ArticleRegression‐Based Normative Data for Independent andCognitively Active Spanish Older Adults: Digit Span, Lettersand Numbers, Trail Making Test and Symbol DigitModalities TestClara Iñesta 1, Javier Oltra‐Cucarella 1,2,*, Beatriz Bonete‐López 1,2, Eva Calderón‐Rubio 1 and Esther Sitges‐Maciá 1,2SABIEX, Universidad Miguel Hernández de Elche, Avda. de la Universidad, 03207 Elche, Spain;[email protected] (C.I.); [email protected] (B.B.‐L.); [email protected] (E.C.‐R.);[email protected] (E.S.‐M.)2 Department of Health Psychology, Miguel Hernandez University of Elche, 03202 Elche, Spain* Correspondence: [email protected]: Iñesta, C.; Oltra‐Cucarella,J.; Bonete‐López, B.; Calderón‐Rubio,E.; Sitges‐Maciá, E. Regression‐BasedNormative Data for Independentand Cognitively Active SpanishOlder Adults: Digit Span, Lettersand Numbers, Trail Making Testand Symbol Digit Modalities Test.Int. J. Environ. Res. Public Health 2021,18, 9958. Editor: Paul B.Abstract: In this work, we developed normative data for the neuropsychological assessment of in‐dependent 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 commu‐nity‐dwelling adults aged 55 or older (67% women). Raw data for Digit Span (DS), Letters and Num‐bers (LN), the Trail Making Test (TMT), and the Symbol Digit Modalities Test (SDMT) were re‐gressed on age, sex, and education. The model predicting TMT‐B scores also included TMT‐Ascores. Z‐scores for the discrepancy between observed and predicted scores were used to identifylow scores. The base rate of low scores for SABIEX normative data was compared to the base rateof low scores using published normative data obtained from the general population. Results: Theeffects of age, sex, and education varied across neuropsychological measures. Although the propor‐tion of low scores was similar between normative datasets, there was no agreement in the identifi‐cation of cognitively impaired individuals. Conclusions: Normative data obtained from the generalpopulation might not be sensitive to identify low scores in cognitively active older adults, incor‐rectly classifying them as cognitively normal compared to the less‐active population. We provide afriendly calculator for use in neuropsychological assessment in cognitively active Spanish peopleaged 55 or older.TchounwouKeywords: cognitively active; cognitive impairment; neuropsychological assessment; normativedata; older adultsReceived: 29 July 2021Accepted: 18 September 2021Published: 22 September 2021Publisher’s Note: MDPI stays neu‐tral with regard to jurisdictionalclaims in published maps and institu‐tional affiliations.Copyright: 2021 by the author.Licensee MDPI, Basel, Switzerland.This article is an open access articledistributed under the terms andconditions of the Creative IntroductionThe population aged 65 years or older is expected to rise worldwide in the comingdecades. The United Nations [1] predicted an increase from 9% in 2020 to around 16% in2050. As reported by the Eurostat database, 20% of people in Europe are aged 65 or olderand this percentage is estimated to increase to 30% by 2070. According to the SpanishNational Statistics Institute [2], Spain is one of the countries with the highest rate of olderpeople in Europe, with 18.58% of people aged 65 years or older.Since age is the main risk factor for dementia [3,4], the increase in the proportion ofolder people is associated with an increase in the incidence and prevalence of cognitiveimpairment and dementia [5,6]. The number of people living with dementia worldwideis currently estimated at 50 million, with dementia being the leading cause of disabilityand dependence during aging [7]. A recent meta‐analysis reported a 12.4% prevalence ofdementia in Europe and 5–9% in Spain in people older than 65 [8]. Previous research hasfound that people diagnosed with Mild Cognitive Impairment (MCI) are at a higher riskInt. J. Environ. Res. Public Health 2021, 18, 9958. /journal/ijerph

Int. J. Environ. Res. Public Health 2021, 18, 99582 of 18of developing dementia [9]. Thus, in the absence of effective pharmacological and non‐pharmacological treatments for dementia [10–12], early detection of cognitive impairmentduring aging has become a major research topic.Neuropsychological assessment is essential to identify pathological cognitivechanges during aging [13,14]. Standardized tests are administered in order to assess thefunctioning of different cognitive domains such as attention, memory, language,visuospatial abilities, and executive functions. Performance is interpreted by comparingindividuals’ scores with scores from a reference group [13]. As raw scores in cognitivetests are affected by demographic variables such as age, sex, or educational level [15–17],normative data are used to transform them into relative measures corrected for the influ‐ence of these variables [16,18] and to provide a framework in which theses scores can belocated and interpreted. Thus, selecting appropriate normative datasets is necessary foraccurately interpreting the results of the neuropsychological assessment, and for reducingthe probability of false diagnoses of cognitive impairment [15,19].Different approaches to developing normative data have been reported. The simplestprocedure is based on the tests’ score distribution to generate norms from the means andstandard deviations. This strategy can be used with the entire sample or stratifying thesample by age [20,21], sex [22], and education [20,21,23]. Means and standard deviationswithin each subgroup are used to transform raw scores into easily interpretable measuressuch as Z scores, T‐scores, scaled scores, or percentile ranks [22]. This method has somelimitations: First, it is based on a series of arbitrary strata [24], assuming which personvariables are predictive of the test score; second, the estimated population means and var‐iances can be less reliable when dividing the sample into subgroups than using the wholesample [25]. A more advanced procedure to develop normative data is using multiplelinear regression models to estimate an individual’s predicted level of performance, basedon sociodemographic variables such as age, sex, and education. The difference betweenthe predicted and the observed score (residual values) is then standardized and inter‐preted [26–28]. A different procedure for clinical classification is the Receiver OperatingCharacteristic curve (ROC) analysis, which is used to determine the cut‐off score with theoptimal balance between sensitivity and specificity [29,30]. The area under the ROC curve(AUC) offers an index of the test’s overall discrimination accuracy, with values close to 1suggesting a high diagnostic accuracy.1.1. Active AgingAlthough brain changes during normal aging entail changes in some cognitive abili‐ties [31], certain activities are considered protective factors against cognitive decline, suchas continued learning and engagement in socially and cognitively stimulating activitiesduring aging [32,33]. This protective link is mostly attributed to an increased cognitivereserve, which compensates for brain changes in normal aging and delays the clinical ex‐pression of cognitive impairment despite underlying brain pathology caused by neuro‐degenerative processes [34,35]. Supporting these hypotheses, frequent participation incognitive activities has been associated with slower late‐life cognitive decline [36] and areduced risk of developing MCI and dementia [37].As a response to the challenges of population aging, the concept and policies of “Ac‐tive Aging” emerged. The Active Aging Framework promotes the optimization of oppor‐tunities for health, participation, and security with the aim of improving the quality of lifeas people age [38,39]. This notion emphasizes the importance of an active lifestyle and thebenefits of life‐long learning [40,41]. From this perspective, university programs for sen‐iors (UPS) have become an important resource for increasing opportunities for active ag‐ing, improving several aspects such as health, psychological well‐being, cognitive func‐tioning, autonomy maintenance, and social participation [41–43]. In recent decades, UPShave spread worldwide [40,44] and have prompted an increase in the number of olderadults that undertake university courses. In Spain, according to the State Association ofUniversity Programs for Older Adults (AEPUM), the number of adults aged 55 or older

Int. J. Environ. Res. Public Health 2021, 18, 99583 of 18enrolling in these programs increased from 23,000 during the 2005–2006 academic year to63,173 in 2018–2019 ( (accessed on 20 June, 2021).Older people who participate in university courses live independently in their eve‐ryday life and seek continued personal development and social interactions through theseeducational programs [45]. It has been reported that the motivations to attend these pro‐grams are to feel active, to invest in personal development, and to gain new knowledgeand social contacts [45,46]. The evidence also suggests that individuals who engage in UPSare cognitively more active than same‐age people in the general population. The tendencyto engage in these courses has been related to a larger number of individual and commu‐nity‐based active practices. Thus, cognitively active people read more frequently, do morephysical exercise, attend more cultural events, and participate more in social activities[42,47].It has been reported that cognitively stimulating activities in mid‐life [48] and latelife [49] contribute to cognitive reserve independently of education. Christensen et al. [50]found that the level of activity in everyday life influenced cognitive performance and ac‐counted for a greater proportion of variance in older people’s cognitive functioning thanthe level of education. In line with these results, in a post‐mortem study, Reed et al. [51]found that cognitive activities during adulthood have a higher influence than the level ofeducation in determining cognitive reserve. Thus, active aging is related to a series ofpractices in everyday life that differs from same‐age adults from the general population,contributing to a higher cognitive reserve that may preserve or enhance their cognitivefunction.1.2. Active Aging and Neuropsychological AssessmentThere is evidence suggesting that active older adults’ lifestyles will affect perfor‐mance on neuropsychological assessment irrespective of years of formal education. Activeolder people are likely to outperform non‐active individuals on cognitive tests [50,52].Even though normative data are demographically corrected by education, they do notaccount for the characteristics of active aging, and therefore, they might be less sensitivefor identifying cognitive impairment among cognitively active older adults with higherperformance levels. To the authors’ knowledge, there are no normative data for Spanishactive older adults. This implies that active older adults might present a diagnostic chal‐lenge in conditions such as cognitive impairment and AD, as pathological changes mightgo undetected in the neuropsychological assessment. To fill that gap, this study developednormative data on the assessment of attention, processing speed, and working memorythrough four cognitive tests widely used as part of the neuropsychological assessment.The Digit span forward (DSF) and backward (DSB) [53] are two frequently usedmeasures of attention and working memory. The Spanish edition of the WAIS‐III includesnormative data for Spanish individuals. There are also normative data of this test for sub‐jects over 50 in Spain within the NEURONORMA Project [54]. Some studies with healthycontrols and MCI and AD patients have reported the effectiveness of both subtests toidentify subtle impairments and to detect MCI [55], and to differentiate people with MCIand AD [56]. Lortie et al. [57] found that individuals with MCI declined in performanceover 6 months, suggesting that both subtests are a reliable measure for monitoring thedisease progression. The backward digit span subtest has also been reported to be a keyvariable discriminating between dementia subtypes such as AD and Dementia with LewyBodies [58].The Letters and Numbers subtest [53] is used as a measure of working memory. Somestudies provided normative data for this test in Spain [54] and Latin America [59] withadults from the general population. Kessel et al. [60] found that MCI and AD patientsperformed worse on this subtest compared with healthy controls. Worse performance inAD compared with MCI patients was also revealed, meaning it is suitable to differentiatepeople with MCI and AD.

Int. J. Environ. Res. Public Health 2021, 18, 99584 of 18The Symbol Digit Modalities Test (SDMT) [61] is used as a measure of informationprocessing speed. Norms obtained with healthy adults from the general population havebeen published in Spain [54] and in Latin America [62]. Smith also included normativedata in Spanish for an age range of 18 to 85 years for two schooling groups [61]. Perfor‐mance on the SDMT is a significant predictor of conversion from cognitively normal toMCI [63] and of progression from MCI to AD [64]. It is also one of the most commonlyused tests in the assessment of Multiple Sclerosis [65], Huntington’s Disease [66], and Par‐kinson’s disease [67].The Trail Making Test (TMT) [68] is widely used as a measure of attention and pro‐cessing speed [15]. TMT normative data have been reported for adults over 50 in Spainwithin the NEURONORMA Project [54] and for adults aged 18–95 in Latin America [69].The TMT is used to screen for neurodegenerative diseases in older adults, su