METHODS
The literature on the respiratory complications of TEF is based on
small, single institution, retrospective case series and case reports.
There are no systematic reviews or randomized controlled trials to
formulate evidence-based guidelines on diagnosis and treatment. To
address this deficiency the RCWG chose a modification of the RAND
Appropriateness Method (RAM)21 developed by the RAND
Corporation–University of California Los Angeles, that helps synthesize
the available empirical evidence using the collective experience of a
panel of experts. The RAM approach has been widely used in the
development of clinical practice guidelines, including in adult
respiratory care22, in primary pediatric
care23 and in rare pediatric conditions such as
Duchenne Muscular Dystrophy24–26.
The RAM rates diagnostic and therapeutic interventions on the basis of
their “appropriateness” and “necessity”. The appropriateness of an
intervention is based on its known or expected medical benefit without
consideration of any logistical or financial obstacles. Appropriateness
does not automatically imply necessity (e.g. bronchoalveolar lavage is
an appropriate diagnostic modality for the evaluation of pneumonia, but
it is not necessary for every case of pneumonia). The “necessity” is
based on whether a) the specific intervention has been deemed
appropriate; b) there is reasonable expectation it will benefit the
patient in a substantial way; c) according to prevailing standards of
care, it would be inappropriate not to offer this intervention to the
patient. The RAM enables individual expert opinion to be independently
and anonymously expressed and identifies areas of agreement as well as
of disagreement or uncertainty that are in need of further study.
The development of the recommendations consisted of the following steps
(Figure 1):
Step 1 . Review of the literature andformulation of statements . Questions relating to the domains ofRespiratory Physiology (RP), Diagnostic Methods (DM) andTherapeutic Interventions (TI) for the management of respiratory
complications in EA-TEF were generated by literature reviews limited to
the English language. Each question was subsequently discussed and
reformulated into specific statements during a face-to-face meeting. The
agreed upon statements were entered into matrices containing clinical
scenarios with a list of potential interventions in 3 areas: Respiratory
Pathophysiology (RP) Diagnostic Methods (DM) and Therapeutic
Interventions (TI).
Step 2. Rating of statements for “appropriateness” . All
statements were individually rated for appropriateness on an ordinal
scale of 1-9 as follows: INAPPROPRIATE: ratings 1-3; UNCERTAIN: ratings
4-6; and APPROPRIATE: ratings 7-9. Pathophysiologic mechanisms were
rated as “appropriate” according to the degree to which the statement
was supported by generally acknowledged pulmonary physiology, currently
available literature, and/or by the clinical experience of the
responders. Statements in which all responses were in agreement i.e.
within the same category were set aside whilst the remaining statements
underwent a second and third round of rating. The results of each round
were shared within the RCWG without identifying the responders.
Step 3. Rating of statements for “necessity” .
Necessity ratings were applied only to statements pertaining to
Diagnostic Methods and Therapeutic Interventions and were rated on an
ordinal scale 1-9 as follows: UNNECESSARY: 1-3; UNCERTAIN: 4-6;
NECESSARY: 7-9. The median, range and the mean (±SD) were calculated. In
the final collation, the responses were classified as follows:
- “Strong” agreement : both the median and the mean values of
the responses for a given item were within the highest category (7-9)
- “Moderate” agreement : both the median and the mean values
of the responses for a given item were within the middle category
(4-6)
- “Weak” agreement : The median value was in the lowest
category (1-3) and/or the difference between the highest and lowest
value was ≥ 3.
Step 4 . Formulation of the recommendations. The
recommendations were classified as Level A (based on strong agreement),
B (moderate agreement) and C (weak agreement). The ratings on Necessity
were formulated into specific recommendations presented in 10 sections
(6 for Diagnostic Methods and 4 for Therapeutic Interventions) with
their supportive evidence.