Development of a screening checklist to diagnose functional memory symptoms: a Delphi study

Introduction Currently there is no standardized way to diagnose functional cognitive disorder (FCD) in clinical practice. We aimed to develop a bedside screening tool to differentiate between FCD and neurodegenerative causes of memory complaints. Methods A Delphi methodology was used to obtain consensus based on expert opinion through a series of three consecutive rounds. International experts with an interest and experience in cognitive disorders and functional neurological disorders were selected. The first and second rounds included both clinical vignettes and a list of diagnostic clinical signs favoring a FCD diagnosis (obtained from a systematic review of the literature), to be ranked using a 7-point Likert scale. Qualitative and quantitative analysis were undertaken. Finally, experts reviewed the component items for which consensus had been achieved and provided recommendations for modifications to better discriminate between FCD and dementia. Results 45 experts, from 12 different countries, participated in the first round. A retention rate of at least 80% was obtained in subsequent rounds. 47% of the participants diagnosed between 10 and 50 patients with FCD over the last year. 69% reported that FCD comprises between 5 and 30% of the patients with memory complaints seen in the clinic. 91.4% would be interested in using a newly developed resource of this kind with the patients. A checklist was built and adapted into a scale, based on 11 statements for which strong (discrepancy between the level of symptoms reported and observed everyday functioning; patient’s ability to detail memory complaints with specific examples; and variable and/or fluctuating cognitive symptoms; 2 points each), moderate (ability to detail the list of drugs and/or previous interactions with other doctors; history of a non-cognitive functional neurological disorder or somatic symptoms; 1 point each) or low consensus (patient more aware of the cognitive changes than others; normal/incongruent cognitive performance; absence of symptom progression; dating symptom onset with precision; presence of psychological stressors; and ability to answer compound questions; 1 point each) was achieved. There was poor inter-rater agreement (Fleiss’ kappa=.035, p<.000). Regarding the vignettes, there was a major consensus gap in diagnosing FCD (<50% agreement) and neurodegenerative syndromes including Alzheimer’s disease (60–95% agreement). Conclusions This 11-item checklist is a simple complementary clinical tool aiming to improve diagnostic accuracy and direct patients to early post-diagnostic treatment. Further retrospective and prospective validation will be needed before it can be used clinically. This study is part of a wider project trying to improve the diagnosis of FCD.