Measures were identified by a process that included a literature search and structured review and discussion of the literature. The initial phase consisted of a MEDLINE search of English language articles for 1986-1996 using the Ovid (Ovid Technologies) search engine. An initial set of references was developed by combining the search for respiratory distress with a search for severity measures. Search terms for respiratory distress were developed by the investigators and consisted of: dyspnea, shortness of breath, respiration disorders, asthma, respiratory tract diseases, lung diseases, obstructive lung diseases, emphysema, reactive airways disease, croup, pulmonary edema, congestive heart failure, pneumonia, pulmonary embolism, cyanosis, anoxia, and tachypnea. Search terms for severity included: score, severity of illness index (MESH term), and predictive value of tests (MESH term). This initial set of references was then limited to articles that pertained to human subjects and published in Abridged Indexed Medicus journals.
This created a final reference set of 2,836 references. A title search of this reference list was then performed. The titles were reviewed by all of the investigators in a structured manner, identifying titles that dealt with the development or evaluation of a severity measure. Titles, which focused on a development or evaluation of the measure, were included for further review. Titles using the measures in clinical trials or for evaluations of intervention were not included. A title was included for further review if any single investigator chose it. A unanimous rejection of a title was required for the title to be eliminated from further review.
Four hundred ninety-seven (497) titles were selected for further review. Abstracts of the selected references were obtained and reviewed. The abstracts were reviewed by all the investigators. Again, an abstract required unanimous rejection by the investigators to be excluded from further evaluation. One hundred seventy-five (175) abstracts were selected for further review. For each abstract selected for further evaluation, the full-length article was obtained. Examination of these articles resulted in 75 papers focused on the development or evaluation of a severity or outcome measure.
Articles were then sorted into groups based on the measure they were addressing (for example, dyspnea scales, or measurements of pulmonary function). A single investigator reviewed each group of articles pertaining to a specific measure. The reviews of these articles were conducted in a structured fashion. The areas considered were: time taken to complete, cost and training, scaling, and reliability. Reviews were conducted independently.
After review, the reviewers met to discuss their findings. Each reviewer presented the results of his review and made recommendations regarding the appropriateness of the measure for prehospital outcomes research. After each presentation, a discussion ensued that resulted in a consensus as to whether or not these measures should be recommended. After the measures were selected, another literature search was conducted to ensure all relevant articles regarding these measures were identified. Using the methods described above, titles, abstracts and articles were identified. These articles were then reviewed by Dr. Maio using the structured review described above. A careful review of all article references was also completed, appropriate references identified, and the complete article reviewed.
Results of the review were then discussed with the other reviewers and a consensus reached on measures to recommend. These findings were reviewed with EMSOP consultants at a project meeting. Following suggested modifications, final recommendations were made.