For my own use: 1) Research the disease - know the various established treatments using guidelines, first-line drugs etc. 2) Assess for bias - internal validity (whether the outcome is due to intervention) - blinding, randomization, having control groups, ensuring correct dosages were used, outcomes objectively assessed aka the study methods. Also want to know who funded the study and what their role was in study design, data collection, data analysis. 3) Assess inclusion & exclusion criteria - does inclusion criteria reflect the population the drug will be used on, did the study exclude those who might use this drug in practice. Must be able to rationalize exclusion criteria, else if too many excluded may weaken external validity. 4) Consider other factors - drug cost vs existing treatments, efficacy worth its price?, affordability, should dose have been titrated when the drug was started/discontinued, did the study assess adherence and how? 5) Assess whether the right statistical tests were used based on the outcomes that were studied, type of data, study design, whether there were confounders. "Stratification, covariates, adjusted for" = confounders are present. Internal validity improves with the right statistic tests chosen 6) Supplemental material eg Annexes - may contain useful information 7) Assess results for clinical AND statistical significance - confidence interval. P value only tells you statistical significance. Narrow confidence interval = more precise. Learn how to interpret OR HR RR CI, Power, Type I & II errors. 8) Dont ignore tables, graphs, footnotes. 9) Assess for risk of error - If p-value 0.05 ie not statistically significant, then check if the results were falsely negative ie check power (if the power was sufficient (>80%), sufficient sample size, effect size ie the difference in effect they expected to see). If yes to all 3 qns, then the negative results could mean that the drug dose was not effective. But if have a no to any 1 of the 3 qns, then say results inconclusive due to low power. 10) Don't read discussion and conclusion until you have created your own ideas on the study as they are the authors own opinion of their study
For my own use:
1) Research the disease - know the various established treatments using guidelines, first-line drugs etc.
2) Assess for bias - internal validity (whether the outcome is due to intervention) - blinding, randomization, having control groups, ensuring correct dosages were used, outcomes objectively assessed aka the study methods. Also want to know who funded the study and what their role was in study design, data collection, data analysis.
3) Assess inclusion & exclusion criteria - does inclusion criteria reflect the population the drug will be used on, did the study exclude those who might use this drug in practice. Must be able to rationalize exclusion criteria, else if too many excluded may weaken external validity.
4) Consider other factors - drug cost vs existing treatments, efficacy worth its price?, affordability, should dose have been titrated when the drug was started/discontinued, did the study assess adherence and how?
5) Assess whether the right statistical tests were used based on the outcomes that were studied, type of data, study design, whether there were confounders. "Stratification, covariates, adjusted for" = confounders are present. Internal validity improves with the right statistic tests chosen
6) Supplemental material eg Annexes - may contain useful information
7) Assess results for clinical AND statistical significance - confidence interval. P value only tells you statistical significance. Narrow confidence interval = more precise. Learn how to interpret OR HR RR CI, Power, Type I & II errors.
8) Dont ignore tables, graphs, footnotes.
9) Assess for risk of error
- If p-value 0.05 ie not statistically significant, then check if the results were falsely negative ie check power (if the power was sufficient (>80%), sufficient sample size, effect size ie the difference in effect they expected to see). If yes to all 3 qns, then the negative results could mean that the drug dose was not effective. But if have a no to any 1 of the 3 qns, then say results inconclusive due to low power.
10) Don't read discussion and conclusion until you have created your own ideas on the study as they are the authors own opinion of their study
this was extremely helpful thank you so much!
Thanks 🙏 Doctor
very nice video! very helpful
hi, can i know what is dental journal clubs??
Madam, Your drawing is as good as your discussion.
Thank you
How we get the access to read journal?
Nice video!