Transcript Sample Slide Heading Image

Corso di clinical writing
What to expect today?
Core modules
•
Introduction
•
Correction of abstracts submitted by participants
•
Practical session 2 – Abstract drafting
•
Results drafting
•
Discussion drafting
•
Tables and Figures drafting
•
Peer review and publication
•
Synthetic example – database creation, analysis and
Results drafting
Results
What were the findings?
Results
What were the findings?
The answer is in the Results.
Expanded IMRAD algorithm
Introduction
Background
Limitations of current evidence
Study hypothesis
Methods
Design
Patients
Procedures
Follow-up
End-points
Additional analyses
Statistical analysis
Results
Baseline and procedural data
Early outcomes
Mid-to-long term outcomes
Additional analyses
Discussion
Summary of study findings
Current research context
Implications of the present study
Avenues for further research
Limitations of the present study
Conclusions
Results
1. Logically answer the research question
2. Begin with most important result
3. Correlate with the methods
4. Use data from this study only
5. Present all the representative data (with exact P
values and confidence intervals)
6. Use tables, graphs, photographs, and drawings
for data
Results
 Show subject characteristics as n/N (%) and means±
standard deviations (or median [interquartile range)
 In an intervention:
• show pre-test means±SD as subject characteristics;
• show change-score means and SDs to give an
impression of any individual responses;
• show differences in mean changes, with 95%
confidence intervals;
• calculate any individual responses as a standard
deviation
Results
 Show scattergrams only for a good reason (outliers, nonuniformity of error, non-linear trend): be aware that they
can unmask you!
 Summarize multiple outcomes in a figure or table
 (Almost) never show test statistics (t, F, 2)
 Try to show chances of clinical/practical substantiveness
 Avoid repetition of outcomes in figures, tables, or text
 MOST IMPORTANTLY: do not discuss the findings or
interpret them qualitatively!
Results
• Supplement rather than repeat data in visuals and
tables:
– Data must agree within the section and with data
given in other sections and visuals
• Avoid too much dryness and overwhelming the reader
with data:
– The mean BMI was 10% lower in the 30 tennis
players than in the 20 control subjects (respectively
22.5±0.8 vs 24.9±0.7 mm Hg, P=0.035).
• Should be simply stated (past tense)
Patient and procedural data
Zehetner et al,
Surg Endosc 2005
Patient and procedural data
O’Brien et al,
Ann Intern Med 2006
Patient and procedural data
Flum et al,
JAMA 2005
Exclusions
O’Brien et al,
Ann Intern Med 2006
Early outcomes
Nelson et al, NEJM 2004
Early outcomes
O’Brien et al, Ann Intern Med 2006
Early outcomes
Ceelen et al, Ann Surg 2003
Late outcomes
Nelson et al, NEJM 2004
Late outcomes
Ceelen et al, Ann Surg 2003
Late outcomes
Nelson et al,
NEJM 2004
Late outcomes
Nelson et al,
NEJM 2004
Do NOTs
• Use big words that you do not really mean
– Attributable
– Causality
– Preferential
– Significant (without statistical evidence)
– Validity
• Mix incidence and prevalence
• Mix frequency, rate, proportion, ratio
Questions?
Take home messages
1. Results should provide a coincise yet
poignant synthesis of what you,
apparently
undexpectedly
and
independently, have found
2. Results and Discussion should
appear as written by two separate
individuals!
And now let’s move to Discussion…