Epidemiology: Open Access
Open Access

Nutritional epidemiology; Covariates; Confounding factors; Meta-epidemiology; Dietary studies; Methodological review

Introduction

Nutritional epidemiology stands at the intersection of two pivotal fields - nutrition and epidemiology. It is dedicated to unraveling the intricate relationship between dietary habits, health outcomes, and disease risks within populations. Central to the precision and validity of findings in nutritional epidemiology studies is the judicious selection of covariates. These additional variables, often representing demographic, lifestyle, or clinical characteristics, serve as essential control measures, mitigating the influence of potential confounding factors on observed associations between dietary factors and health outcomes [2 ].

This meta-epidemiological review embarks on a systematic evaluation of covariate selection methods across a diverse spectrum of nutritional epidemiology studies. By synthesizing data from multiple investigations, we aim to discern prevalent practices, scrutinize their merits and limitations, and proffer recommendations for optimal strategies in covariate selection. This comprehensive analysis seeks to refine methodological standards in the field, ultimately amplifying the quality and reliability of research endeavors in nutritional epidemiology.

In this review, we embark on a journey to dissect the nuances of covariate selection in nutritional epidemiology studies. By comprehensively evaluating practices employed across a wide array of studies, we hope to shed light on the multifaceted nature of this methodological decision. The significance of covariate selection reverberates throughout the research process, from study design to data analysis [3 ]. A well-informed selection ensures that observed associations genuinely reflect the influence of diet on health outcomes, unclouded by potential confounders.

Through this meta-epidemiological exploration, we aspire to contribute to the refinement of methodology in nutritional epidemiology. By enhancing our understanding of the factors influencing covariate selection, we endeavor to equip researchers and practitioners with valuable insights that will fortify the rigor and robustness of studies in this critical field [4].

Methods

1. Inclusion criteria:

• A comprehensive literature search was conducted using electronic databases such as PubMed, Web of Science, and relevant nutritional epidemiology journals. Studies included in this meta-epidemiological review met the following criteria:

• Primary research studies in the field of nutritional epidemiology.

• Clearly reported methods for covariate selection.

• Published in peer-reviewed journals.

2. Search strategy:

• The search encompassed studies published between [specific time frame] to [specific time frame], with no language restrictions. The search terms included combinations of keywords such as "nutritional epidemiology," "covariates," "confounding factors," and "dietary studies."

3. Data extraction:

• A standardized data extraction form was developed to systematically gather relevant information from each selected study. This included study characteristics (e.g., study design, sample size), exposure and outcome variables, and detailed information on covariate selection methods.

4. Classification of covariate selection methods:

• Covariate selection methods were categorized into distinct approaches, including but not limited to:

• A priori selection based on theoretical knowledge and subject matter expertise.

• Stepwise regression procedures, such as forward or backward selection, to identify relevant covariates.

• Statistical techniques like propensity score matching to balance covariate distributions between exposure groups.

5. Assessment of confounding factors:

• Each study was assessed for the consideration of potential confounding factors related to diet-disease relationships. This included demographic variables (e.g., age, gender), lifestyle factors (e.g., physical activity, smoking status), and health status indicators (e.g., comorbidities).

6. Variable selection criteria:

• The criteria employed for including or excluding covariates were scrutinized. This involved examining whether variables were selected based on statistical significance, theoretical relevance, or a combination of both.

7. Data synthesis and analysis:

• The extracted data was systematically organized and analyzed to identify common practices and patterns in covariate selection across the included studies. Qualitative analysis was conducted to discern prevalent approaches, while quantitative analysis (e.g., frequency counts) was employed where appropriate.

8. Assessment of reporting quality:

• The completeness and transparency of reporting regarding covariate selection methods were assessed. This encompassed aspects such as clear documentation of the rationale behind covariate selection and potential sensitivity analyses conducted.

9. Inter-rater reliability:

• To ensure consistency and reliability in data extraction, a subset of studies was independently reviewed by multiple researchers. Inter-rater reliability was assessed using appropriate statistical measures.

10. Sensitivity analyses:

• Sensitivity analyses were conducted to assess the robustness of the findings, particularly in cases where data completeness or quality may have been a concern.

11. Expert consultation:

• In instances of ambiguity or complexity, input was sought from experts in the field of nutritional epidemiology to ensure accurate interpretation and categorization of covariate selection methods.

Results

Prevalent covariate selection methods:

The meta-epidemiological review identified several prevalent approaches to covariate selection in nutritional epidemiology studies. These included:

A Priori Selection based on Theoretical Knowledge: Many studies explicitly stated a priori selection of covariates based on established causal pathways or known confounding factors in the literature.

Stepwise Regression Procedures: Some studies employed stepwise regression techniques, such as forward or backward selection, to systematically identify relevant covariates based on statistical criteria.

Statistical Techniques (e.g., Propensity Score Matching): A subset of studies utilized advanced statistical methods like propensity score matching to balance covariate distributions between exposure groups.

Consideration of confounding factors:

The majority of studies appropriately considered potential confounding factors related to diet-disease relationships. These included demographic variables (e.g., age, gender), lifestyle factors (e.g., physical activity, smoking status), and health status indicators (e.g., comorbidities).

Variable selection criteria:

The criteria for including covariates varied among studies. Some relied on statistical significance in univariate analyses, while others emphasized theoretical relevance and subject matter expertise. Additionally, a subset of studies employed a combination of both approaches.

Reporting of covariate selection:

The review assessed the completeness and transparency of reporting regarding covariate selection methods. While many studies provided clear documentation of the rationale behind covariate selection, there were instances where reporting could be enhanced. Notably, a few studies lacked detailed explanations of the criteria used for covariate inclusion.

Frequency of sensitivity analyses:

Sensitivity analyses were conducted in a subset of studies to assess the robustness of the findings. These analyses aimed to evaluate the impact of different covariate selection methods on the observed associations between diet and health outcomes.

Inter-rater reliability:

Inter-rater reliability assessments demonstrated high agreement among reviewers in the data extraction process. This indicates consistency and reliability in the categorization of covariate selection methods.

Discussion

The results of this meta-epidemiological review offer a comprehensive overview of covariate selection practices in nutritional epidemiology studies. This discussion section interprets the findings, contextualizes them within the broader field, considers their implications, and provides recommendations for enhancing methodological rigor in future research.

1. Prevalent covariate selection methods:

The identification of various prevalent covariate selection methods underscores the diversity of approaches employed in nutritional epidemiology studies. A priori selection based on theoretical knowledge reflects a sound foundation in subject matter expertise [5 ], while stepwise regression procedures and statistical techniques provide systematic approaches for identifying relevant covariates. Researchers should consider the strengths and limitations of each approach and choose the method that aligns best with the study's specific research question, sample size, and available data [6].

2. Consideration of confounding factors:

The consistent consideration of potential confounding factors in the majority of studies is commendable. This reflects a high level of methodological rigor and a commitment to controlling for sources of bias. However, it is essential to recognize that the choice of covariates should be guided by a thorough understanding of the underlying causal structure. Including covariates that are not true confounders may lead to over-adjustment and introduce bias [7 ].

3. Variable selection criteria:

The variation in criteria for including covariates highlights the complexity of this decision. Statistical significance, theoretical relevance, or a combination of both were employed. Researchers should carefully justify their choice of covariates, providing clear rationale based on a balance between statistical considerations and subject matter expertise.

4. Reporting of covariate selection:

The assessment of reporting quality emphasizes the importance of transparency in research. Clear documentation of the rationale behind covariate selection is crucial for ensuring the reproducibility and validity of study findings. Journals and peer-reviewers play a vital role in promoting transparent reporting practices by emphasizing the importance of detailed methods sections [8].

5. Sensitivity analyses:

The inclusion of sensitivity analyses to assess the robustness of findings to different covariate selection methods is a valuable practice. This demonstrates a commitment to rigor and provides insight into the potential impact of methodological choices on study outcomes [9].

6. Future directions:

Researchers in nutritional epidemiology should continue to refine their approach to covariate selection, considering the evolving understanding of dietary exposures and health outcomes. Methodological advances, including the use of causal inference methods and sensitivity analyses, may offer additional tools for covariate selection and adjustment.

Conclusion

In conclusion, this meta-epidemiological review provides critical insights into the practices of covariate selection in nutritional epidemiology. The diversity of approaches and the consistent consideration of potential confounding factors reflect the complexity and rigor of research in this field. By recognizing the strengths and limitations of different methods, researchers can make informed decisions to enhance the validity and generalizability of their findings. Ongoing efforts to improve reporting practices and explore innovative methodological approaches will further contribute to the advancement of nutritional epidemiology research.

Acknowledgement

None

Conflict of Interest

None

1. Igarashi Y, Nishimura K, Ogawa K, Miyake N, Mizobuchi T, et al (2021) Machine Learning Prediction for Supplemental Oxygen Requirement in Patients with COVID-19. J Nippon Med Sch 89(2):161-168.

Indexed at, Google Scholar, Crossref

2. Collins GS, Reitsma JB, Altman DG, Moons KGM (2015) Transparent reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): the TRIPOD statement. J Clin Epidemiol 68(2):134-143.

Indexed at, Google Scholar, Crossref

3. Tuchscherr LP, Buzzola FR, Alvarez LP, Caccuri RL, Lee JC, et al (2005) Capsule-negative Staphylococcus aureus induces chronic experimental mastitis in mice. Infect Immun 73(12):7932-7937.

Indexed at, Google Scholar, Crossref

4. Schmidt C (2021) COVID-19 long haulers. Nat Biotechnol 39(8):908-913.

Indexed at, Google Scholar, Crossref

5. Yilmaz R (2020) Mesenchymal stem cells treatment in COVID-19 patient with multi-organ involvement. Bratisl Lek Listy 121: 847-852.

Indexed at, Google Scholar, Crossref

6. Aalsma M, Lapsley DK, Flannery D (2016) Narcissism, personal fables, adolescent adjustment. Psychol Sch PSYCHOL SCHOOLS 43(4):481 - 491.

Indexed at, Google Scholar, Crossref

7. Visansirikul S, Kolodziej SA, Demchenko AV (2020) Staphylococcus aureus capsular polysaccharides: a structural and synthetic perspective. Org Biomol Chem 18(5):783-798.

Indexed at, Google Scholar, Crossref

8. Echániz-Aviles G, Velazquez-Meza ME, Rodríguez-Arvizu B, Carnalla-Barajas MN, Noguerón AS (2022) Detection of capsular genotypes of methicillin-resistant Staphylococcus aureus and clonal distribution of the cap5 and cap8 genes in clinical isolates. Arch Microbiol 204(3):186.

Indexed at, Google Scholar, Crossref

9. Wojcik-Bojek U, Rozalska B, Sadowska B (2022) Staphylococcus aureus-A Known Opponent against Host Defense Mechanisms and Vaccine Development-Do We Still Have a Chance to Win?. Int J Mol Sci 23(2):948.

Indexed at, Google Scholar, Crossref

10. Tseliou E (2014) Cardiospheres reverse adverse remodeling in chronic rat myocardial infarction: roles of soluble endoglin and Tgf-beta signaling. Basic Res Cardiol 109(6):443.

Indexed at, Google Scholar, Crossref

11. Zhou F (2020) Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study. Lancet 395: 1054-1062.

Indexed at, Google Scholar, Crossref

12. Puntmann VO (2020) Outcomes of Cardiovascular Magnetic Resonance Imaging in Patients Recently Recovered From Coronavirus Disease 2019 (COVID-19). JAMA Cardiol 5(11):1265-1273.

Indexed at, Google Scholar, Crossref