METHODS
In response to the mandate from the SAAQ, the Task Force focused on risk factors, diagnosis, treatment and prognosis of soft-tissue injury to the neck from whiplash. In terms of the Quebec Classification of Whiplash-Associated Disorders (WAD), Grade 0 WAD was not considered because no disorder is manifest. Grade IV WAD, involving bony-tissue or spinal-cord injury, was not considered except for studies evaluating diagnostic procedures for differentiating between bony-tissue and soft-tissue injuries. Manifestations of soft-tissue injuries of the neck from other causes, such as "shaken baby syndrome" and diving mishaps were also excluded. While this definition limited the generalizability of the Task Force recommendations, it allowed a focussed review of the literature on a clinical entity homogeneous in cause and directly relevant to motor-vehicle collisions (MVC). Unfortunately, very little acceptable material was found.
Studies Reviewed
In the first stage of identifying material for review, a wide net was cast to optimize our ability to find relevant articles. A bibliography of thousands of titles was amassed from the sources described below.
1. A broad search of computerized bibliographic databases (Medline, TRIS, NTIS) was conducted for articles on head, neck, spinal cord problems and specific problems such as cervicobrachialgia, with or without mention of traffic collisions. This was updated through September 1993.
2. Task Force members suggested investigators to contact and supplied unpublished and published reports of studies of which they were aware.
3. Agencies such as the Insurance Institute for Highway Safety, and others were contacted and provided information on studies that they had conducted or supported.
4. Reference lists of review and original articles included in the Task Force Review were scanned for relevant articles.
References resulting from the above search strategy were subjected to a two-stage screening process by the Task Force Secretariat. The purpose of this screening was to reduce the pool of articles, selecting for review those articles dealing with the content areas identified by the Task Force mandate.
The inclusion criteria for the pool of potentially relevant articles (preliminary screening) from the sources described above were:
1. an article or report published beginning in 1980 and appearing in Medline or found by other means through September 1993 (important articles or "seminal papers" published earlier were included as exceptions);
2. an English or French language report;
3. an original research report, brief report, work in progress, or letter;
4. an article published in a journal, conference proceedings, or technical report, or an unpublished manuscript;
5. the study subjects included subjects with acceleration-deceleration injury to the neck from MVCs;
6. the study examined: a) risk factors; b) prognosis; c) diagnostic methods; d) efficacy of treatments; e) clinical manifestations.
The following exclusion criteria were used at the preliminary screening.
1. The study subjects did not include subjects with neck injuries due to MVC. Such studies were occasionally not excluded, as exceptions, when the substantive area clearly included clinical manifestations similar to those arising in WAD.
2. The study subjects were restricted to those with bony tissue injury of the neck, spinal cord injury, head trauma, penetrating wounds, infections, multiple injuries, or with other comorbidities.
3. The study subjects were animal, cadaver, or simulation studies. A few cadaver and simulation studies were considered as background reading. However, they were not candidate studies for full Task Force review.
Those articles that passed the preliminary screening were subjected to a second screening to select articles for in-depth review. The purpose of this screening was to identify articles whose substantive area was within, or informative, for the focus of the Task Force. This involved confirmation that the article met the inclusion and exclusion criteria of the preliminary screening and invoked further the following exclusion criteria:
1. review articles or clinical reviews (These were excluded from the best evidence synthesis but were, as "seminal" papers cited in a number of articles reviewed, distributed as background reading),
2. single case reports, and
3. laboratory and simulation studies.
After the closing date for the search strategies, we continued to monitor the literature and to receive articles from Task Force members, with the goal of identifying recently published highly meritorious articles for the Task Force review. Current Contents was scanned in October 1993, and in February 1994 an additional Medline search was conducted for specific treatments, because of concern regarding the dearth of articles on specific treatments for WAD. Articles identified by these means through April 1994 were included in the Task Force review process described below.
A final search of Medline using the broad search strategy was conducted in July 1994. Articles identified by these searches or supplied by Task Force members through September 1994, if deemed highly relevant on preliminary screening, were reviewed by the Secretariat or by individual Task Force members. The objective of this review was to identify any highly relevant and meritorious articles that would enhance the knowledge base for the Task Force Recommendations, even though they were found too late for the best evidence synthesis. Such late-find articles are footnoted in the section on results, later in this section.
Critical Review Process
The articles identified by the two-stage screening process described above were subjected to critical appraisal by the Task Force. Before undertaking the critical review of the literature, the principles of research design, hypothesis testing and causal inference were reviewed by the Task Force. These principles formed the basis for assessment of the robustness of evidence from individual studies and for synthesizing evidence from the studies reviewed. To assure a common understanding of the relevant substantive issues (e.g., anatomy, pathophysiology), a clinicians' subgroup, with representation from several subspecialties, prepared and presented didactic sessions on the basic clinical sciences, prior to the critical review of the literature.
The critical appraisal of the articles was guided by using structured Critical Appraisal Forms adapted and augmented from forms used by the Canadian Task Force on the Periodic Health Examination, the Quebec Task Force on Spinal Disorders, and the Working Group on Passive Smoking. These forms prompt the reviewer to record information regarding the study design, methodologic features, analysis, clinical relevance, scientific merit, strengths and weaknesses of the study. Each selected article was reviewed independently by two Task Force members. The articles were then presented by their reviewers and discussed by the full Task Force. A consensus review was developed as a result of the discussion. When consensus as to relevance, quality of evidence, or key results could not be reached in discussion, a third reviewer did a critical appraisal. The article was returned for discussion at a later date.
Each article was classified as to its substantive relevance and scientific acceptability. Because the review of the literature took place over a period of two years, there was an evolution in the Task Force definition of clinical relevance and acceptability. At the time that the synthesis of acceptable evidence took place (i.e., during the drafting of the report), any Task Force member could return an article for reconsideration if there was dissatisfied with its classification. Thus, an early consensus decision could be affirmed or changed, based on the matured consensus of the Task Force.
The references identified from all search strategies described above (first screening) were compiled and stored using a computerized bibliographic database management system. The critical appraisal and consensus information on articles reviewed were stored in a proprietary database. The database structure, data entry interface, error checking and other features were developed for the Task Force with commercially available database management software.
