An AAOS Work Group, consisting of eight physician members, was assembled specifically for the development of this guideline. The Work Group consisted of a diverse group of physician specialists with expertise in treating patients with carpal tunnel syndrome.
The Work Group, with the assistance of the AAOS staff, began by formulating “simulated recommendations”. The simulated recommendations were used to define the scope of the guideline and to refine the literature searches that were conducted. The Work Group, with the assistance of the AAOS medical librarian and staff, completed a systematic review of the relevant literature. Details of the systematic review are provided below.
During the process of developing this guideline, the Work Group participated in a series of conference calls and meetings. When published information of sufficient quality was not available, consensus opinion was employed.
The final draft of the guideline was reviewed by an outside advisory panel (peer review), reviewed internally by the AAOS Board of Directors, Council on Research Quality Assessment and Technology, Board of Councilors, and Board of Specialty Societies (public commentary) and approved by the AAOS Evidence Based Practice Committee, Guideline Oversight and Technology Committee, Council on Research Quality Assessment and Technology, and the Board of Directors.
Peer review of the draft guideline is completed by an outside Peer Review Advisory Panel. Outside Advisory Panels are convened for each AAOS guideline and consist of experts in the guideline’s topic area. These experts represent professional societies other than AAOS and are nominated by the guideline Work Group prior to beginning work on the guideline. Non-editorial comments received from each reviewer are documented, reviewed by the Work Group and approved by the Work Group Chairperson. AAOS staff sends each reviewer the approved documentation for his/her comments. For this guideline, thirteen outside peer review organizations were invited to review the draft guideline and all supporting documentation. Seven societies participated in the review of the CTS Treatment guideline draft.
Following response to all reviews, the guideline draft was sent to thirty-one individuals, who were members of the AAOS Board of Directors, Council on Research Quality Assessment and Technology, Board of Councilors, and Board of Specialty Societies for public commentary. Following this period of public commentary, the guideline was submitted for approval.
Within AAOS, multiple iterations of written review were conducted by the participating Work Group, AAOS Guidelines and Technology Oversight Committee, AAOS Evidence Based Practice Committee, and the AAOS Council on Research, Quality Assessment and Technology prior to final approval by the AAOS Board of Directors. The total number of AAOS reviewers within these governing bodies is fifty-eight. The approval process is documented in Appendix VI.
Voting on guideline recommendations will be conducted using a modification of the nominal group technique (NGT), a method previously used in guideline development. Briefly, each member of the guideline Work Group ranked his or her agreement with a guideline recommendation on a scale ranging from 1 to 9 (where 1 is “total disagreement” and 9 is “total agreement”). Consensus is obtained if the number of individuals who do not rate a measure as 7, 8, or 9 is statistically non-significant (as determined using the binomial distribution). Because the number of Work Group members who are allowed to dissent with the recommendation depends on statistical significance, the number of permissible dissenters varies with the size of the work group. The number of permissible dissenters for several work group sizes is given in the table (See PDF, pg. 5).
The NGT is conducted by first having members vote on a given recommendation without discussion. If the number of dissenters is “permissible”, the recommendation is adopted without further discussion. If the number of dissenters is not permissible, there is further discussion to see whether the disagreement(s) can be resolved. Three rounds of voting are held to attempt to resolve disagreements. If disagreements are not resolved after three voting rounds, no recommendation is adopted.
Article Inclusion and Exclusion Criteria
Inclusion and exclusion criteria were developed a priori. Articles were retrieved and included only if they met these specific inclusion and exclusion criteria (see Appendix II: Article Inclusions and Exclusions). Supplemental searches were conducted to identify national rates and other information relevant to performance measures.
Work Group members were given the opportunity to supplement the searches of electronic databases with articles not identified by those searches. No additional articles were added by the Work Group for this guideline. Had articles been added, they would have been subjected to the same a priori inclusion and exclusion criteria specified in Appendix II.
A total of three hundred thirty-two articles were reviewed for this guideline. Ninety-four articles met all a priori inclusion criteria. Two hundred and thirty eight articles were excluded for various reasons. Tracking these numbers in the flowcharts is not possible because a study could be included in more than one flowchart (i.e. some of the surgical studies were included in the infection flowchart). These numbers can be verified using the evidence tables and counting the references in the technical report. The flowcharts in Appendix II: Article Inclusions and Exclusions illustrate the number of articles retrieved for specific recommendations as well as the number of articles used to update systematic reviews.
For all recommendations except recommendation 3, we included only studies that diagnosed patients with a combination of electro-diagnostic tests and signs and symptoms. For recommendation 3, which addresses workplace issues, we required only that patients be diagnosed with signs and symptoms (see Appendix II: Article Inclusions and Exclusions). We relaxed the inclusion criteria for studies addressing CTS in the workplace because these patients are typically symptomatic and rarely receive electro-diagnostic tests to confirm their diagnosis. Even though we relaxed our inclusion criteria we were unable to find relevant literature which, to us, indicates a critical need for future research in this area.
We did not search for, or include, all available evidence. Wherever appropriate, we searched for and included the best available evidence. Hence, if Level II evidence was available, we did not search for or include Level III evidence or lower unless there was very little Level II evidence, and a great deal of Level III evidence.
Our analyses focused on patient oriented outcome measures. These measures are defined in clinical research as “outcomes that matter to patients including reduced morbidity, reduced mortality, symptom improvement, or improving patients’ quality of life”. By critically focusing on patient-oriented outcomes, the recommendations in this guideline are expected to improve overall patient care in the treatment of carpal tunnel syndrome.
We searched four electronic databases, MEDLINE, EMBASE, CINAHL and the Cochrane database of systematic reviews, to identify literature for this guideline. Search strategies were reviewed by the work group prior to conducting the searches. A list of the electronic databases we searched and the search strategies we used are provided in Appendix I: Literature Searches. All literature searches were supplemented with manual screening of bibliographies in publications accepted for inclusion into the evidence base. In addition, the bibliographies of recent review articles were searched for potentially relevant citations. All included articles met the specified a priori inclusion/exclusion criteria.
Assigning a Level of Evidence
The quality of evidence was rated using the evidence hierarchy shown in Appendix III: Rating Evidence Quality. A complete description of the hierarchy is included in the AAOS Evidence Report for this guideline. This hierarchy is also on the American Academy of Orthopaedic Surgeons (AAOS) website at: http://www.aaos.org/Research/Committee/Evidence/loetable1.pdf
Six reviewers independently completed data extraction for all studies. Evidence tables were constructed to summarize the best evidence pertaining to each recommendation and all evidence can be found in the accompanying Evidence Report to this guideline.
Grading the Recommendations
Each guideline recommendation was graded using the following system:
A: Good evidence (Level I Studies with consistent findings) for or against recommending intervention.
B: Fair evidence (Level II or III Studies with consistent findings) for or against recommending intervention.
C: Poor quality evidence (Level IV or V) for or against recommending intervention.
I: There is insufficient or conflicting evidence not allowing a recommendation for or against intervention.
The Committee used the following language in constructing the recommendations:
We recommend Treatment X: (for Grade A recommendations)
We suggest Treatment X: (for Grade B recommendations)
Treatment X is an option: (for Grade C recommendations)
These definitions help clarify the intent of the Work Group by reflecting the assessment of the importance of adherence to the recommendation based on the grade of the recommendation.
The statistical analyses performed help compare the treatment options available to patients with carpal tunnel syndrome. In order to assess specific treatments, comparisons were made between similar populations of patients receiving the treatment to patients receiving a control, placebo, or a second treatment. The goal of most treatment comparisons is to demonstrate that a treatment has a significant effect or that there is a significant difference between two treatments.
Small sample sizes in clinical trials present serious concerns because a lack of statistical power means that small but clinically important differences may go undetected. We calculated the minimal detectable difference to determine if a study was sufficiently powered for the given outcome. In our power calculations, we used 80% power, 95% confidence intervals and the number of patients per group. This allowed calculation of the minimal detectable effect size which was compared to the calculated effect size to determine if the study had enough power to detect the observed effect. If the trial was found to lack sufficient power for a given outcome, its results were taken as inconclusive. Power calculations were performed using G Power 3 (Version 3.0.5).14 Results are listed in the Evidence Report and Evidence Tables.
For recommendations one through eight in this guideline, several measures of association including the odds ratio (OR) and the natural log of the odds ratio (log OR) were used to compare treatments. In addition, the standardized mean difference (SMD) was used for computing standardized measures of effect size. Effect sizes were calculated when applicable; OR and log OR for dichotomous data and the SMD for continuous data. The larger the OR is, the larger the effect size. The SMD can be evaluated as follows: 0.2 for a small effect, 0.5 for a moderate effect and 0.8 for a large effect.
Studies had to have treatments, outcome measures, and durations of follow up in common to perform meta-analysis of the data. Given the paucity and heterogeneity of the data for specific recommendations, we did not apply formal meta-analytic techniques in
all circumstances. Log OR and Cohen’s h were computed for dichotomous outcome measures and SMD was computed for continuous outcome measures that were pooled for meta-analyses. When the event rate was zero for dichotomous outcome measures, a continuity correction was added. The Log OR and SMD values were then meta-analyzed using standard DerSimonian and Laird random effects model meta-analysis.17 When possible, effect sizes were pooled across different studies, and heterogeneity was assessed with the I-squared statistic.16 Summary statistics were presented when heterogeneity was less than 50%. All meta-analyses and effect size calculations were performed using STATA 10.0 (StataCorp LP, College Station, Texas) and the “metan” command.
Recommendation nine addressed the applicability of various instruments for the evaluation of carpal tunnel syndrome treatment in patients. Instruments are generally evaluated using eight key component areas: appropriateness, reliability, validity, responsiveness, precision, interpretability, acceptability, and feasibility. For this evaluation, we did not assess appropriateness, precision, interpretability, acceptability or feasibility. The physician should consider whether the instrument used was appropriate to measure CTS outcomes, that it contained the appropriate number of distinctions with regard to the dimensions being measured for precision and that the instruments were generally acceptable and feasible for use in the identified patient population. An overall summary of the properties assessed in each instrument (by primary study) is illustrated in the Evidence Tables.
Reliability, validity, and responsiveness were the three primary key components addressed in the studies for this recommendation. Reliable instruments are internally consistent and internal consistency is commonly measured by Cronbach’s alpha. This statistic measures how comparable the results of the instrument would be if the instrument were split into two versions or the average level of agreement of all the possible ways of performing split-half tests.
Validity was quantified using Spearman and Pearson correlation coefficients. Instruments with similar concepts should have large correlations and instruments with different concepts should have small correlations. A correlation coefficient of 0.5 and above is a large correlation, indicating two converging instruments, and a correlation coefficient below 0.5 is a smaller correlation, indicating two diverging instruments. A negative correlation indicates two instruments that score in the opposite direction.
Responsiveness was measured using the standardized response mean (SRM). The SRM is expressed as the change score divided by the standard deviation of the change score. A standardized response mean of 0.2 is indicative of a small change, 0.5 a medium change, and 0.8 a large change.
This guideline represents a cross-sectional view of current intervention methods and will become outdated when more sophisticated tests, more objective assessments and more rigorous differential diagnosis are possible. Linkage to other disorders, genetic diagnosis, and occupational and human factors literature will contribute to our understanding of the early stages of this condition and the means of differential treatment.
Because of the high profile of CTS in the workplace and the high level of interest in this topic, the guideline will be revised in accordance with changing practice, rapidly emerging opinion, new technology, and new evidence. It is anticipated that this guideline will be revised in 2011.