Diagnosis
Collins S.P., et alPrevalence of Negative Chest Radiography Results in the Emergency Department Patient with Decompensated Heart Failure
Annals of Emergency Medicine
Jan 2006; 13-18
- Submitted by:Alan Van Opstal - Medical Student
- Institution:Grand Rapids MERC/Michigan State University Program in Emergency Medicine, Spectrum Health–Butterworth Hospital, Grand Rapids, MI
- Date submitted:24th August 2007
1
Objectives and hypotheses
1.1
Are the objectives of the study clearly stated?
Yes. That author states, "This secondary analysis of a case registry examined the sensitivity of the emergency department (ED) chest radiograph and the sensitivity of the ED final diagnosis in a cohort of admitted patients who either had a primary hospital discharge diagnosis of acute decompensated congestive heart failure or were treated for this condition during their hospital stay."
2
Design
2.1
Is the study design suitable for the objectives
The design is appropriate, although the exclusions and lack of a gold standard described below limit the authors claims.
2.2
Who / what was studied?
The author states, "This secondary analysis of a case registry examined the sensitivity of the emergency department (ED) chest radiograph and the sensitivity of the ED final diagnosis in a cohort of admitted patients who either had a primary hospital discharge diagnosis of acute decompensated congestive heart failure or were treated for this condition during their hospital stay."
2.3
Was this the right sample to answer the objectives?
Not entirely. Only subjects whose care was initiated in the ED, and only those who were admitted from the ED were included. This does not fully represent the entire spectrum of patients who present with decompensated heart failure.
2.4
Is the study large enough to achieve its objectives? Have sample size estimates been performed?
"There were 85,376 patients meeting inclusion criteria for this analysis." No formal discussion was given in regards to the needed study size for a desired power for the study.
2.5
Were all subjects accounted for?
"There were 121,658 patients enrolled; 26,962 were excluded because they did not originate in the ED, and 9,320 did not have a yes/no answer for radiographic signs of congestion."
2.6
Were all appropriate outcomes considered?
Outcomes were not considered in this study, as the authors set out to determine the sensitivity of chest radiography.
2.7
Has ethical approval been obtained if appropriate?
Yes. "The institutional review board at each center approved the registry protocol, and all patient data are kept confidential through encrypted treatment."
2.8
Was an independent blinded gold standard test applied to all subjects?
No. "No uniform or standardized criteria were used to define radiographic findings of congestion, and radiographs were not evaluated by a core group of radiologists blinded to patient information."
3
Measurement and observation
3.1
Is it clear what was measured, how it was measured and what the outcomes were?
Yes. Interpretations of chest radiographs (Y/N for pulmonary congestion) by radiologists and primary discharge diagnoses of patients were compared to determine the sensitivity of CXR for correctly identifying decompensated heart failure.
3.2
Are the measurements valid?
To a certain degree, as chest radiographic findings are not specific to heart failure.
3.3
Are the measurements reliable?
This is questionable as "no uniform or standardized criteria were used to define radiographic findings of congestion, and radiographs were not evaluated by a core group of radiologists blinded to patient information." And, "The treating physicians did not follow standardized criteria when determining the cause of each patient's dyspnea, which reflects actual practice in the majority of hospitals. Not following standardized criteria could lead to under- and overdiagnosis of decompensated heart failure because diagnostic criteria would not be consistent from one physician to the next."
3.4
Are the measurements reproducible?
See above.
4
Presentation of results
4.1
Are the basic data adequately described?
Yes. Patient demographic comparison done in Table 1. Patients with heart failure and non-heart failure diagnoses compared in regards to radiographic findings done in Tables 2 and 3.
4.2
Are the results presented clearly, objectively and in sufficient detail to enable readers to make their own judgement?
Yes. Results section walks through the findings along with 95% confidence intervals.
4.3
Are the results internally consistent, i.e. do the numbers add up properly?
Yes.
5
Analysis
5.1
Are the data suitable for analysis?
Yes. The diagnostic test of chest radiography is stratified into Positive and Negative in regards to heart failure findings and compared with discharge diagnoses, thus determining the sensitivity of using CXR.
5.2
Are the methods appropriate to the data?
Yes. "Data are described using means and SDs for age and frequencies and proportions for other variables. The 95% confidence intervals (CIs) for proportions have been computed using the score method with continuity corrections."
5.3
Are any statistics correctly performed and interpreted?
Yes. The authors also feared that "inclusion of the subgroup of patients identified as having heart failure by 'a primary focus of treatment was heart failure' may have introduced selection bias to the results. A sensitivity analysis excluding this patient group was conducted, and the prevalence of radiographic signs of congestion in the cohort of patients identified by a "primary ICD-9 code indicating a discharge diagnosis of acute decompensated heart failure" is reported in addition to the primary results."
6
Discussion
6.1
Are the results discussed in relation to existing knowledge on the subject and study objectives?
There is mention that "our findings are consistent with previous studies, which found chest radiography unreliable in ED patients who had signs and symptoms of decompensated heart failure. Twenty percent of cardiomegaly observed on echocardiography is missed on chest radiography, and pulmonary congestion can be minimal or absent in patients with significantly elevated pulmonary artery wedge pressures."
6.2
Is the discussion biased?
The authors mention possible sources of selection, spectrum, and incorporation bias.
7
Interpretation
7.1
Are the authors' conclusions justified by the data?
Yes. The data suggests that chest radiography has a sensitivity of 81%, therefore, "clinicians should consider the prevalence of chest radiography that lacks signs of congestion when evaluating patients for possible acute decompensated heart failure."
7.2
What level of evidence has this paper presented? (using CEBM levels )
3b due to the lack of consistently applied reference standards.
7.3
Does this paper help me answer my problem?
Not really. It tells me that "despite a nearly 20% negative rate in patients ultimately found to have acute decompensated heart failure, chest radiography may be one of the more sensitive, readily available tests available to the emergency physician."
8
Implementation
8.1
Can the test be implemented in practice?
Yes, readily in the ED with patients presenting with possible decompensated heart failure.
8.2
What aids to implementation exist?
Readily available radiography in the ED.
8.3
What barriers to implementation exist?
Few since radiography is relatively cheap and quick to obtain.
8.4
Are my patients the same as the patients tested?
Yes, although one must consider also the patients with decompensated heart failure who are discharged from the ED who were not included in this study.
8.5
Will the test improve diagnosis in my patients?
Yes, but "when used together, multiple diagnostic tests can improve the diagnostic ability of the emergency physician." Therefore the use of chest radiography alone may not, but in conjunction with BNP and physical findings, it may.