Guideline Development Process

Members of the Systematic Review Group and Guideline Writing Group members were required to disclose all potential conflicts of interest before appointment, before and during guideline development, and on publication.

The Guideline Writing Group was initially composed of eight psychiatrists with general research and clinical expertise. To achieve a multidisciplinary group, some experts from other disciplines (i.e., nursing, neurology, and geriatrics) were added to the group. In addition, individuals nominated as experts on the topic were surveyed, as described under the section “Expert Opinion Survey Data: Results” in Appendix B. The Guideline Writing Group was diverse and balanced with respect to its members’ expertise as well as other characteristics, such as geographical location and demographic background. Methodological expertise (i.e., with respect to appraisal of strength of research evidence) was provided by the Systematic Review Group. The Alzheimer’s Association was involved in reviewing the draft and provided perspective from patients, families, and other care partners.

An expert opinion survey was fielded to 593 experts on the topic of the guideline. These experts were peer-nominated by current and past APA Council and work group members, chairs of academic departments of psychiatry, directors of psychiatry residency programs in the United States and Canada, leadership of other medical organizations, and the APA Assembly. Nominators were asked to identify two types of experts to serve on the panel: researchers and clinicians. “Research experts” were defined as individuals who have significant research activities, scholarly publications, or academic reputation in the treatment of Alzheimer’s disease and other dementias, including the use of antipsychotic medications for the treatment of behavioral/psychological symptoms. “Clinical experts” were defined as individuals who have substantial clinical experience in the treatment of Alzheimer’s disease and other dementias, including the use of antipsychotic medications for the treatment of behavioral/psychological symptoms. The experts were contacted via email to complete the survey online.

Survey questions were adapted from clinical questions developed by the AHRQ for its 2011 review on off-label use of antipsychotics (Maglione et al. 2011). The survey included questions to address appropriate antipsychotic use, duration of treatment, and clinical experience of using antipsychotics to treat agitation or psychosis in patients with dementia in given clinical circumstances.

Most of the experts, 66.2%, were nominated once, 14.7% were nominated twice, and the remainder were nominated up to 19 times. The composition of the portion of the experts who responded to the survey corresponds closely with that of the entire panel, within 0%–5% (i.e., in the number of times panel members were nominated and whether they were identified as clinical or research experts or both).

The response rate for the survey was 34.4% (n = 204); 3.9% of the responses were partial, meaning that at least one question was completed. The experts who responded to the survey comprised approximately 61% clinical experts, 11% research experts, 24% experts in both categories, and 4% unspecified experts.

Quantitative data from the survey are shown in the section “Review of Supporting Research Evidence” in Appendix A. The survey also collected many free text comments, which were reviewed during development of the draft guideline. Key themes from qualitative data have been incorporated into the implementation section of the guideline.

This guideline is based on a systematic search of available research evidence. The search terms and limits used are available at http://psychiatryonline.org/pb-assets/books/Practice%20Guidelines/PG_Dementia_Search.pdf.

Initial searches of MEDLINE, PsycINFO, and Cochrane databases were conducted in February 2013 and included search terms for SGAs and for off-label indications for SGA use (including dementia), extending the search conducted for the AHRQ systematic review "Off-Label Use of Atypical Antipsychotics: An Update" (Maglione et al. 2011). These searches yielded 1,624 articles in MEDLINE, 657 articles in PsycINFO, and 1,457 articles in the Cochrane database. Two individuals (R.R. and L.J.F.) screened the 2,141 articles from the different searches when duplicate references were eliminated. Included articles were a clinical trial (including a controlled or randomized trial), observational study, meta-analysis, or systematic review that was clinically relevant to the off-label use of SGAs. The identified articles were subsequently restricted to the topic of dementia, and this yielded 12 articles (3 randomized trials, 9 observational studies).

Subsequent systematic searches were conducted in January 2015 and included terms for all antipsychotic medications and for all types of dementia, cognitive disorders, and cognitive impairment. Searches were limited to English language articles in adult humans and to clinical trials, observational studies, meta-analyses, and systematic reviews. All searches were done for the years from 1900 through 2014.These searches yielded 1,483 articles in MEDLINE, 470 articles in PsycINFO, and 335 articles in the Cochrane database. After duplicate articles and unpublished meeting abstracts were removed, two individuals (S-H.H. and L.F.) screened an additional 1,719 articles for relevance to the use of antipsychotic medications in individuals with dementia. Articles were included if they were randomized controlled trials that related to antipsychotic treatment of behavioral and psychological symptoms of dementia (BPSD). Because the AHRQ review (Maglione et al. 2011) only incorporated studies related to SGAs, we did not include randomized trials that only studied FGAs. We also excluded post-hoc analyses of pooled data and randomized trials that addressed acute use of intramuscular antipsychotic agents for the treatment of agitation. Observational studies, including administrative database studies, were included if they had a sample size of at least 500 individuals and addressed antipsychotic treatment of BPSD or harms of antipsychotic treatment in geriatric populations with or without dementia.

Results of this second search included all relevant articles that had been identified in the AHRQ review (Maglione et al. 2011) or in the initial search. Overall, 45 randomized controlled trials and 52 observational studies met the above criteria and were included in the guideline. An additional 4 studies appeared to meet these criteria upon screening the article title, but no abstracts were available and the full article could not be located. An additional 382 articles were related to dementia and antipsychotic treatment but did not meet the criteria noted above. Of these, 46 were meta-analyses or post-hoc analyses of pooled data and 13 were randomized controlled trials that only included an FGA. The remaining articles included 359 that were related to antipsychotic treatment but not dementia, 317 related to dementia but not antipsychotic treatment, and 560 that were unrelated to either dementia or antipsychotic treatment.

“Strength of supporting research evidence” describes the level of confidence that findings from scientific observation and testing of an effect of an intervention reflect the true effect. Confidence is enhanced by factors such as rigorous study design and minimal potential for study bias. Three ratings are used: high, moderate, and low.

Ratings are determined by the Systematic Review Group, after assessment of available clinical trials across four primary domains: risk of bias, consistency of findings across studies, directness of the effect on a specific health outcome, and precision of the estimate of effect. These domains and the method used to evaluate them are described above under “Systematic Review Methodology.”

In accordance with the AHRQ’s Methods Guide for Effectiveness and Comparative Effectiveness Reviews (Agency for Healthcare Research and Quality 2014), the ratings are defined as follows:

• High (denoted by the letter A) = High confidence that the evidence reflects the true effect. Further research is very unlikely to change our confidence in the estimate of effect.

• Moderate (denoted by the letter B) = Moderate confidence that the evidence reflects the true effect. Further research may change our confidence in the estimate of effect and may change the estimate.

• Low (denoted by the letter C) = Low confidence that the evidence reflects the true effect. Further research is likely to change our confidence in the estimate of effect and is likely to change the estimate.

Each guideline statement is separately rated to indicate strength of recommendation and strength of supporting research evidence.

“Strength of recommendation” describes the level of confidence that potential benefits of an intervention outweigh potential harms. This level of confidence is informed by available evidence, which includes evidence from clinical trials as well as expert opinion and patient values and preferences. As described in the introduction to this guideline (see “Rating the Strength of Supporting Research Evidence”), the rating is a consensus judgment of the authors of the guideline and is endorsed by the APA Board of Trustees.

There are two possible ratings: recommendation or suggestion. These ratings correspond to ratings of “strong” or “weak” (also termed “conditional”) as defined under the GRADE method for rating recommendations in clinical practice guidelines (described in publications such as Guyatt et al. 2008 and others available on the website of the GRADE Working Group at http://gradeworkinggroup.org/index.htm). “Recommendation” (denoted by the numeral 1 after the guideline statement) indicates confidence that the benefits of the intervention clearly outweigh harms. “Suggestion” (denoted by the numeral 2 after the guideline statement) indicates uncertainty (i.e., the balance of benefits and harms is difficult to judge or either the benefits or the harms are unclear).

When a negative statement is made, ratings of strength of recommendation should be understood as meaning the inverse of the above (e.g., “recommendation” indicates confidence that harms clearly outweigh benefits).

When there is insufficient information to support a recommendation or a suggestion, a statement may be made that further research about the intervention is needed.

The Guideline Writing Group determined ratings of strength of recommendation by a modified Delphi method using blind, iterative voting and discussion. In weighing potential benefits and harms, the group considered the strength of supporting research evidence, the results of the expert opinion survey, and their own clinical experiences and opinions. For recommendations, at least 9 of the 10 members of the group must have voted to “recommend” the intervention or assessment after four rounds of voting. On the basis of the discussion among the members of the group, adjustments to the wording of recommendations could be made between voting rounds. If this level of consensus was not achieved, the group could agree to make a “suggestion” rather than a recommendation. No suggestion or statement was made if three or more group members voted “no statement.” Differences of opinion within the group about ratings of strength of recommendation, if any, are described in the section “Potential Benefits and Harms” earlier in this guideline.

This guideline was made available for review on July 31, 2015 by stakeholders, including the APA membership, scientific and clinical experts, allied organizations (including patient advocacy organizations), and the public. A total of 44 individuals and 11 groups/organizations submitted comments on the guideline. The chair and co-chair of the Guideline Writing Group reviewed and addressed all comments received; substantive issues were reviewed by the Guideline Writing Group.

The guideline was approved by the APA Board of Trustees on December 13, 2015.