Introduction need to adapt their skills, knowledge and professionalism


Quality Commission (CQC, 2018) identifies patient care needs to be consistently
of high quality due to a more complex population, health care practitioners
need to adapt their skills, knowledge and professionalism to adapt to this
changing need. Evidence based research needs to be based on the most current,
valid and relevant evidence and physiotherapists need to implement
evidence-based policies and have a critical attitude to their own practice and
current research, Dawes et al (2005).

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

critical review will be focusing on Dyer et al (2012) “Ambulatory oxygen
improves the effectiveness of pulmonary rehabilitation in selected patients
with chronic obstructive pulmonary disease”, research paper.

essay appraises a quantitative research article which is relevant to Physiotherapy
practice. It focuses on a worldwide chronic condition; Chronic Obstructive
Pulmonary Disease (COPD). Physiotherapy is recommended for patients that have
COPD as it is extremely effective as specified in the NICE guideline pathway,
NICE (2018).

The essay
will follow the Consolidate Standards of Reporting Trials (CONSORT) which is a
tool used to help assess the validity within research. CONSORT (2010) is a
“evidence-based, minimum set of recommendations” for conducting randomised
clinical trials. By following these recommendations, it will help to improve
quality of research and help assist in the decision-making within health care.

Critical Analysis

title that Dyer et al (2012) uses is a declarative title as it states the
outcome of research paper. This type of title is attractive to readers as it
conveys the conclusion of the study. However, this can be misleading as many
studies have limitations, weaknesses and biases. Gjersvik
and Nylenna (2014), provided research showing the increase and upward trend in
declarative titles from years 1974-2014. CONSORT (2010) states that the word
“randomised” should be used in the research title as it can be easily

basic design of the study was stated in the abstract a by Dyer et al (2012). It
clearly reports that this research paper is a focused issue. The authors do not
state the hypothesis in the abstract. The predictions are not clearly written
other than that of the declarative title. CONSORT (2010) state that inaccuracy
information given in the abstract compared to the full body of text can
“mislead someone’s interpretation of the trial findings”. Dyer et al (2012) makes
a mistake within the abstract that a total of 51 patients completed the study,
whereas the results state that only 47 completed the study. This inaccuracy can
cause readers to suggest this is an untrustworthy piece of research.

authors of the study do not provide qualifications or any registration of
governing body. Qualifications and the authors affiliation with a recognised
University or research institution is evidence that the author is knowledgeable
and qualified. Dyer et al (2012) fails to document all authors credentials
therefore making the reader cautious of the results.

conducted by Petticrew et al (2012) comparing non-Cochrane reviews to Cochrane
reviews revealed that Cocchrane reviews are good quality, reliable and trustworthy.
Dyer et al (2012) uses Bradley and O’Neill (2005) Cochrane review to emphasise
that oxygen improves exercise tolerance. However, Bradley and O’Neill (2005)
concluded that the two systemic reviews of a trial consisting of 535
participants. Comparing sample sizes Dyer et al (2012) uses a smaller sample
than the systemic review by Bradley and O’Neill (2005). This recommendation
from Bradley and O’Neil (2005) of increasing the sample size to increase
validity of results should have been achieved prior to conducting the study.

recommended by CONSORT (2010), Dyer et al (2012) does not record any hypothesis
in the introduction section, but provides the reader with the aim of the study.

trial design was based on an ambulatory oxygen assessment. As explained by Main
and Denehy (2016) the incremental shuttle walking test (ISWT) is to be
completed first as the speed used in endurance shuttle walking test (ESWT) is
set by the results in ISWT. The patients were then eligible for the
intervention if they demonstrated improvement in ESWT of at least 10%
ambulatory oxygen. This form of selection prior to the simple randomisation of
being put into two groups is biased, as Dyer et al (2012) only selects patients
that demonstrated improvement with ambulatory oxygen prior to the intervention.

Dyer et al (2012) reports that ethical approval was granted by the Surrey
Research Ethics Committee. This shows that the researcher demonstrated that
they adhere to and accepted the ethical standards.

population is very specific; patients with COPD who attended pulmonary
rehabilitation were selected. Dyer (2012) failed to describe the type of
setting or location where the study was implemented. This is another suggestion
from CONSORT (2010), as the use of different venues could have an impact on the

pulmonary rehabilitation was set by evidence based guidelines, Reis et al
(2007). However, Dyer (2012) states that the programme was tailored to suit
individual needs comprising of strengthening and endurance exercise. This
results in an uneven application of the intervention and has difficulty in
replicating and interpreting study findings (Melnyk and Morrison-Beedy, 2012).
Furthermore, participants randomised to the ambulatory oxygen group were asked
to use their oxygen at their prescribed flow rate for the exercise session. As
documented by the NHS (2017) supplemental oxygen specific to the user must be
prescribed by a doctor. As mentioned previously the credentials of the
researchers was not stated nor any information regarding a medical doctor
prescribing oxygen. This is considered as unethical as, CONSORT (2010)
recommends that for any drug trials it should have details of the prescriptions
and the prescriber.

only quantitative outcome measurement that Dyer et al (2012) uses is the ESWT.
The researcher justifies this by using Sandland (2008) research as it states
that ESWT is more sensitivity to change in exercise tolerance. However, in the
same research article it explains that ESWT detected a greater number of
responders to supplementary oxygen therapy. Therefore, buy choosing only one
outcome measure that shows increase sensitivity and favourability to oxygen
therapy is bias.

other outcome measures are four questionnaires. The only questionnaire
documented and produced quantitative results in the paper was the subjective perceptions
of oxygen. The result of perception of oxygen at the start of the intervention
was that 70% of the participants felt that oxygen would help them to walk further,
emphasising the placebo effect. 

et al (2012) uses sequential sampling and is clearly stated. Sequential
sampling is a non-probability sampling technique, the author most likely chose
this technique as the research picks a single group, Denscombe (2014). The type
of non-probability sampling used is purposive sampling, patients are selected
according to the needs of the study. The advantage of this type of sampling is
it is allowing the researcher to collect a small-scale sample to then relate to
the generalise population. The disadvantage using this type of sampling is that
is open to selection bias and error, but due to the aims of the study Dyer et
al (2012) provides clear reasons why it is necessary to use this type of
sampling technique (Daniel, 2012). To reduce selection bias, Dyer et al (2012)
uses simple randomisation from the purpose sampling to put into two groups.
Alternatively, the researchers could have used restricted randomisation to
produce balance groups and eliminate whether Body Mass Index (BMI) influences
the outcome. Dyer et al (2012) does explain and give evidence that BMI does not
influence the outcomes, however results by Pires et al (2007) comparing walking
distance covered compared to age and BMI report that subjects that have a lower
BMI index walked greater distances. This contradiction would have been
eliminated with restrictive sampling.

power calculation was calculated and was identified the need for a sample size
of 52, however the participants that completed the study was 47, this study
shows to be underpowered. This was due to some participants withdrawing because
of either exacerbation of their condition or other medical conditions, which is
something that cannot be controlled. Therefore, the power analysis was based on
unrealistically high expectation of population effect size, Card (2012).
Research by Sullivan et al (2012) indicates linear relationship between effect
size and sample size. By having low power and biases reduce the reproducibility
of findings and negatively affect the validity of results.

In the
abstract Dyer et al (2012) explains that they conducted a “single-blind,
randomised control study”. This is defined by Chin and Lee (2008) as either the
patients or the researchers do not know which the control group or the
intervention group are. The researcher was blinded to group allocation as the
room air group was given oxygen cylinders and nasal cannulas but the oxygen was
not switched on for the room air group.

are two questionable bias here, firstly whilst conducting the final ESWT the
researcher was blinded by both groups. This is untrue as when oxygen being
delivered it makes a distinguishable noise, Solomon et al (2010). Dyer et al
(2012) fails to explain this therefore, it is not guaranteed that they are
fully blinded by all participants. 
Secondly, as sham oxygen was not available the participants knew which
group they were allocated, this can cause the placebo effect with the
ambulatory oxygen group, this cannot be measured but can have an adverse effect
on results, Dulan (2010). This also creates the observer effect which raises
the prospect that those who know they are in the experimental group will expect
to experience something new (Denscombe, 2014).

et al (2012) explains the reasons why efficacy approach was necessary as it
maximises the likelihood of observing an intervention effect and can be
relatable to clinical practice. However, NICE (2018) state that intention to
treat analysis are used for clinical effectiveness as they mirror actual
practice as not everyone adheres to treatment. This is a limitation to Dyer et
al (2012) research as we cannot relate the intervention to the general

et al (2012) reports that there was a weak negative correlation and high p
value, between the acute response to ambulatory oxygen at baseline and change
in ESWT after pulmonary rehabilitation. 
This means that the intervention group were at an advantage as there
were no statistical evidence that the use of ambulatory oxygen pre-and post ESWT
helped increase distances walked.

flow was achieved by the use of a flow diagram and reasons for participants
withdrawing from the study was given. However, there is unequal completion rate
was identified between groups, an additional person completed the study in the
ambulatory oxygen group, overall increasing the average in distance

the discussion section, CONSORT (2010) recommend that, limitation,
generalisability and interpretation are the minimum that should be included
within this section. The limitations and interpretation were reported in Dyer
et al (2012) discussion section. The researcher was systemic as possible and
was based on a comprehensive search but was shown to been bias as there was no
evidence conflicting his research.

(2010) recommends for randomised control studies should report information
regarding registration, protocol and funding. Dyer et al (2012) reports funding
from the British Lung Foundation but does not provide any information of how
much involvement they had within the study. Registration and protocol were not
documented, and no reason was given. This makes it difficult for health care
professionals to locate the unique trial identification number, in addition
unable to look up the protocol for a replication study.


It is
apparent from this quantitative study that limitations and biases have an
impact on results. The main limitation is the study being underpowered and not
using the intention to treat analysis, therefore results cannot be trustworthy
and unable to relate and generalise to the whole population. Furthermore, the
implication of not having a medical professional who is qualified to prescribe
oxygen has serious ethical ramifications. In conclusion, this study has weak
methodology, limitation and biased results and it would be better to find other
research confirming these findings before using it as evidence for
physiotherapy practice.