In order to establish criteria for the appropriate referral of patients with Parkinson's disease for the consideration of DBS, we used the RAND Appropriateness Method (RAM).1, 2 This method is particularly helpful when scientific evidence is lacking or is not detailed enough to establish clear indications for a particular treatment or diagnostic procedure. Since its development in the mid-eighties by the American RAND Corporation and the University of California, the method has been used in various fields and has been extensively tested for its internal consistency and external validity.3, 4

The RAM consists of a modified Delphi approach in which a panel of experts assesses the appropriateness of particular clinical decisions in an iterative way. Initially, a literature study is done to critically appraise and summarize the evidence from clinical studies. Using the results of this overview and the additional comments of experts, particular patient characteristics are selected that may be relevant for the clinical decision under investigation. By combining these characteristics, a set of hypothetical profiles (cases) is generated, and a panel of experts is then asked to rate the appropriateness of certain clinical decisions for each of these cases. A decision is called 'appropriate' when the expected benefits (e.g. symptom reduction) exceed the expected risks (e.g. adverse events). The extent of appropriateness is expressed using a 9-point scale in which 9 = extremely appropriate, 1 = extremely inappropriate, and 5 = equivocal or uncertain. After the panelists have individually rated all indications, they meet to discuss the results. The aim of this discussion is not primarily to reach consensus, but to investigate whether disagreement is due to differences in the interpretation of the definitions or cases that are used. After this discussion, a second individual rating round takes place, in which all or part of the indications are rated a second time. Based on the median score and the extent of agreement, for each of the patient profiles a panel statement (appropriate, uncertain, and inappropriate) is calculated. The expert panel has agreed during the kick off meeting, held in Geneva, Switzerland, on 20th August 2013 that, in the tool, “appropriate” will be used as “strongly recommended, “uncertain” as “recommended” and “inappropriate” as “not recommended” categories in order to provide clearer directions to clinicians. The results of RAM studies may be used retrospectively (e.g. for audits), or prospectively (e.g. for guideline development).

1Brook RH, Chassin MR, Fink A, et al. Int J Technol Assess Health Care 1986;2:53-63.

2Fitch K, Bernstein SJ, Aguilar MD, et al. The RAND/UCLA Appropriateness Method. User's manual.

3Shekelle P. Med Decis Making 2004; 24: 228-31

4Lawson EH, Gibbons MM, Ko CY, Shekelle PG. J Clin Epidemiol 2012;65:1133-43.