Information for professionals
Explanation of the SORT method
Collection of scientific research
Scientific studies were searched for in Pubmed by a researcher from the Louis Bolk Institute. They were then assessed for quality by two researchers. First, search terms were chosen for the six indications and the 13 therapies (MeSH terms). Initially, meta-analyses, systematic reviews and randomised controlled trials (RCTs) were searched for. If no good studies were found, other types of studies were also searched for, such as observational studies and case studies. These studies were then included in the assessment. Literature was also provided by the professional organisations. The studies found per indication and per therapy were assessed to determine whether they were suitable for inclusion in the assessment based on the title, then based on the abstract, or based on the entire text.
Recommendations according to the SORT method
The included studies were then assessed separately according to the Strength of Recommendation Taxonomy (SORT) method of Ebell et al. 2004. This assessment method is a practice-oriented method, which is appropriate for the working methods of primary care and complementary care. The method attaches great importance to patient-oriented outcomes.
Patient-centered outcomes are outcomes that are important to the patient, examples include:
- Experienced improvement of complaints
- Experienced quality of life
The assessed studies according to the SORT method then give recommendations for a therapy for an indication to a certain extent. The first step is to assess the quality of the studies found, see table 1.
The second step is to come to a recommendation based on the quality of the studies found. The latter is indicated with an A, B or C, see table 2.
According to the SORT method, all recommendations are substantiated on:
- Quality of the individual studies (Table 1),
- Minimum number of good quality studies,
- Consistency of evidence.
Table 1.
Classification of assessment of quality of found studies
| Level | Assessment of quality per study | Study design |
|---|---|---|
| 1 | Good quality patient-centered evidence | Systematic review/ meta-analysis en 2 RCTs with consistent outcomes* |
| 2 | Limited quality patient-centered evidence | Systematic review/ meta-analysis with inconsistent outcomes, 1 RCT, low quality clinical trial, cohort, case-control* |
| 3 | Other evidence | Case series, case studies, benchmark studies, opinions* |
* For more information on the quality assessment of the individual studies, see Figure 3 from Ebell et al. 2004.
If a study does not report patient-oriented outcomes, it is classified as Level 3. For meta-analyses, the quality of the individual studies included is assessed. If the quality of these studies is high, the evidence is rated as Level 1. If the quality is judged to be low, it is rated as Level 2. Individual RCTs were assessed for quality by a scientific researcher from the Louis Bolk Institute.
Table 2.
Classification of recommendations by level of evidence
| Level | Recommendation by level of evidence |
|---|---|
| A | Recommend based on consistent and good quality patient-centered evidence |
| B | Recommend based on inconsistent or limited quality patient-centered evidence |
| C | Recommend based on consensus, normal practice, opinion, disease-specific evidence |
| N | Do not recommend based on no research or evidence found |
For the assessment of the recommendations, the ‘patient oriented outcomes’ are checked again. If they are not there, it becomes a recommendation C. The most difficult decision is between A or B. If there is at least 1 RCT, a systematic review or a meta-analysis with ‘patient oriented outcomes’, it automatically becomes an A or B. It becomes an A if it is a Cochrane review with a recommendation, or two good quality RCTs (or cohort studies) with consistent findings, or one meta-analysis or systematic review with consistent findings.
Consistency across studies
Results between studies can be consistent or inconsistent. A result is consistent if:
- Multiple studies show the same results,
- When possible differences between studies can be explained well,
- Good quality systematic reviews and/or meta-analyses exist.
A result is inconsistent if:
- The results between studies differ and cannot be explained,
- Good quality systematic reviews and/or meta-analyses exist but with varying results.
No research or evidence found
During the search for evidence by the Louis Bolk Institute for the six indications for various complementary therapies, studies were also found that showed that there was no effect. Due to the evidence of no effect found, the recommendation ‘N’ for ‘Do not recommend’ was added.
Performed SORT analyses
By clicking on the indication below you can read the recommendations according to the SORT method for various forms of therapy.
Disclaimer
All information distributed by RBCZ through this Complementary Care Decision Aid has been compiled with care. No rights can be derived from the content of this Decision Aid in any way. RBCZ accepts no liability for direct or indirect damage arising from or related to incorrect or incomplete information, use of this website, malfunctions, or otherwise. RBCZ is not responsible for medical decisions made based on the information in this Decision Aid.
