Research Methodology

📚 04 PSM & Preventive Pediatrics • 🎯 78 High-Yield MCQs • Nelson 22nd Edition Standard

Research Methodology, Study Design, and Literature Search Fundamentals

  • Study Protocol Eligibility Criteria: ((RM001)): Inclusion and exclusion criteria must be very specific to ensure appropriate population selection, improve internal validity, minimize confounding, and maximize reproducibility.
  • Pilot Studies: ((RM002)): Small-scale preliminary investigations used to evaluate research instruments, sampling frames, and feasibility. Crucial: Ethical approval for pilot and main studies are obtained independently; successful completion of a pilot study is NOT a mandatory prerequisite for main study approval.
  • Mandatory Ethical Approval: ((RM003)): Ethics Committee approval is universally mandatory before initiating any research involving primary data collection from human participants to guarantee rights, safety, dignity, and well-being.
  • Effect Modifiers & Literature Review: ((RM004)): A thorough review of literature is the best evidence-based source for identifying effect modifiers (variables altering the exposure-outcome relationship) and their biological plausibility.
  • Implementation Research: ((RM005)): Scientific inquiry that evaluates the acceptability, functionality, feasibility, and real-world performance of a health program under routine community conditions.
  • Bench-Based Research: ((RM006)): Preclinical or basic science research conducted strictly in a laboratory setting using animals, cell cultures, or molecular techniques to explore fundamental mechanisms.
  • Translational Research Scope: ((RM007)): Translational study is not a formal study design; it describes a broad research continuum moving scientific discoveries from "bench to bedside" and into public health application.

Research Errors, Biases, and Validity

  • Random vs. Systematic Errors: ((((RM008)) · ((RM010)))):
  • Random errors:: Arise from chance variation and lead to imprecision; they cannot be completely eliminated (though reduced by larger sample sizes).
  • Systematic errors (bias):: Push study results consistently in one direction. Minimized by instrument calibration, standardization, and rigorous design.
  • Critical Appraisal: ((RM015)): The systematic application of analytical principles to evaluate research evidence for validity, reliability, and risk of bias before clinical application.

Research Questions, Objectives, and Information Seeking

  • Analytical Study Objectives: ((RM011)): The verb “determine” is the preferred terminology in objectives of analytical studies (evaluating associations/causal links), whereas “estimate” is used for descriptive studies.
  • Research Life Cycle: ((RM012)): The first step in the research life cycle is identifying data needs (recognizing missing information and determining if new data collection is required) prior to framing the research question.
  • FINER Criteria for Research Questions: ((RM013)): Evaluates quality using: Feasible, Interesting, Novel, Ethical, and Relevant. (Note: Reliable is not a component of FINER).
  • Descriptive Research Question: ((RM014)): Focuses solely on observations to measure a quantity (prevalence, incidence, demographic patterns) without comparison groups or hypothesis testing.
  • Information Seeking & MeSH: ((((RM016)) · ((RM017)) · ((RM020)))):
  • Information seeking:: Efficiently scanning literature via manual or computerized methods.
  • MeSH (Medical Subject Headings):: Controlled vocabulary produced by the National Library of Medicine (NLM) structured hierarchically for MEDLINE/PubMed. MeSH is not used for EMBASE (which uses EMTREE).
  • Literature Search & Manuscript Flow: ((((RM018)) · ((RM019)))): Sequence: Organize information $\rightarrow$ Synthesize results $\rightarrow$ Identify lacunae (gaps) $\rightarrow$ Develop research question. Literature review is essential for Introduction, Methods, and Discussion, but never used in the Results section.

Epidemiological Measures, Confounding, and Sampling

  • Disease Prevalence Limitations: ((RM021)): Point prevalence is the ratio of existing cases to the population at risk. Prevalence is inappropriate for measuring the disease burden of high-fatality conditions because affected patients die quickly and drop out of the existing case pool.
  • Confounding and Analytical Control: ((((RM022)) · ((RM023)))): Confounders distort true associations and can exaggerate, mask, or diminish the effect (though they do not always change the direction of the effect). Multivariate analysis is the primary method to control confounding during data analysis.
  • Qualitative Research Methods: ((((RM024)) · ((RM025)))): Characterized by being subjective, credible, inductive, and focused on interpretation. Used for in-depth understanding of behaviors, social phenomena, and unexplored areas; not used for familiar, heavily researched topics.
  • Sampling Error & Sample Size Calculations: ((((RM026)) · ((RM027)) · ((RM028)) · ((RM029)) · ((RM030)))):
  • Sampling error:: Can only be measured in probability samples (where selection probability is known).
  • Cluster sampling:: Requires a list of clusters (villages, wards), not a complete list of individual units.
  • Quota sampling:: A recognized form of nonprobability sampling.
  • Effect size vs Sample size:: An inverse relationship—as the expected effect size increases, the required sample size decreases.
  • Accessible population:: The segment of the target population defined by geographical and temporal characteristics from which the sample is drawn.

Study Execution, Tools, and Ethical Conduct

  • Analysis Plan: ((RM031)): Ensures analyzable information is gathered and prevents irrelevant comparisons, but does not determine data quality (which depends on data collection tools, training, and standard operating procedures).
  • Study Questionnaire Design: ((RM032)): Questions must be concise, precise, neutral, free of double negatives, and devoid of scientific or technical jargon.
  • Data Cleaning: ((RM033)): Includes checking for duplicates, handling missing data, and verifying legal ranges. Does not include calculating the strength of association (which belongs to the formal statistical analysis phase).
  • Informed Consent (IC): ((RM034)): Must be voluntary, informed, competent, and properly documented. Consent obtained via coercion is legally and ethically invalid.
  • Concept Paper: Background and Justification Sequence: ((RM035)): Structured logically as: Importance of the study problem $\rightarrow$ Known and unknown aspects $\rightarrow$ Information that needs to be generated.

Research Methodology, Ethics, and Data Management

  • Protocol Development Sequence: ((RM036)): Research documents follow a strict developmental hierarchy. The process must begin with drafting a one-page concept proposal (concept note) to outline study ideas and objectives. Only after the concept note is accepted does the researcher develop the full, detailed study protocol required for ethics committee approval.
  • Publication Ethics & Negative Findings: ((RM037)): Negative, null, or non-significant findings must be published. Suppressing negative results leads to publication bias, skews systematic reviews/meta-analyses, and causes unnecessary duplication of failed experiments.
  • Publication Misconduct vs. Conflict of Interest (COI): ((RM038)): Conflict of interest (COI) (financial, professional, or personal ties) is not publication misconduct provided it is fully disclosed. True publication misconduct includes plagiarism (copying ideas/text), ghost authorship (uncredited contributors), and data fabrication/falsification.
  • Validity Types in Epidemiology & Research: ((((RM039)) · ((RM054)) · ((RM061)))):
  • Construct Validity: : Assumes a causal relationship between variables and determines whether the theoretical explanation best fits the measured constructs ((RM039)).
  • Internal Validity: : The degree to which a study accurately reflects the true relationship between exposure and outcome within the study population, maximized by minimizing bias, controlling confounders, and reducing measurement errors ((RM054)).
  • External Validity: : The degree to which study inferences can be generalized to broader populations or other external settings ((RM061)).
  • Study Designs & Epidemiological Measures: ((((RM040)) · ((RM043)) · ((RM045)) · ((RM047)) · ((RM053)))):
  • Field Trials: : Evaluates the effect of an intervention (e.g., vaccine introduction like pentavalent vaccine) under real community/population conditions over time ((RM040)).
  • Descriptive Research Questions: : Focuses solely on measuring the frequency, pattern, and distribution of disease (e.g., estimating prevalence or incidence) without comparing groups ((RM043)).
  • Analytical Research Questions: : Assesses exposure–outcome relationships, associations, or causal effects between groups (e.g., assessing the effect of tobacco cessation on TB treatment outcomes) (((RM045)) · ((RM047))).
  • Incidence: : The most useful epidemiological measure for evaluating the impact of a primary prevention programme as it captures newly diagnosed cases over time ((RM053)).
  • Bias, Confounding, & Effect Modification: ((((RM041)) · ((RM042)) · ((RM055)) · ((RM056)) · ((RM067)))):
  • Bias: : Systematic errors introduced during participant recruitment or data collection that distort the true association ((RM041)).
  • Prevarication: : A specific type of information bias where participants intentionally provide false or misleading information (e.g., underreporting alcohol use) ((RM055)).
  • Confounding: : Occurs when a third factor distorts the true association. Identified via stratified analysis when stratum-specific risk ratios are identical to each other but ==significantly different (e.g., shifted to null, RR = 1)== from the distorted crude estimate (((RM056)) · ((RM067))).
  • Effect Modifier: : A third variable that genuinely changes the magnitude of the effect of the exposure on the outcome across different subgroups (e.g., anticholinergic drugs modifying antipsychotic-induced extrapyramidal side effects) ((RM042)).
  • Literature Review & Information Retrieval: ((((RM049)) · ((RM050)) · ((RM051)) · ((RM052)))):
  • Information Retrieval: : The process of using keywords, search filters, and databases (PubMed, Scopus, Web of Science) to locate relevant scientific literature (((RM049)) · ((RM052))).
  • PubMed: : A completely free, non-paid biomedical search engine maintained by NCBI at the U.S. National Library of Medicine, containing millions of citations with full-text links via PubMed Central (PMC) ((RM051)).
  • Purpose of Literature Review: : Used to identify knowledge gaps, prevent research duplication, and shape research questions. It cannot directly generate or replace the final conclusions of a new study ((RM050)).
  • Sampling, Sample Size, & Questionnaire Design: ((((RM059)) · ((RM060)) · ((RM063)) · ((RM064)))):
  • Systematic Random Sampling: : Selecting every $k^{th}$ subject after a random start; yields a sample that is evenly spread and potentially more representative than simple random sampling ((RM059)).
  • Sample Size for Proportions: : Calculated using $n = \frac{Z^2pq}{d^2}$ (where $p$ = anticipated proportion, $d$ = absolute precision, $Z$ = confidence level) ((RM060)).
  • Questionnaires: : Must incorporate skip patterns/branching for applicability and flow ((RM063)).
  • Pilot Testing: : Must never be conducted on the same participants who will later be enrolled in the main study to avoid contaminating the sample ((RM064)).
  • Qualitative Research Methods: ((((RM057)) · ((RM058)))):
  • Content Analysis (Deductive): : Analyzing qualitative data by coding transcripts using a pre-existing theoretical framework or predefined categories ((RM057)).
  • Focus Group Discussions (FGDs): : Ideal qualitative method for exploring shared collective practices, beliefs, and challenges when direct subjects (e.g., toddlers) cannot articulate responses themselves ((RM058)).
  • Data Analysis Preparation: ((((RM065)) · ((RM066)))): - Initial profiling of a study population requires calculating frequency distributions for categorical variables ((RM065)). - Proper data management includes recoding variables into simplified categories prior to launching statistical tests ((RM066)).
  • Research Ethics & Safety Monitoring: ((((RM068)) · ((RM069)) · ((RM070)))):
  • Autonomy: : The ethical principle granting participants the absolute right to withdraw from a study at any point without coercion; violated if an investigator pressurizes continuation ((RM068)).
  • Data Safety Monitoring Board (DSMB): : An independent committee with access to unblinded interim trial data tasked with periodically reviewing participant safety, trial progress, and making recommendations on continuation, modification, or termination ((RM069)). Tracking operational parameters like enrollment targets is not a DSMB safety function ((RM070)).

Research Methodology & Scientific Integrity

  • Concept Paper Fundamentals: ((RM071)): A concept paper is a brief, focused preliminary document outlining the problem, knowledge gaps, key objectives, and basic methodology. It must not include detailed discussions of anticipated study findings ((RM071)) (since research has not yet been conducted).
  • Operational Definitions: ((RM072)): A dedicated section in a research proposal where key exposure and outcome variables are precisely defined using measurable and standardized criteria (e.g., specific cutoffs for high salt intake or hypertension) to ensure uniform interpretation by all data collectors ((RM072)).
  • Human Subject Protection: ((RM073)): Ethical safeguards, participant rights, safety, privacy, and confidentiality measures must be explicitly stated in the methods section of a concept paper or research proposal ((RM073)).
  • ICMJE Authorship Criteria: ((RM074)): Authorship is justified by making a meaningful intellectual contribution (e.g., study design, refining objectives, tool selection, or drafting the manuscript) ((RM074)). Departmental boundaries do not restrict valid authorship ((RM074)).
  • Salami Slicing: ((RM075)): Unethical publication practice involving the intentional division of a single substantial study into multiple smaller papers without scientific justification, solely to artificially inflate the publication count ((RM075)).
  • Plagiarism & Citation: ((RM076)): Paraphrasing another author's ideas into one's own words still requires citing the reference ((RM076)). Plagiarism cannot be resolved through rephrasing alone without proper attribution ((RM076)).
  • Target Population for Generalization: ((RM077)): The target population represents the entire group of individuals to whom the researcher ultimately wants to apply or generalize the study findings, which is broader than the studied sample ((RM077)).
  • Study Reporting Guidelines: ((RM078)): - CONSORT = Randomized controlled trial ((RM078)) - STROBE = Observational study (cohort, case-control, cross-sectional) ((RM078)) - PRISMA = Systematic review and meta-analysis ((RM078))