Treatment
Nendyah Roestijawati
Ideas and evidence
Studies of treatment effects
Observational and experimental
studies
Experimental study cohort study
which conditions of study are
specified for making unbiased
comparisons
Selection treatment group
Nature of interventions
Management follow up
Measuring outcomes
Randomized Controlled Trials
Standard of excellence of scientific
study of the effects of treatment
Same as cohort study except that
treatment is assigned by
randomization rather than
physician and patient
Structure of RCT
Ethics
No compelling reason to believe that either
randomly allocated treatment is better than other
equipoise
Experimental intervention might be better than
control
Primary outcomes must be benefit cannot
randomly one is more harmful for other
Patients must fully understand the consequences of
participating the study
Trial must be stopped if there is effectiveness, harm,
futility in continuing
Sampling
To increase homogeneity, strengthen
internal validity, distinguish signal
and noise
Inclusion criteria : patients really do
have the condition being studied
strict diagnostic criteria
Sampling
Exclusion criteria :
Comorbidity
Not expected live long enough
Contraindications of one treatments
cannot randomized
Refuse to participate
Do not cooperated
Intervention
Generalizability can be easily
described and reproduced in other
setting
Complexity : single weak effects,
multifaceted more effective and
amenable
Strength : different from alternative
managements and reasonable to
expect that the outcome will be
affected
Comparison groups
No intervention
Being part of study Hawthorne
effect
Usual care
Placebo treatment placebo
effect
Another intervention current use
Changes of outcomes
Allocating treatment
Random allocation/randomization
free of confounding
Baseline characteristics
comparing frequency in treat and
control group
Stratified randomization patients
gathered into groups with similar
levels of prognostic factors
Baseline characteristic
Characteristics Liberal Restricted
Age 81.8 81.5
Male 24.8 23.7
Any cardiovascular disease 63.3 62.5
Tobacco use < 600 mg/d 11.6 11.3
Anaesthesiology risk score 3.0 2.9
General anaesthesia 54.0 56.2
Lived in nursing house 10.3 10.9
Stratified randomization
Differences arising after randomization
Patients may not have disease
Compliance
Cross-over
Co-interventions
Blinding
To minimize change behaviour
because they aware which
treatment received
Single blind and double blind
ambiguous
Open label/trial : no attempt to
blind
Assessment of outcomes
Composite outcome
Performance status Definition
0 Asymptomatic
1 Symptomatic, fully ambulatory
2 Symptomatic, in bed <50% of
the day
3 Symptomatic, in bed >50% of
the day
4 Bedridden
5 Dead
Assessment of outcomes
Treatment effects
Measure Definition
Relative Risk Control event rate - treatment event rate
Reduction Control event rate
Absolute Rate Control event rate treatment event rate
Reduction
Number needed 1
to treat or --------------------------------------------------------
number needed Control event rate treatment event rate
to harm
Efficacy and effectiveness
Efficacy trials (ideal) : patient offered
interventions, follow instruction, get best
possible care, do not care other disease
randomized
Effectiveness trials (ordinary): patient not
take the assigned treatment, doctor
and facility not the best
Implementation gap : difference
between efficacy and effectiveness
Result analysis
Best choice
treatment
Experiment
treatment is
better
Superiority and non inferiority trials
Superiority trials : one treatment better
than another
Non-inferiority trials : new treatment
unlikely to be less effective than current
treatment but in other studies to be more
effective than placebo inferiority
margin
Variations on randomized trials
Cluster randomized trials doctors,
hospitals, clinics, communities
Cross over trials control for
differences response not related
treatment effect
Tailoring results to individual patients
Subgroups
Effectiveness in individual patients
efficacious, effective, working,
benefits worth discomfort and risk
Trials N=1 rarely done and
publish
Limitations randomized trials
Does not settle specific question
Expensive
Phases of clinical trials
I. Identify a dose range that
tolerated and safe
II. Efficacious and relationship
between dose and efficacy
III. Randomized trials : efficacy and
rates of common side effects
Post marketing surveillance
Terimakasih
