Critical Appraisal for Therapy Articles

STUDENT NAME :
STUDENT NUMBER :

Title of article: Twice weekly fluticasone propionate added to emollient

maintenance treatment to reduce risk of relapse in atopic dermatitis:
randomised, double blind, parallel group study

What question did the study ask?

Patients : Adult (aged 12-65) patients with moderate to severe
atopic dermatitis who were experiencing a flare.
Intervention : Fluticasone propionate (0.05% cream or 0.005%
ointment) applied twice weekly in addition to emollient maintenance treatment.
Comparison : Emollient alone applied twice weekly in addition to
emollient maintenance treatment.
Outcome(s) : Time to relapse of atopic dermatitis, risk of relapse,
adverse events.

Are the results of the trial valid? (Internal Validity)

1a. R- Was the assignment of patients to treatments randomized?

What is best? Where do I find the information?
Centralized computer randomization is ideal and The article states that "patients were randomised (by
often used in multi-cantered trials. Smaller trials SAS software) and stratified within each centre into
may use an independent person (e.g, the hospital high (12-15) and low (8-11) GMSPS groups".
pharmacy) to “police” the randomization.
This paper: Yes

Comment:
Yes, the assignment of patients to treatments was randomized.

1b. R- Were the groups similar at the start of the trial?

What is best? Where do I find the information?
If the randomization process worked (that is, The article states that "patients were randomised (by
achieved comparable groups) the groups should be SAS software) and stratified within each centre into
similar. The more similar the groups the better it is. high (12-15) and low (8-11) GMSPS groups". The
There should be some indication of whether article also reports the baseline characteristics of the
differences between groups are statistically patients in each group, such as age, sex, disease
significant (ie. p values). severity, and previous use of topical corticosteroids,
and shows that there were no significant differences
between the groups. This means that the groups were
comparable and balanced at the beginning of the
trial, and any differences in the outcomes can be
attributed to the treatments.
This paper: Yes
Comment:
Yes, the groups were similar at the start of the trial.

2a. A – Aside from the allocated treatment, were groups treated equally?
What is best? Where do I find the information?
Apart from the intervention the patients in the The article states that "patients applied emollient on
different groups should be treated the same, eg., a daily basis with a bath oil as needed and either the
additional treatments or tests. same formulation of fluticasone propionate or its
placebo base (emollient alone) twice weekly to the
areas that were usually affected". This means that
the groups received the same care and instructions,
except for the addition of fluticasone propionate or

1
University of Oxford, 2005
Critical Appraisal for Therapy Articles

placebo. The article also states that "the groups

showed no differences in adverse events", which
implies that the treatments did not cause any harm or
imbalance between the groups.
This paper: Yes
Comment:
Yes, aside from the allocated treatment, the groups were treated equally.

2b. A – Were all patients who entered the trial accounted for? – and were they analyzed
in the groups to which they were randomized?
What is best? Where do I find the information?
Losses to follow-up should be minimal – preferably The article states that "patients were randomised (by
less than 20%. However, if few patients have the SAS software) and stratified within each centre into
outcome of interest, then even small losses to high (12-15) and low (8-11) GMSPS groups".
follow-up can bias the results. Patients should also GMSPS stands for Glasgow meningococcal sepsis
be analyzed in the groups to which they were prognostic score, which is a tool to assess the
randomized – ‘intention-to-treat analysis’. severity of meningococcal sepsis. The article also
reports the flow of patients through the trial, and
shows that there were no losses to follow-up or
exclusions from the analysis. This means that the
trial followed the intention-to-treat principle, which
is a method to avoid bias and preserve the benefits
of randomization.
This paper: Yes
Comment:
Yes, all patients who entered the trial were accounted for and analyzed in the groups to which they were
randomized.

3. M - Were measures objective or were the patients and clinicians kept “blind” to
which treatment was being received?
What is best? Where do I find the information?
It is ideal if the study is ‘double-blinded’ – that is, The article states that "the study was double blind,
both patients and investigators are unaware of with the investigator, patient, and GlaxoSmithKline
treatment allocation. If the outcome is objective personnel unaware of the treatment allocation". The
(eg., death) then blinding is less critical. If the article also states that "the primary outcome measure
outcome is subjective (eg., symptoms or function) was time to relapse of atopic dermatitis during the
then blinding of the outcome assessor is critical. maintenance phase, defined as a score of 2 or more
on the investigator's global assessment". The
investigator's global assessment is a validated scale
that rates the severity of atopic dermatitis based on
objective criteria, such as erythema, oedema,
excoriation, and lichenification. Blinding and
objective measures are methods to reduce the risk of
bias and increase the validity of the result.
This paper: Yes
Comment:
Yes, the measures were objective and the patients and clinicians were kept blind to which treatment was
being received.

2
University of Oxford, 2005
Critical Appraisal for Therapy Articles

What were the results?

1. How large was the treatment effect?

The precision of the estimate of the treatment effect can be assessed by looking at the confidence intervals,
which are ranges of values that are likely to contain the true value of the effect. The narrower the confidence
interval, the more precise the estimate is. The wider the confidence interval, the more uncertainty there is
about the estimate.

The confidence intervals for the relative risk of relapse were:

- Fluticasone propionate cream: 0.10 to 0.29

- Fluticasone propionate ointment: 0.31 to 0.83
- Emollient alone: 1 (reference group)

These confidence intervals indicate that the estimate of the treatment effect for fluticasone propionate cream
was more precise than the estimate for fluticasone propionate ointment, as the range of values was narrower
and did not overlap with the reference group. The estimate for fluticasone propionate ointment was less
precise, as the range of values was wider and slightly overlapped with the reference group. However, both
estimates were statistically significant, as the confidence intervals did not include 1, which is the value of no
effect.
What is the measure? What does it mean?
Relative Risk (RR) = risk of the outcome in Relative risk (RR): The ratio of the probability of relapse
the treatment group/risk of the outcome in the in one group to the probability of relapse in another group.
control group. A RR of 1 means that the treatments have the same effect,
a RR of less than 1 means that the treatment reduces the
risk of relapse, and a RR of more than 1 means that the
treatment increases the risk of relapse.
- RR for fluticasone propionate cream = 0.17
- RR for fluticasone propionate ointment = 0.53
- RR for emollient alone = 1 (reference group)
In our example, the RR = 0.10/0.15 = 0.67 This means that the patients who used the cream were 83%
less likely to have a relapse than the patients who did not
use the cream.

Absolute Risk Reduction (ARR) = risk of Absolute risk reduction (ARR): The difference in the
the outcome in the control group – the risk of probability of relapse between two groups. A positive
the outcome in the treatment group. This is ARR means that the treatment reduces the risk of relapse,
also known as the absolute risk difference. and a negative ARR means that the treatment increases the
risk of relapse.
- ARR for fluticasone propionate cream = 1 - 0.17 =
0.83
- ARR for fluticasone propionate ointment = 1 - 0.53 =
0.47
- ARR for emollient alone = 1 - 1 = 0 (reference group)
In our example, the ARR = 0.15 - 0.10 = 0.05 This means that the risk of relapse decreased by 83
or 5% percentage points in the patients who used the cream
compared to the patients who did not use the cream.

Relative Risk Reduction (RRR) = absolute Relative risk reduction (RRR): The proportion of the risk
risk reduction/risk of the outcome in the of relapse that is reduced by the treatment, compared with
control group. An alternative way to calculate the control group. A positive RRR means that the
the RRR is to subtract the RR from 1 (eg. treatment reduces the risk of relapse, and a negative RRR
RRR = 1 - RR) means that the treatment increases the risk of relapse.

3
University of Oxford, 2005
Critical Appraisal for Therapy Articles

- RRR for fluticasone propionate cream = (1 - 0.17) / 1

= 0.83
- RRR for fluticasone propionate ointment = (1 - 0.53) /
1 = 0.47
- RRR for emollient alone = (1 - 1) / 1 = 0 (reference
group)
In our example, the RRR = 0.05/0.15 = 0.33 or This means that the treatment reduced the risk of relapse
33% by 83% of the original risk.
Or RRR = 1 - 0.67 = 0.33 or
33%
Number Needed to Treat (NNT) = inverse of Number needed to treat (NNT): The number of patients
the ARR and is calculated as 1 / ARR. who need to receive the treatment to prevent one relapse,
compared with the control group. A lower NNT means
that the treatment is more effective, and a higher NNT
means that the treatment is less effective.
- NNT for fluticasone propionate cream = 1 / 0.83 = 1.2
- NNT for fluticasone propionate ointment = 1 / 0.47 =
2.1
- NNT for emollient alone = 1 / 0 = undefined
(reference group).
In our example, the NNT = 1/ 0.05 = 20 This means that for every 1.2 people who use the cream,
one relapse is prevented.
2. How precise was the estimate of the treatment effect?
In the article, the estimate of the treatment effect was based on the comparison of the time to relapse of
atopic dermatitis between two groups: one that applied fluticasone propionate twice weekly in addition to
emollient maintenance treatment, and another that applied emollient alone. The authors reported the results
using hazard ratios, which are a type of ratio measure that indicate how much the risk of relapse changes
over time for one group compared to another.

One way to measure the precision of the estimate of the treatment effect is to look at the confidence interval,
which indicates the range of values that are compatible with the observed data. The narrower the confidence
interval, the more precise the estimate is. The article reported that the hazard ratio for fluticasone propionate
cream was 0.17 (95% confidence interval 0.10 to 0.28), and the hazard ratio for fluticasone propionate
ointment was 0.52 (95% confidence interval 0.34 to 0.80). This means that the confidence interval for the
cream was narrower than the confidence interval for the ointment, indicating that the estimate of the
treatment effect for the cream was more precise than the estimate for the ointment.

Another way to measure the precision of the estimate of the treatment effect is to look at the P value, which
indicates the probability of obtaining a result as extreme or more extreme than the observed data, assuming
that there is no true difference between the groups. The smaller the P value, the less likely the result is due
to chance, and the more precise the estimate is. The article reported that the P value for fluticasone
propionate cream was less than 0.001, and the P value for fluticasone propionate ointment was 0.010. This
means that the P value for the cream was smaller than the P value for the ointment, indicating that the
estimate of the treatment effect for the cream was more precise than the estimate for the ointment.

In summary, based on the confidence interval and the P value, the estimate of the treatment effect for
fluticasone propionate cream was more precise than the estimate for fluticasone propionate ointment.
However, both estimates showed a significant reduction in the risk of relapse for the fluticasone propionate
group compared to the emollient alone group.

4
University of Oxford, 2005