Reliability And Agreement Of The Cognitive Domains And Functional Assessment Questionnaire - Informant Version (CDFAQ-IV)

Iara Souza Santos^1,2, João Vitor da Silva Viana^1,2, Aline Siqueira de Souza², Giovanna Correia Pereira Moro^1,2, Caio Peixoto Tavares², Isabelle de Aguiar Maia³, Guilherme Vieira Rodrigues da Costa³, Alberto Fonseca Lopes Rodrigues³, Joice Coutinho de Alvarenga², Gabriela Tomé Oliveira Engelmann², Marco Aurélio Fagundes Angelo⁴, Débora Marques de Miranda⁵, Marco Aurélio Romano-silva⁵, Jonas Jardim de Paula⁶, Luiz Armando Cunha de Marco⁵, Maria Aparecida Camargos Bicalho^6,7, Bernardo de Mattos Viana⁶.
¹Undergraduate Medicine, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, ²Cog-Aging Research Group, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, Minas Gerais, Brazil, ³Older Adult’s Psychiatry and Psychology Extension Program (PROEPSI), School of Medicine, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, Minas Gerais, Brazil, ⁴Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil, ⁵Neurotec R National Institute of Science and Technology (INCT-Neurotec R), Faculty of Medicine, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, Minas Gerais, Brazil, ⁶Molecular Medicine Postgraduate Program, School of Medicine, Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, Minas Gerais, Brazil, ⁷Department of Internal Medicine, School of Medicine, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil.

The CDFAQ-IV demonstrated excellent inter-rater reliability and moderate test–retest stability, supporting its use as a reliable informant-based tool for assessing cognitive and functional decline.

BACKGROUND Informant-based instruments assess cognitive and functional decline in neurocognitive disorders. The CDFAQ-IV evaluates six cognitive domains and functional abilities through ordinal informant reports; however, its inter-rater and test-retest reliability have not been extensively examined. Our objective was to evaluate the inter-rater agreement and reliability, and test-retest reliability of the CDFAQ-IV.

METHODS This cross-sectional study from the Cog-Aging cohort comprised two analyses. Inter-rater reliability was assessed in 16 informants of participants (cognitively unimpaired, mild cognitive impairment, or mild dementia) by evaluators independently recording responses and alternating interview conduction. Test-retest reliability involved 36 independent informants reassessed by the same evaluator after 2-12 weeks. The CDFAQ-IV covers six domains - Complex Attention (CA), Executive Function (EF), Learning and Memory (LM), Language (L), Perceptual-motor (PM), and Social Cognition (SC) - each with five items, yielding section and total scores. Analyses used weighted Cohen’s Kappa (item-level) and Intraclass Correlation Coefficients (ICC[2,1] for inter-rater; ICC[3,1] for test-retest), with absolute agreement and single measures, for section and total scores.

RESULTS In the inter-rater sample, participants (50% female) averaged 71.9±6.6 years of age and 10.1±6.4 years of education. Informants (68.8% female; mainly children or spouses) averaged 55.6±17.6 years of age and 13.1±6.2 years of education. Item-level agreement was mostly almost perfect (κ≥ 0.81). Items L2 and L3 showed substantial agreement, whereas PM3-PM5 were non-estimable due to absent variability. Section-level ICCs ranged from 0.960 to 1.000, and total score reliability was excellent (ICC=0.995; 95% CI: 0.986-0.998). In the test-retest sample, participants (63.9% female) averaged 75.8±8.3 years of age and 8.5±6.4 years of education. Informants (69.4% female; mainly children or spouses) averaged 53.9±15.2 years of age and 13.4±5.2 years of education. Item-level stability ranged from poor to substantial, with predominantly fair-to-moderate agreement, especially in EF, L, and SC. Section-level reliability was mostly moderate (0.552-0.719), with lower stability in PM (ICC=0.488). The total score demonstrated moderate test-retest reliability (ICC=0.692; 95% CI:0.474-0.830).

CONCLUSION Preliminary results indicate that CDFAQ-IV had excellent inter-rater reliability and predominantly almost perfect item-level agreement, with fair-to-moderate temporal stability.

REFERENCES 1. Diniz, M.F. (2021). Construção de um questionário de avaliação de domínios cognitivos e funcionalidade baseados nos critérios diagnósticos do DSM-5 para transtornos neurocognitivos. Universidade Federal de Minas Gerais, Master's Dissertation. 2. American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders (DSM-5). 5th ed. American Psychiatric Association. 3. Koo, T.K. & Li, M.Y. (2016). A guideline of selecting and reporting intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine, 15(2), 155–163.

CONTACT Iara Souza Santos Email: iarass2307@gmail.com Bernardo de Mattos Viana, MD, Ph.D Email: bernardomviana@yahoo.com