A methodological review protocol of the use of Bayesian factor analysis in primary care research

Articles from MEDLINE, Embase, Cochrane Library, CINAHL, and Scopus will be identified. All relevant articles published before January 1, 2020, will be considered.

Google Scholar will be manually scanned for the first 200 to 300 records for supplementary information [12]. Reference lists and the future citation of the retrieved articles will be manually searched with two additional rounds. As currently no well-established guidelines exist on conducting methodological reviews, we will follow the general recommendations in the literature [13], the guidelines under development [14], and the PRISMA-P statement for the methodological review [15, 16]. Review of review articles will serve for identifying the literature covered within the reviews.

Quantitative and empirical research studies, methodological studies using Bayesian factor analysis, review articles, conference abstracts, and thesis or dissertation documents will be included. Research studies using similar model structures such as structural equation models and latent variable models, as well as item response theory, factor loadings, and item domain correlations, will be included. Some conferences publish full-text papers, i.e., conference proceedings alongside with abstracts. Only conference abstracts with respective full-text access will be considered in the methodological review.

The inclusion criteria of the review require literature to match the following three themes: “in the context of primary care practice,” “Bayesian methods,” and “factor analysis.” The definition of primary care follows that of the American Academy of Family Physicians as being “comprehensive,” “first contact,” and “continuing”; meanwhile, it covers “any undiagnosed sign, symptom, or health concern” [17]. The term “Bayesian methods” refers to any inferential method that employs prior distributions in conjunction with observed information to arrive at an estimate for the parameter of interest.

Editorials, commentaries, book reviews, hypotheses, critical appraisals, reflections, surveys, case reports or studies, or studies that do not employ Bayesian methods will be excluded. Studies that include some of the keywords but use them under different connotations or references will be excluded. Examples of ineligible use of keywords include “primary studies,” “prior to,” “human epidermal growth factor,” and “genetic factor.”

Bayesian methods used in other types of analyses, such as Bayes rule, Bayes or Bayesian factor studies, variational Bayes, Bayesian Information Criterion/Criteria, Bayesian random effects models, Bayesian/Bayes network, belief network, and Bayes(ian) model or probabilistic directed acyclic graphical model will be excluded. Studies not in family medicine or primary care but using related terminology will be excluded. Examples are “a family of methods” and “exponential family.”

Titles and abstracts of studies will be sequentially screened using the search strategy by at least two independent reviewers using the software Rayyan [18]. If no information is given in the title or abstract about any of the three inclusion criteria, i.e., no indication about whether the study is applying Bayesian methods, using factor analysis, or in primary care, those studies will be included at the initial stage of screening. In indecisive situations, for example, when the term “factor analysis” is mentioned but not specified whether it is Bayesian or not Bayesian, the article will be kept for the next round of full-text review. The full text of articles that meet the inclusion criteria will be retrieved and examined independently by k>2 reviewers, each reviewing one out of the k portions of the identified articles that are randomly assigned to each reviewer. All articles will be also reviewed by the main author. Any disagreement between the reviewers and the main author about the eligibility of specific studies will be discussed, and an additional reviewer will be involved if necessary, until a consensus is reached. For studies with multiple publication records, the most comprehensive or up-to-date record will be used.

Data extraction and data preparation will be facilitated using Microsoft Excel and the statistical software package R. All records will be coded and categorized under the predefined themes in the codebook from the Canadian Institute of Health Research (CIHR) grants and rewards guide [19]. Despite existent guidelines and recommendations on reporting of general Bayesian methods, confirmatory factor analysis, and questionnaire development, no single comprehensive recommendation was found on the reporting of Bayesian confirmatory factor analysis [20‐22]. Where applicable, the following data will be extracted: type of journal, publication date, geographical location, sample size, number of items or questions used for the Bayesian factor analysis, number of factors, domains or constructs, reported item-domain correlations, regression parameters, factor loadings, parameters of structural equation models, use of prior information and assumed prior distributions, and the primary care settings. A standardized predesigned data collection form will be used for data extraction. The assessment criteria below will be followed:

The data extraction form used to summarize information obtained from the identified articles will be pilot tested to identify possible sources of error or imprecision. For this purpose, all reviewers involved will extract the data from a selected set of articles using the data extraction form. The extracted data will then be compared and sources for potential mismatches or errors discussed and resolved.

The risk of bias in individual studies is not applicable to and will not be assessed in the review since the goal is to summarize the use and reporting of Bayesian questionnaire validation methods, i.e., there is no single effect parameter that is of primary interest. The data collected across studies will indicate the presence or absence of each of the seven criteria for assessing the appropriateness of design, conduct, and reporting. The quality of implementation and reporting of Bayesian methods for each eligible study will be assessed and rated on an ordinal scale with the following levels: very low, low, moderate, and high on the following aspects: reporting about methodology, Bayesian model, estimated parameters, prior elicitation, and basic contextual information provided. The quality assessment will be conducted independently by two expert statisticians (H.Z. and T.S.) and presented in tables in the final publication of the methodological review. No available critical tools exist to appraise the use of Bayesian factor analysis; however, the proposed quality appraisal (i.e., a methodological “peer review”) by the authors will help to identify prevalent issues and initiate discussions of better reporting standards.