In this work, we developed normative data for the neuropsychological assessment of independent and cognitively active Spanish older adults over 55 years of age. Method: Regression-based normative data were calculated from a sample of 103 non-depressed independent community-dwelling adults aged 55 or older (67% women). Raw data for Digit Span (DS), Letters and Numbers (LN), the Trail Making Test (TMT), and the Symbol Digit Modalities Test (SDMT) were regressed on age, sex, and education. The model predicting TMT-B scores also included TMT-A scores. Z-scores for the discrepancy between observed and predicted scores were used to identify low scores. The base rate of low scores for SABIEX normative data was compared to the base rate of low scores using published normative data obtained from the general population. Results: The effects of age, sex, and education varied across neuropsychological measures. Although the proportion of low scores was similar between normative datasets, there was no agreement in the identification of cognitively impaired individuals. Conclusions: Normative data obtained from the general population might not be sensitive to identify low scores in cognitively active older adults, incorrectly classifying them as cognitively normal compared to the less-active population. We provide a friendly calculator for use in neuropsychological assessment in cognitively active Spanish people aged 55 or older.

The aim of this work was to develop normative data for neuropsychological tests for the assessment of independent and cognitively active Spanish older adults over 55 years of age. Methods: regression-based normative data were calculated from a sample of 103 nondepressed independent community-dwelling adults aged 55 or older (66% women). The raw data for the Free and Cued Selective Reminding Test (FCSRT), the Rey–Osterrieth Complex Figure Test (ROCF) and the Judgement of Line Orientation Test (JLO) were regressed on age, sex and education. The model predicting the FCSRT delayed-recall (FCSRT-Del) scores also included the FCSRT immediate-recall (FCSRT-Imm) scores. The model predicting the ROCF immediate-recall (ROCF-Imm) scores included the ROCF copy-trial (ROCF-C) scores, and the model predicting the ROCF delayed-recall (ROCF-Del) scores included both the ROCF-C and the ROCF-Imm scores. In order to identify low scores, z-scores were used to determine the discrepancy between the observed and the predicted scores. The base rates of the low scores for both the SABIEX normative data and the published normative data obtained from the general population were compared. Results: the effects of the different sociodemographic variables (age, sex and education) varied throughout the neuropsychological measures. Despite finding similar proportions of low scores between the normative data sets, the agreement was irrelevant or only fair-to-good. Conclusions: the normative data obtained from the general population might not be sensitive enough to identify low scores in cognitively active older adults, incorrectly classifying them as cognitively normal compared to the less active population.

An increased cognitive reserve is associated with changes in the pattern of cognitive decline during aging. Thus, normative data adapted to the characteristics of the target population are needed to reduce the possibility of false diagnoses. The aim of this work was to develop normative data for the Phonemic Verbal Fluency test, the Semantic Verbal Fluency test and the Boston Naming Test (BNT). Method: Regression-based normative data were calculated from a sample of 118 non-depressed, cognitively active, independent community-dwelling adults aged 55 or older (64.4% women) from SABIEX (University for Seniors at the Universidad Miguel Hernández de Elche). Raw scores were regressed on age, sex, and education. Results: The effects of age and education varied across neuropsychological measures. No effect of sex was found in any of the tests assessed. Statistically significant differences were found in the proportion of low scores using SABIEX or population-based normative datasets. The level of agreement identifying individuals labeled as showing one or more low scores was only fair-to-good. Conclusions: Normative data obtained from the general population might not be sensitive to identify low scores in cognitively active older adults, increasing the risk of misdiagnoses. A friendly calculator is available for neuropsychological assessment.

Regression-based normative data for neuropsychological variables are increasing in popularity over the last years. However, some use raw data while others use transformation when the observed response variable is skewed. This work analyzes how well the linear models fit for each type of variable. We used real data from a sample of n = 163 cognitively healthy individuals and compared the fit of linear regression models for raw scores and for corrected scaled scores. We then simulated a population of 1,000,000 individuals and drew 1,000 random samples of different sizes (n = 100, 200, 5,000, 1,000, 10,000) for seven different scenarios, analyzed the percentage of individuals scoring in the lowest 5%, and analyzed the agreement between models with the Cohen's κ statistic. Linear models for raw scores and for scaled scores were similar when the model included all the covariates, but barely identified low scores when scaled scores were corrected with covariates taken from different regressions (κ = 0.58). Models with raw scores showed that the expected number of individuals scoring low was close to the expected 5%, whereas models with scaled scores with covariates taken from different regressions were close to 0%. The two models agreed only when the response variable was random symmetrical and uncorrelated with the covariates. When calculating normative data using linear regressions, raw scores should be the preferred choice. If residuals analysis shows that the model does not fit the data well, researchers should consider using nonlinear models. Transforming data for normality of the observed response is discouraged. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

Objective: The use of screening tests is part of the neuropsychological assessment. Recently, the Attention, Memory and Frontal Abilities Screening Test (AMFAST) has been developed and translated to Spanish. However, its diagnostic properties have never been compared to other tests used to identify cognitive impairment. The goal of this pilot study is to compare the diagnostic properties of the AMFAST with those of the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD)-Neuropsychological Battery (CERAD-NB) in their ability to distinguish pre-defined clinical groups in a heterogeneous clinical sample of Spanish older adults. Methods: Ninety-four participants (66 cognitively unimpaired – CU; 28 cognitively impaired – CI) aged 50 or older were assessed with the AMFAST and the CERAD-NB, and differences between groups were analyzed with beta-binomial models. The diagnostic properties of both tests were calculated with adjusted-ROC curves, and areas under the curve (AUCs) were compared. Results: The CU group outperformed the CI group both on the AMFAST (b = −1.67, std. error (SE) = 0.19, p < .001) and on the CERAD-NB (b = −1.02, SE = 0.12, p < .001). The adjusted-ROC curves showed no differences in sensitivity and specificity compared to the model with test scores only, neither for the AMFAST (z = −1.03, p = .305) nor for the CERAD (z = −1.16, p = .246), with no statistically significant differences in the AUC (AMFAST = 0.954, CERAD = 0.954, p = .989). Conclusions: Our preliminary results based on a small, diagnostically heterogeneous sample suggest that the AMFAST and the CERAD-NB did not differ significantly to identify objective cognitive impairment. Future studies with larger, more homogenous samples are needed.