Bayesian Information Criteria (BIC) to determine the best-fit model? These need to be reported.
Ethical consideration: The clarified that clearance was obtained from the University of Zambia Biomedical Research Ethics Committee with additional clearance from two other places, which is appropriate. However, it is not enough to assume that readers will know that e,g, Maseno University is Kenya. As the study was conducted in Kisumu East sub-county in Kenya, it is important to clarify that the additional clearances were obtained were from Kenya.
Results:
The authors say “a total of 434 caregivers were included, with each of the seven health facilities contributing 62 caregivers”, but they had a sample size of 446. Meanwhile, they did not clarify under statistical analyses or data management how many of the 446 women were eliminated at what stage and why. Please clarify what happened with the 12 caregivers. It is also interesting each of the health facilities contributed equally to the proportion of sample size included in the analyses. It is important to provide a table of the health facilities, their respective OPD attendance for malaria, and the respective proportions of recruited to make up the sample size of 446. This information will give a better perspective on how many were eliminated from which health facility viz-a-viz a clarification on why they were not included in the analyses.
Terminology: terms like “A substantial majority”, “minority”, “A larger proportion”, “A significant proportion”, etc. should be avoided at the results stage, as they already denote an interpretation instead of describing.
The results are presented clearly, but there is some unnecessary repetition in reporting the key findings. For instance, the association between the ability to recognize symptoms and timely treatment is mentioned multiple times.
Table 1:
The title “Descriptive Statistics” is vague and does not comprehensively address the content of the Table.
Ensure that all variable items add up to 100. E.g. the sum of percentages under the variable “Malaria Knowledge” is 0.01 less than 100.
Provide the USD equivalence of Ksh in Household income to give a contextual yet global perspective of that index.
Delete: “Regarding physical factors, caregivers primarily utilized two modes of transport for seeking medical care for their children which were foot and motorbike/tuktuk” as it is clearly inherent in the subsequent sentences.
Revise “…accounting to” to „accounting for‟.
Under “Determinants of timely malaria treatment”, revise “compared to those who were those that were unable able to tell…”
The text under “Knowledge and health seeking practices” does not call out or refer to any table and/or figure. Please revise.
Table 3 and the related text describe the adjusted odds ratios (AORs). Ensure that all AORs are reported with consistent precision (e.g., two decimal places) and that the confidence intervals are clearly presented.
Discussion:
The discussion appropriately interprets the findings, but there are areas where the interpretation could be strengthened. For example, the finding that no physical factors were associated with timely treatment contrasts with other studies. This discrepancy should be explored in more detail.
The study's limitations are partially addressed, but the impact of potential recall bias and selection bias should be discussed more explicitly. The cross-sectional design and its limitation in establishing causality should also be noted. Do not forget the impact of the 20% significance level considered in the design.
Conclusion:
References:
Minor Corrections:
"Only 18.43% (80/434) sought treatment from herbalist or traditional healer before presenting to health facility" should be "Only 18.43% (80/434) sought treatment from a herbalist or traditional healer before presenting to a health facility."
Paragraph openers (see Discussion): It is confusing to use “Finally” and later use “Moreover” for a subsequent paragraph. Finally can only mean finally.
In conclusion, this manuscript addresses an important public health issue and provides valuable insights into the determinants of timely malaria treatment in a malaria-endemic region. While the study design and methods are generally sound, several areas need revision, particularly in terms of clarity, consistency, and the acknowledgment of limitations. These issues need to be addressed to significantly strengthen the manuscript and make it more suitable for publication.
Creative Commons Attribution 4.0 International License (CC-BY 4.0)
