Prediction of peak flow values followed by
feedback improves perception of lung
function and adherence to inhaled
corticosteroids in children with asthma
SCH Journal Club
26th March 2013
Aim
To determine whether predicting peak flow
value and comparing with actual peak flow
value in asthmatic children will improve
perception of their lung function
Objectives
• Search literature for relevant paper
• Critically appraise the paper using CASP framework
• Determine validity of paper
• Assess whether able to apply to clinical practice
Clinical Scenario
• 10 year old boy known to have poor control of
his asthma attends the respiratory clinic for
routine review.
• He is noted to have had several admissions to
hospital since his last review despite maximal
treatment, always attending with either acute
severe or life-threatening symptoms.
Background
• Many factors can contribute to the poor control of
asthma
• One such factor which has been shown to contribute is
the discrepancy between objective measures of airway
obstruction and symptom perception.
• This can result in a delay in seeking treatment for acute
exacerbations due to an under-perception of the degree
of bronchoconstriction
Question Arising
• Does our patient’s perception of his symptoms correlate
with the level of his lung function and is he able to
recognise when his lung function has become
compromised?
• If not is there a proven way to improve his perception of
lung function thereby improving the control of his asthma
and reducing his acute admissions to hospital?
Clinical Question
Population
In children with poorlycontrolled asthma
Intervention
is there a simple intervention to
Comparison
Outcome
improve perception of lung
function and thereby improve
control of symptoms?
Literature Search
[asthma]
AND [perception OR awareness]
AND [(lung AND function) OR (airway AND
obstruction)]
AND [(hospital OR (secondary AND care) OR
(emergency AND healthcare)].
Limited to [child 0 – 18 years AND English
language]
Prediction of peak flow values followed by feedback
improves perception of lung function and adherence
to inhaled corticosteroids in children with asthma
Jonathan M Feldman, Haley Kutner, Lynne
matte, Michelle Lupkin, Dara Steinberg,
Kimberly Sidora-Arcoleo, Denise Serebrisky,
Karen Warman
Thorax 2012;67:1040 - 1045
Method
• Patients recruited from asthma clinics, primary
care clinics, ED and mailings
• Restricted to Puerto Rican, African-American
and Afro-Caribbean ethnic groups
• Confirmed diagnosis of asthma with breathing
difficulties in past year
• Exclusion criteria included cognitive learning
disability, vocal cord dysfunction, inability to
conduct spirometry
1st Study Group
FEEDBACK GROUP
2nd Study Group
NON-FEEDBACK GROUP
• Initial training on use of spirometer to
ensure correct technique
• Initial training on use of spirometer to
ensure correct technique
•Over a six week period recruits
assessed their lung function twice a day
at set times
•Over a six week period recruits assessed
their lung function twice a day at set times
•Each recruit predicted their PEF and
recorded this on their spirometer
• They then performed three consecutive
PEF and documented the “best of 3” PEF
• Each recruit predicted their PEF and
recorded this on their spirometer
• They then performed three consecutive
PEF and documented the “best of 3” PEF
• Blinded to value of PEF
• Results of PEF value were fed-back to
recruits to be able to compare with their
predictions
• However they were unable to alter
predicted PEF
Figure 1 Asthma risk grid (adapted
with permission from OceanSide
Publications, Inc from Klein et al18).
Accurate zone: boxes 1, 5, 9 and
±10% wedge; magnification zone:
boxes 2, 3 and 6; danger zone: boxes
4, 7 and 8. PEF, peak expiratory flow.
Outcome Measures
1. Asthma Risk Grid (Figure 1) was used to
determine the % of predictions from each recruit
in (i) accurate, (ii) under-perception and (iii)
over-perception zones as a measure of the
perception of respiratory compromise
2. Adherence to ICS inhalers during study period
was monitored using doser-devices (95/192; 10)
3. Use of quick-relief metered dose inhaler was
monitored during study period using doserdevices (181/192; -16)
Figure 2 Participant enrolment in the
peak expiratory flow (PEF) feedback
and no feedback conditions
A/ Are the results of the study valid? –
Screening Questions
1. Did the study address a
clearly focused question?
Yes
■
Can’t Tell
□
No
□
“We hypothesised that children
who receive feedback on PEF
predictions would have less
• the population studied
under-perception of respiratory
• the risk factors studied
compromise, better ICS
• the outcomes considered
adherence and less quick relief
• is it clear whether the study tried to medication use than children who
detect a beneficial or harmful effect? do not receive feedback”
HINT: A question can be focused in
terms of?
Page 1041 end of first paragraph
A/ Are the results of the study valid? Screening Questions
Can’t Tell
□
No
□
2. Did the author use an
appropriate method to answer
their question?
Yes
■
• is a cohort study a good way of
answering the question under the
circumstances?
• why not use a randomised control
trial?
- but two groups similar in baseline
characteristics (table 1)
- difficult to blind to intervention
•Did it address the study question?
• Yes – the study determined
perception of lung function in each
group, comparing the accuracy of
predictions with actual readings of lung
function and then compared the
feedback group with the non-feedback
group; also looked at inhaler use
A/ Are the results of the study valid Detailed Questions
3. Was the cohort recruited in an Yes
Can’t Tell
acceptable way?
□
■
No
□
• was the cohort representative of a
defined population
• Yes – defined ethnic population with
confirmed diagnosis of asthma
• was there something special about the
cohort?
• No – baseline characteristics similar
across feedback and non-feedback
group (Table 1)
• was everybody included who should
have been included?
• Not sure – used convenience sampling
(this is noted as a limitation of the study
in the discussion)
• Limited inclusion criteria specified –
e.g. age groups not mentioned
•No information on any individuals who
declined to participate
4 Was the exposure accurately
measured to minimize bias?
Yes
■
Can’t Tell
□
No
□
The exposure was the process of
predicting and recording actual PEF
measurements with feedback on results
for the feedback group and blinding of
results with non-feedback group.
• Did they use subjective or objective
measurements?
 subjective measurement was the
recruits’ prediction of lung function
 objective measurement was the
recorded PEF
* Potential for bias in blinded group through
less motivation resulting in less effort with
PEF performance but PEF results
comparable between two groups
• Do the measures truly reflect what you
want them to (have they been validated)?
• This method for perception of pulmonary
function has been used in previous studies
• Were all the subjects classified into
exposure groups using the same
• Yes – procedure the same for each recruit
over the study periods
Can’t Tell
□
No
□
5. Was the outcome accurately
measured to minimize bias?
Yes
■
HINT: We are looking for measurement
or classification bias:
• Did they use subjective or objective
measurements?
• Do the measures truly reflect what you
want them to (have they been
validated)?
• Has a reliable system been
established for detecting all the cases
(for measuring disease occurrence)?
• Were the measurement methods
similar in the different groups?
• Were the subjects and/or the outcome
assessor blinded to exposure (does this
matter)?
The outcome is the percentage of PEF
measurements falling into each
category on the Asthma Risk Grid to
determine whether those with feedback
are more accurate at predicting lung
function and the use of ICS and quick
relief inhalers.
 The measurements were the same in
the two study groups
 There was no blinding to exposure –
impossible to do with subjects as this
was the exposure being measured.
6 A Have the authors identified all
important confounding factors?
Yes
□
Can’t Tell
□
List the ones you think might be important,
that the author missed.
See Table 2, page 1043
No
■
DID NOT consider:
 Duration of diagnosis of asthma
 Asthma within primary care-giver
? Environmental factors
B. Have they taken account of the
confounding factors in the design
and/or analysis?
HINT:
• Look for restriction in design, and
techniques eg modelling, stratified-,
regression-, or sensitivity
analysis to correct, control or adjust for
confounding
Yes
■
Can’t Tell
□
No
□
Used analysis of covariance for analysis of
data and included age and ethnicity as
covariates
Repeated analysis excluding higher age
group from feedback group and got similar
results
7. A. Was the follow up of subjects Yes
complete enough?
□
Can’t Tell
■
No
□
7. B. Was the follow up of subjects Yes
long enough?
□
Can’t Tell
■
No
□
HINT:
• The good or bad effects should have
had long enough to reveal themselves
• The persons that are lost to follow-up
may have different outcomes than those
available for assessment
• In an open or dynamic cohort, was
there anything special about the
outcome of the people leaving, or the
exposure of the people entering the
cohort?
• Effect with predicting PEF seen after
initial training and maintained for study
period of six weeks
• BUT was effect continued? – ICS take
6 weeks to take effect so may have been
beneficial to have longer study period to
see real improvement in control of
asthma (and therefore reduction in use
of reliever medication)
• Those lost to follow-up had similar
demographic details
B/ What are the results?
8 What are the results of this
study?
Refer to Table 3 on page 1043
• What are the bottom line results?
• 60.7% +/- 2.3 of predictions in the PEF
feedback group fell within the accurate zone
compared to 48.2% +/- 2.5 of predictions in
no-feedback group (p <0.001)
• Have they reported the rate or the
proportion between the
exposed/unexposed, the ratio/the rate
difference?
• How strong is the association between
exposure and outcome (RR,)?
• 15.3% +/- 2.1 of predictions in the PEF
feedback group fell within the danger zone
compared to 41.6% +/- 2.3 of predictions in
the no-feedback group (p <0.001)
• What is the absolute risk reduction
(ARR)?
• Increased adherence to ICS for duration of
trial in feedback group (48.8% vs 27.5%)
• Increased use of quick-reliever inhalers for
duration of trial in feedback group (41.9 puffs
vs 21.8 puffs)
9. How precise are the results?
• Size of the confidence intervals
? One standard error of the mean used (=
68% sample mean within 1 standard error;
i.e. 68% of the means will fall within
percentages given) therefore results not
that precise
10. Do you believe the results?
Yes
□
• Big effect is hard to ignore!
• Can it be due to bias, chance or
confounding?
• Are the design and methods of this study
sufficiently flawed to make the results
unreliable?
• Consider Bradford Hills criteria (e.g. time
sequence, dose-response gradient,
biological plausibility, consistency).
• Don’t fully understand the statistics
Can’t Tell
■
No
□
• There does appear to be an improvement
in the perception of lung function within the
feedback group which I think is due to the
use of feedback
• Use of inhalers less clear as smaller
numbers
C/ Will the results help me locally?
11. Can the results be applied to Yes
the local population?
□
Can’t Tell
■
No
□
HINT: Consider whether
• The subjects covered in the study
 very different ethnic group
could be sufficiently different from your compared to local population but it is
population to cause concern
possible that the results would be
similar in different ethnic groups
• Your local setting is likely to differ
(previous studies have had similar
much from that of the study
results in a different ethnic group)
• Can you quantify the local benefits
and harms?
12. Do the results of this study
Yes
fit with other available evidence? ■
Can’t Tell
□
No
□
Previous studies have used a similar
protocol with Latino children and found
similar results
(end of first paragraph in Discussion,
page 1044)
Previous studies have also shown the
lack of incremental learning effect
across time
(second paragraph in Discussion, page
1044)
Summary and Conclusions
• The study supports the suggestion that
predicting peak flow values followed by
feedback does improve the perception of lung
function in children with asthma from a specified
ethnic group in USA.
• However there are flaws in the methodology and
results are not clear compromising validity of
paper and results
• Not sure that it could be applied to local
population - ideally need to look at local
population to determine similar effect.
Bottom Line
• An interesting concept which has potential to improve
perception of lung function but needs further research to
determine whether applicable to local population and
whether there are benefits on the control of symptoms.