Hydroxychloroquine in COVID-19 Prevention Review

Question 1 – start on new page.

Read the systematic review and meta-analysis paper below, then answer the questions.The following papers are available on the Leganto reading list on Moodle or links below.

Garcia-Albeniz et al . Systematic review and meta-analysis of randomized trials of hydroxychloroquine for the prevention of COVID-19. European Journal of Epidemiology. 2022; 37(8):789-796. https://unsw.alma.exlibrisgroup.com/leganto/public/61UNSW_INST/citation/57859859140001731?auth=SAML

To answer Q3.1 you will also need to download the paper by Naggie et al. below, which is included in the above systematic review: https://unsw.alma.exlibrisgroup.com/leganto/public/61UNSW_INST/citation/57859912830001731?auth=SAML (NB. This paper was published in the International Journal of Infectious Diseases in 2023; an earlier pre-print version was included in the review).

Before answering the questions below, you are encouraged to note key elements of the systematic review and meta-analysis protocol, including the:

• Research question
• Study designs included in the systematic review
• Treatment strategies (intervention/s and comparator)

Primary outcome

A key task when undertaking a systematic review is completing the risk of bias assessments for all included studies. This assignment question aims to give you some experience with a small part of this task by asking you to assess the risk of bias in the measurement of the outcome for one of the papers included in the Garcia-Albeniz et al . 2022 systematic review.

Step 1: You will need to use the ROB v2.0 template questions 4.1 to 4.5 to assess the risk of bias in measurement of the outcome (Domain 4) for the Naggie et al. paper included in the review (see link to the paper above). In the Comments section of the template, write about the information in the paper that supports your rating for each question. You do not need to submit the form for this assignment, but you will use this to draft your answer to this question below.

You can download the RoB v2.0 template with the questions for the risk of bias assessment at the link below or from the A3 assignment section on Moodle: https://drive.google.com/file/d/18Zks7k4kxhbUUlbZ51Ya5xYa3p3ECQV0/view

More information about the ROB v2.0 and how to use it can be found here (optional): https://sites.google.com/site/riskofbiastool/welcome/rob-2-0-tool?authuser=0

Step 2: Review the Domain 4 (questions 4.1 to 4.5) risk of bias assessment judgements for the Naggie et al paper published in the supplementary material by Garcia-Albeniz and colleagues (available in Excel file uploaded in A3 assignment section on Moodle).

Step 3: In the space below, explain why you think Garcia-Albeniz and colleagues made the Domain 4 risk of bias judgements for each of the RoB2 questions 4.1 to 4.5 that they reported for the Naggie et al. study , which led to the overall rating of ‘Some Concerns’

for the Domain 4 risk of bias assessment for the measurement of the outcome. Refer to specific information reported in the Naggie et al . paper to support your answer.

Negi et al. (2023) demonstrated a thorough methodology in measuring the result using a composite primary outcome that included confirmed and suspected COVID-19 clinical infections. The study acknowledged difficulties in areas where laboratory confirmation for COVID-19-like illness was not widely accessible, which could affect the accuracy of outcome assessments, even though it was not fundamentally improper. The study population's lower-than-expected rates of asymptomatic shedding and seroconversion raised questions regarding the outcome measurement's sensitivity. The fact that the trial was not powered to detect a tiny treatment effect further highlights the difficulties in detecting meaningful variations in the outcome, underscoring the significance of being aware of these potential limitations when interpreting the data.

A composite primary outcome that combined verified and suspected cases based on symptoms and test findings was used in the study's method to outcome ascertainment. The research acknowledged constraints associated with laboratory validation and variances in testing regulations among various geographical areas. Yet, it remained committed to a double-blind strategy to reduce biases in outcome determination. However, the trial's partially distant design and limitations on laboratory confirmation created complications that might have impacted how differently the intervention groups' outcomes were ascertained.

Most importantly, the use of a double-blind design in the study ensured that researchers evaluating the results were unaware of the intervention that study participants had received. This careful blinding technique, which included assessors and participants, was designed to remove biases in outcome evaluation from assessors' knowledge of the intervention. This meticulous methodology enhances the study's internal validity by guaranteeing the objectivity of the primary outcome evaluation.

The trial was further blinded to avoid any potential impact on outcome evaluation based on knowledge of the received intervention. By keeping both participants and result assessors in the dark regarding the intervention (hydroxychloroquine or placebo), this strategic blinding upholds the study's dedication to reducing biases. It improves the validity of the outcome assessment.

The study effectively reduced the possibility that the outcome assessment was impacted by knowledge of the intervention received by the study participants because of the thorough double-blind design and the lack of information suggesting awareness of the intervention by outcome assessors. Lastly, the study's attempts to guarantee the independence and dependability of the result evaluation from knowledge of the intervention are highlighted by the careful consideration of blinding techniques.

Refer to Figure 1 on page 793 that presents the results of the meta-analysis for the pre-exposure prophylaxis trials.

Report the main findings from the meta-analysis reported in Figure 1 and how this adds to the findings from the original studies.

To support your answer, refer to the specific overall fixed effect estimate for COVID-19 and its interpretation, then discuss in comparison to the results using the alternative meta-analysis methods (i.e. random effects models) and the original studies. Discuss specific details such as the number of studies included, study designs, sample sizes,

study populations, and heterogeneity (formal calculation and visual observation in the Figure) in your summary of the meta-analysis findings.

Findings and Contributions

The thorough meta-analysis by Garcia-Albeniz et al. (2022) which examines seven pre-exposure prophylaxis trials is an important effort to clarify the effectiveness of hydroxychloroquine (HCQ) as a COVID-19 preventive measure. The fixed-effect estimate, which summarised the entire data set, revealed a strong reduction in risk with HCQ, as evidenced by a notable risk ratio of 0.72 (95% CI: 0.58-0.90). This significant decrease in the likelihood of acquiring COVID-19 highlights the possible preventive effectiveness of HCQ compared to those not given the medication. Positively, using different meta-analysis methods (random-effects models, in particular) produced consistent findings. The core findings are strengthened by the pooled risk ratio estimate of 0.72 (95% CI: 0.55-0.95) that was achieved by carefully modifying the Hartung-Knapp random-effects approach. This estimate establishes a consistent protective impact across several trials.

The safety evaluation is another crucial aspect of the investigation since major side effects were similar in the HCQ and non-HCQ groups. This finding indicates that HCQ appears tolerable and strengthens its case as a COVID-19 preventative measure because there hasn't been any appreciable rise in adverse events.

Even while the meta-analysis provides strong support for HCQ, there is something to be said about the minor signs of heterogeneity that may be seen in the forest plot. The visual depiction of the data indicates some variety in effect estimates among the constituent studies, even if formal assessments of heterogeneity were not explicitly given. This emphasises the need for careful interpretation and draws attention to any discrepancies in the treatment effect's magnitude between trial settings.

In summary, the meta-analysis offers a more accurate and nuanced assessment of the treatment impact linked to HCQ in COVID-19 prevention, which substantially contributes to the body of knowledge already in existence. The convergence of results from different analytical techniques and the positive safety profile highlights the potential value of HCQ in reducing the risk of COVID-19. However, the complex examination of heterogeneity highlights the need for ongoing examination and judgment in interpreting these results, enhancing the conversation over HCQ's place in the COVID-19 preventive landscape.

In your own words, describe the main finding/conclusion from the pre-exposure prophylaxis meta-analysis in the context of the original research studies and key implication/s from the meta-analysis.

By combining information from several studies, the meta-analysis by Garcia-Albeniz et al. (2022) significantly improves the accuracy of the estimated treatment effect of hydroxychloroquine (HCQ) in COVID-19 prophylaxis. This method significantly increases statistical power by combining data from several trials, reducing uncertainty and offering a stronger basis for inference. One important aspect that suggests HCQ is well-tolerated is the observed parity in the rate of major side effects between the groups receiving HCQ and those not receiving it. This positive safety profile is key in supporting HCQ's efficaciousness as a COVID-19 prophylactic.

Moreover, the meta-analysis emphasises the importance of considering data from numerous randomised trials to draw thorough conclusions on how well HCQ works to 

prevent COVID-19. This all-encompassing strategy offers a thorough awareness of potential changes across various study settings and facilitates a more nuanced comprehension of the treatment effect. As a result, the meta-analysis plays a critical role in synthesising current knowledge and promoting a better understanding of the function of HCQ in pandemic management.

Nonetheless, these results have consequences that go beyond science and impact pandemic management. The meta-analysis clarifies the possible repercussions of taking early prophylactic trial results as proof that HCQ is ineffective. These interpretations might have interfered with the completion of later experiments, which would have delayed the timely production of accurate estimations necessary for pandemic management. This highlights the significance of assessing treatment measures during public health crises using a fair and evidence-based methodology since drawing hasty conclusions could have unfavourable effects. Essentially, the meta-analysis clarifies the comprehension of the efficacy of HCQ while emphasising the wider significance of systematic and cooperative scientific research in forming successful solutions to global health issues.

Question 2 - start on a new page.

Question 4 aims to assess your understanding of the process of emulating a target trial using observational data. To prepare to answer the questions below, first read the paper, then consider the key elements of the target trial that the study aims to emulate. It is a good idea to write these down specifically so you can consider the methods used in the ideal (target) trial vs. the observational study.

Suarez et al. Association of Antidepressant Use During Pregnancy With Risk of Neurodevelopmental Disorders in Children. JAMA Internal Medicine. 2022; 182(11):1149-1160.

https://unsw.alma.exlibrisgroup.com/leganto/public/61UNSW_INST/citation/57892049300001731?auth=SAML

For all questions below, please focus on:

• the primary exposure
• the aggregate outcome (Any neurodevelopmental disorders, NDD) 

Write the research question in a sentence using the PICO format for the primary exposure (treatment strategies) and the target population for the study. What is the association between antidepressant use during pregnancy and the risk of neurodevelopmental disorders in children?

In your own words, explain your understanding of the motivation for conducting the study and two reasons why the researchers used an observational study design rather than a randomised controlled trial to answer the question.

Motivation for conducting the study

Investigating the relationship between antidepressant usage during pregnancy and the risk of neurodevelopmental disorders (NDDs) in offspring was the driving force behind the research done by Suarez et al. (2022). It was critical to ascertain whether there was a true causal association between antidepressant usage during pregnancy and non-destructive pregnancy disorders (NDDs), given the inconsistent findings of prior investigations.

Reasons for using an observational study design

The first compelling factor was ethics. It would be unethical to conduct a randomised controlled trial (RCT) in which pregnant participants are selected at random to take antidepressants or not, as this could put the control group in danger or deny the treatment group access to essential medication. Second, observational studies provided a realistic, clinical practice-like environment for researchers to investigate the effects of antidepressant use during pregnancy. This offers important details about the advantages and disadvantages of antidepressant use in a variety of demographics.There were two cohorts included in this study. You can find the characteristics of the two cohorts reported in Table 1.

Describe the key definition of the two cohorts in one sentence each. Then explain why you think the study authors chose to use data from two cohorts, rather than one cohort, for this study. Justify your answer with reference to information reported in the paper (most relevant is the information reported in Table 1).

Characteristics of the two cohorts

Medicaid Analytic eXtract (MAX) cohort: Publicly insured pregnant individuals and their children, data collected from 2000 to 2014.

IBM MarketScan Research Database (MarketScan) cohort: Privately insured pregnant individuals and their children, data collected from 2003 to 2015.

Reasons for using two cohorts

The first reason was to increase generalizability, as publicly and privately insured individuals captured a more diverse population and accounted for potential differences in healthcare utilisation and access. Secondly, the cohorts covered different time periods, allowing for a broader range of data and longer follow-up periods. Lastly, pooling the cohorts helped overcome sparse data challenges, ensuring more robust and reliable results.

Before you answer the following questions, you are encouraged to write a summary for yourself on the following key elements of the emulated trial protocol to ensure you have read the relevant information needed to interpret the results reported in the paper:

Causal contrast that is estimated in the study.

• Primary exposure/treatment strategies compared and how were the treatment strategies were measured/defined from the available data sources.
• Treatment assignment in this study and how it was similar/different to the target trial.
• Study outcomes and the source/s of information on the outcome in the observational study.
• Follow-up period for the study.

First, read the ‘Statistical Analysis’ sub-section in the Methods section of the paper and corresponding results in Tables and Figures. Focus on the aggregate outcome of ‘Any neurodevelopmental disorder’ (Any NDD).

Second, look at Figure 1 where the study authors present the descriptive results: focus on 1c that shows the cumulative incidence of ‘Any NDD’ in the exposed and unexposed groups in each cohort.

Third, look at Figure 2 where the study authors report the: (i) unadjusted measure of association (i.e. the hazard ratio) for each outcome, including ‘Any NDD’; and (ii) the results from four analyses that used different approaches to minimising bias due to confounding (i.e. Adjusted, HDPS adjusted, Discontinuer referent and Sibling analysis), using data from the pooled cohorts.

With a focus on the methods to minimise bias due to confounding from the design to analysis stages of the study , describe and compare the methods and results of the unadjusted and the ‘Adjusted’ analysis in Figure 2 (for the ‘Any NDD’ outcome), including why you think there was such a difference in the hazard ratios between the unadjusted and Adjusted analyses. Refer to the results in Figure 2, as well as Table 1 and/or Figure 1, to support your answer.

Methods to minimise bias due to confounding

Unadjusted analysis: The unadjusted analysis in Figure 2 provides the hazard ratio (HR) for the association between antidepressant exposure during pregnancy and the outcome of "Any neurodevelopmental disorder" (Any NDD) without any adjustment for confounding factors.

Adjusted analysis: The "Adjusted" analysis in Figure 2 accounts for potential confounding factors by adjusting for covariates such as age, race/ethnicity, and other relevant variables. This analysis provides a more accurate estimate of the association between antidepressant exposure and Any NDD.

Results of the unadjusted and Adjusted analyses

Firstly, concerning the unadjusted analysis, the unadjusted HR for Any NDD in the unexposed group is 1.0. In contrast, the unadjusted HR for the exposed group is 1.97 (95% CI, 1.92-2.02). This suggests a higher risk of Any NDD in the exposed group than the unexposed group. However, after controlling for confounding variables, the Adjusted analysis shows no significant correlation between antidepressant exposure and Any NDD. The HR for Any NDD in the exposed group is 0.97 (95% CI, 0.88-1.06).

The difference in hazard ratios between unadjusted and Adjusted analyses

The Adjusted analysis's confounding factor accounting can explain the variation in hazard ratios between the unadjusted and adjusted analyses. The Adjusted analysis accounts for covariates, including age, race/ethnicity, and other pertinent variables, to account for potential confounding and yield a more precise estimate of the connection. Confounding factors may have impacted the initial connection seen in the unadjusted study, as suggested by the considerable fall in the HR in the adjusted analysis. 

In Figure 2, the study authors also reported analysis results from three additional approaches to minimising bias due to confounding – the HDPS (high-dimensional propensity score) adjusted, the discontinuer referent and the sibling analysis.

The discontinuer and sibling approaches both use sub-samples of the overall study population to examine the degree of residual confounding that may remain in the Adjusted and HDPS adjusted analyses.

Describe your understanding of the specific design and/or analysis strategies used to minimise bias due to confounding in the discontinuer and sibling approaches . Explain what you think these two extra analyses tell us in addition to the results from the ‘Adjusted’ and ‘HDPS adjusted’ approaches, concerning the results for the ‘Any NDD’ outcome in Figure 2. Approaches/ Strategies

The discontinuer technique is the first one; it compares the exposed group's results to a subgroup of the unexposed group that was previously exposed to antidepressants before becoming pregnant. It aims to evaluate the potential residual confounding in the Adjusted and HDPS-adjusted analyses. Sibling analysis, the second method, contrasts the results of exposed versus unexposed sibling groups. This research accounts for shared familial environmental and genetic variables by comparing siblings who were subjected to antidepressants during pregnancy alongside their unexposed siblings.

Additional insights from the discontinuer and sibling analyses

A comparison between the exposed group and a subgroup of the unexposed group that had past exposure to antidepressants is made possible by the discontinuer analysis, which offers information on the possible influence of prior antidepressant exposure on the outcomes. This aids in assessing how previous exposure contributes to the observed connections. On the other hand, by comparing outcomes among exposed and unexposed sibling groups, sibling analysis aids in controlling for common familial characteristics. After adjusting for genetic and environmental variables, this analysis offers more proof of the link between neurodevelopmental outcomes and antidepressant exposure during pregnancy.

These further analyses shed important light on the possible influence of shared familial variables and previous exposure on the relationships between antidepressant use during pregnancy and neurodevelopmental outcomes that have been found. They contribute to the evaluation of the findings' robustness and offer a more thorough comprehension of the connection between children's exposure to antidepressants and neurodevelopmental abnormalities.

• The discontinuer approach compares outcomes in the exposed groups to a sub-group of the unexposed group who had previously been exposed to antidepressants prior to
• The sibling analysis compared outcomes within exposed and unexposed sibling groups.

In your own words, summarise in plain English the main findings/conclusion from this study and implications for clinical practice. Include the main findings from both the descriptive and statistical modelling for the primary exposure and the ‘Any NDD’ outcome in your answer.

Main Findings and Implications for Clinical Practice

First, after controlling for confounding variables, the study by Suarez et al. (2022) discovered no meaningful correlation between the use of antidepressants during pregnancy and the incidence of neurodevelopmental problems in offspring. Second, each cohort's exposed and unexposed groups had comparable cumulative incidences of "Any neurodevelopmental disorder" (Any NDD), according to the descriptive results in Figure 1. Thirdly, compared to the unexposed group, the exposed group had a greater risk of Any NDD, according to unadjusted analysis. The adjusted study, which considered confounding variables, did not find a significant correlation between antidepressant exposure and any NDD. In addition, studies, such as the HDPS adjusted, discontinuer referent, and sibling analyses, were done to avoid bias owing to confounding. The results of the adjusted analysis, which showed no significant correlation between antidepressant exposure during pregnancy and any NDD, were corroborated by these studies.

Finally, the results of the study imply that the use of antidepressants during pregnancy does not raise the child's risk of neurodevelopmental problems. Antidepressant exposure during pregnancy, however, might still be a crucial indicator of the need for earlyscreening and intervention, given primitive connections. The choice to use antidepressants during pregnancy should be determined by an in-depth evaluation of the individual's mental health requirements, weighing the potential benefits and hazards. These findings are significant for clinical practice overall. When it comes to pregnant people who have used antidepressants, careful observation and prompt intervention can be necessary.

Get your assignment help here: BFD105-Biological Foundations Assessment Answers

Looking For Complete Assignment Sample Get It On WhatsApp