12 Final LMM: Reporting

Report the salient outcomes of the model that help in interpretation and draw appropriate conclusions.

Some reporting guidelines are included here. Note that these are not exhaustive. These are general suggestions.

• Present a clear statement on primary and secondary research objectives.

• Describe the experimental design, sample size, data structure and possible sources of variations.

• Present the statistical modelling approach, fixed and random effects structure and model selection strategies.

• Explain estimation methods and other statistics used to interpret the model outcomes.

• Provide the details of the model development stages and scripts as supplementary material so that your model is reproducible on the same data.

• Present the summary of demography information about the patients.

• Tabulate the estimates of fixed and random effects from the final model with standard error, confidence interval, p-value as appropriate.

• Plot the effect size along with 95% confidence intervals

• Present the summary plots with estimated marginal means along with confidence intervals

• Sometimes, even summary plots of raw data along with estimated marginal means are also useful.

• Interpret the results and draw appropriate conclusions in line with the research objectives.

• Present estimated marginal means for selected main effects.

• If the multiple comparison of means is conducted, provide additional details.

• In general, random effects of each level of the cluster variable is not presented, but in certain scenario, the prediction of random effects will be useful and can be summarised graphcally.

Figure: Summary of outputs from the final fitted linear mixed model on SBP data