ALHT211: Abstract & Background

Abstract provides a concise snapshot of a study: aim, design, main findings, interpretation. ~100–300 words.
Includes sections: Aim, study design, main findings, interpretation, keywords, funding/registration info.
Important: reading an abstract alone is not equivalent to performing a full appraisal.
PRISMA-DTA for Abstracts: a checklist for abstracts of systematic reviews and meta-analyses of diagnostic test accuracy studies.

Purpose: establish the need, ethical considerations, and research gap; frame the study within current literature.
Aims should be clear and specific; PICO (Population, Intervention, Comparator, Outcome) helps align aims with methods and data collection.
Assess whether the narrative leads to a clear statement of aims and whether study variables match aims.
Check for bias in literature selection and relevance to the topic; identify conceptual basis (key prior papers).

Is the study design appropriate for the research question and aims?
Measurement tools: are they reliable and valid?
Sample size and power analysis: sufficient to detect effects?
Sample population and recruitment: description and potential biases; allocation bias in intervention studies; concealment if applicable.
Design type: superiority, equivalence, noninferiority; presence of control or comparison groups; replication potential.
Biases addressed: observer/participant biases, setting consistency, treatment fidelity.
Data analysis: are statistical methods and outcome measures clearly described and justified?
Reproducibility: sufficient detail to replicate the study.
Related guidelines: Polgar (Ch.22) reference; consider CONSORT, CHERRIES, PRIMSA where relevant.

Structure: descriptive statistics followed by inferential statistics.
Transparency: report protocol deviations and how dropouts or missing data were handled (e.g., missing data methods).
Data presentation: use of tables, graphs, and figures; ensure alignment with Methods and aims.
Apply general statistical literacy:
- Significance: p-values and threshold often p \leq 0.05 unless stated otherwise.
- Trends: a reported trend is not equivalent to statistical significance.
- Effect sizes: report and interpret correlation coefficients r and other effect sizes.
- Confidence intervals: discuss CI and ranges of outcomes.
- Biases: remain vigilant for biases in reporting and analysis.

Discussion: author interpretation of results; consider overgeneralisation and both statistical and clinical significance (in light of study power).
Compare findings with previous research.
Strengths and limitations: openly discuss what strengthens the study and what may limit generalizability.
Conclusion: succinctly state what was found, whether aims were answered, and what was proven vs. not proven.
Future directions: what research is needed next; clinical implications: how findings should influence practice.

Correlation indicates association between variables, not causation.
Causation requires evidence of mechanism, temporality, and control for confounding factors; typically requires experimental or robust quasi-experimental design.
In practice, beware misinterpretation of correlational findings as causal conclusions.

Read title/abstract to gauge relevance and believability.
Read full paper if relevant; assess replicability and applicability to practice.
Store/share findings and discuss with stakeholders (clients, colleagues).
Use a structured approach: synthesis, appraisal, application.
Tools and guidelines to keep in mind:
- CONSORT, CHERRIES, PRIMSA for study design reporting.
- CASP (Critical Appraisal Skills Programme) for bias and quality assessment.
- PICO framework to map population, intervention, comparator, and outcomes.
- PRISMA-DTA for abstracts of diagnostic accuracy reviews.

PICO: Population, Intervention, Comparator, Outcome.
Study design types: superiority, equivalence, noninferiority.
Bias and fidelity: allocation bias, observer bias, participant bias, treatment fidelity.
Statistical concepts: p-value, statistical significance, correlation coefficient r, confidence interval CI.
Reporting guidelines: CONSORT (randomised trials), CHERRIES (web-based surveys), PRIMSA (systematic reviews), PRISMA-DTA (diagnostic accuracy).

Abstract sections to look for: Aim, Design, Key findings, Interpretation, Keywords.
Background should justify the study and clarify aims using a narrative that may rely on prior papers (Polgar et al., 2024).
Method should justify design choices and describe sampling, measurements, and analysis clearly enough for replication.
Results should align with Methods, report dropouts/missing data, and present descriptive and inferential statistics with appropriate visuals.
Discussion should weigh significance, comparability to prior work, and limitations, plus practical implications and future directions.
The “Read title/abstract” strategy emphasizes quick relevance checks and practical applicability.