Predicting Premorbid Scores on the Repeatable Battery for the Assessment of Neuropsychological Status and their Validation in an Elderly Sample

Predicting Premorbid Scores on the Repeatable Battery for the Assessment of Neuropsychological... Abstract Objective Assessing cognitive change during a single visit requires the comparison of estimated premorbid abilities and current neuropsychological functioning. Although premorbid intellect has been widely examined, premorbid expectations for other cognitive abilities have received less attention. The current study sought to develop and validate premorbid estimates for the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). Method Using demographic variables and an estimate of premorbid intellect, premorbid performance on the RBANS was predicted in a sample of 143 community-dwelling, cognitively intact older adults. Results On all six Indexes of the RBANS, premorbid intellect was the best predictor of current cognitive functioning, with gender adding to one of the prediction models (R2 = 0.04–0.16, ps < .02). These prediction formulae were then applied to a sample of 122 individuals with amnestic Mild Cognitive Impairment to look for discrepancies between premorbid and current RBANS scores. Despite minimal differences between premorbid and current RBANS scores in the intact sample, large, and statistically significant differences were observed in the impaired sample, especially on the Immediate Memory Index (discrepancy = −29.00, p < .001), Delayed Memory Index (discrepancy = −32.28, p < .001), and Total Scale score (discrepancy = −25.58, p < .001). Conclusion Although validation in larger samples is needed, the current estimates of premorbid RBANS abilities may aid clinicians in determining change across time. Learning and memory, Mild cognitive impairment, Elderly/geriatrics/aging Introduction Assessing cognitive change within a single evaluation requires the clinician to infer cognitive decline based on a discrepancy between current neuropsychological functioning and estimates of premorbid abilities (Babcock, 1930; Lezak, Howieson, & Loring, 2004; McFie, 1975; Wechsler, 1958). Multiple techniques have been developed to estimate premorbid general cognitive functioning. For example, algorithms have been developed based on various demographic variables (Barona, Reynolds, & Chastain, 1984; Psychological Corporation, 2001; Reynolds & Gutkin, 1979; Schoenberg, Lange, Brickell, & Saklofske, 2007; Vanderploeg, Schinka, Baum, Tremont, & Mittenberg, 1998). “Hold” tests of current abilities, including single word reading tests or measures of vocabulary, can also estimate premorbid intellect (Blair & Spreen, 1989; Nelson & O’Connell, 1978; Nelson, 1982; Vanderploeg & Schinka, 1995). Academic records and achievement scores can indicate one’s prior cognitive abilities (Baade & Schoenberg, 2004). Finally, these methods have been combined (e.g., demographics and word reading tests) to shed light on this clinical dilemma (Crawford, Nelson, Blackmore, Cochrane, & Allan, 1990; Krull, Scott, & Sherer, 1995; Lange, Schoenberg, Chelune, Scott, & Adams, 2005; Schoenberg, Scott, Duff, & Adams, 2002; Vanderploeg & Schinka, 1995). At this time, no single method appears to be the “gold standard” (Powell, Brossart, & Reynolds, 2003; Schoenberg, Scott, Ruwe, Patton, & Adams, 2004). Although most estimates of premorbid abilities have focused on intellectual abilities, there is a growing body of work on predicting premorbid memory abilities. Williams (1997) reviewed estimates of premorbid memory, and noted that large discrepancies (approximately 23 standard score points) are needed between premorbid and current memory scores to infer “change.” However, it was also indicated that demographic variables and other performance measures might improve the prediction of prior memory abilities. Because this time, there have been a few attempts to predict premorbid memory (Duff, 2010; Duff, Chelune, & Dennett, 2011; Gladsjo, Heaton, Palmer, Taylor, & Jeste, 1999; Hilsabeck & Sutker, 2009; Isella et al., 2005; Schretlen, Buffington, Meyer, & Pearlson, 2005), which have found mixed relationships between demographic variables, premorbid intellect estimates, and memory functioning. Even fewer studies have attempted to predict non-memory premorbid estimates (Crawford, Moore, & Cameron, 1992; Schretlen et al., 2005). Some of this is due to the complicated relationship between intellect and memory and other cognitive abilities. For example, Binder, Iverson, and Brooks (2009) found that is relatively common for individuals in normative samples to show large discrepancies between at least two test scores, and that this was just as common in those with above average intellect. In both normative and patient samples, intellect and memory (and other cognitive scores) can be discrepant (Glass, Bartels, & Ryan, 2009; Hawkins & Tulsky, 2001; Nadolne, Adams, Scott, Hoffman, & Tremont, 1997; Skeel, Sitzer, Fogal, Wells, & Johnstone, 2004). In the current study, we sought to use demographic variables and an estimate of premorbid intellect to predict premorbid cognitive abilities on the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), a popular brief battery that examines multiple cognitive domains (Randolph, 2012), in cognitively intact older adults. Next, these prediction formulae were applied to a sample of individuals with amnestic Mild Cognitive Impairment (MCI) to look for discrepancies between premorbid and current RBANS scores. It was hypothesized that some combination of demographic variables and premorbid intellect scores would predict current RBANS scores in the intact sample. It was also expected that significant discrepancies would be found between premorbid and current RBANS scores in the impaired sample. Methods Participants Two samples of participants were used in this study. The first sample included one hundred forty-three community-dwelling older adults, who were classified as cognitively intact (see the following). These individuals were recruited from senior centers and independent living facilities to prospectively study practice effects in older adults. Their mean age was 75.3 (SD = 7.2, range = 65–96) years and their mean education was 15.6 (SD = 2.7, range = 8–22) years. Most were female (83%) and all were Caucasian. Premorbid intellect at baseline was average, as assessed by the Wide Range Achievement Test – third edition (WRAT-3) Reading subtest (standard score: M = 107.9, SD = 7.7, range = 81–126). Depression was minimal (Geriatric Depression Scale 30-item raw score: M = 4.02, SD = 3.7, range = 0–14). The second sample included one hundred twenty-two community-dwelling older adults, who were classified as either single or multidomain amnestic MCI (see the following). These individuals were recruited both from cognitive disorders clinics and community settings (e.g., senior centers and independent living facilities) to prospectively study practice effects in older adults. Their mean age was 75.8 (SD = 5.8, range = 65–89) years and their mean education was 16.4 (SD = 2.9, range = 12–25) years. Slightly more than half were male (54%) and all were Caucasian. Premorbid intellect at baseline was average, as assessed by the Wide Range Achievement Test -fourth edition (WRAT-4) Reading subtest (standard score: M = 108.5, SD = 8.5, range = 85–145). Depression was minimal (Geriatric Depression Scale 30-item raw score: M = 4.02, SD = 3.7, range = 0–14). Using results from a research interview with a participant and a knowledgeable informant and a baseline cognitive evaluation, these individuals were classified as either cognitively intact or amnestic MCI based on criteria by Albert et al. (2011) and Petersen (2004). These criteria include objective memory impairment, memory complaint, and largely intact functional activities. Objective memory and non-memory cognitive performances were dichotomized as intact or impaired using the seventh percentile as the cutoff (i.e., −1.5 SD, a common demarcation point for MCI). A self-reported memory complaint was required to be classified as MCI, but could be present or absent to be classified as intact. Functional status (e.g., driving, managing medications, handling finances, completing household chores) needed to be reported to be normal/typical/intact for all participants (intact and MCI). All cognitive data and classifications for both samples were reviewed by a neuropsychologist (KD). General inclusion criteria for both samples included: age 65 years or older and functionally independent (by report of the participant and a knowledgeable informant). General exclusion criteria for both samples included: neurological condition likely to adversely affect cognition (e.g., stroke, head injury with loss of consciousness of >30 min, epilepsy, Parkinson’s disease), dementia, major psychiatric condition (e.g., bipolar disorder, schizophrenia, uncontrolled major depression), current severe depression (i.e., score of >19 on 30-item Geriatric Depression Scale), substance abuse, taking anticonvulsant or antipsychotic medications, and residing in a nursing home or other skilled living facility. Procedures All participants provided informed consent prior to participation, and all procedures were approved by the local Institutional Review Board. During a baseline visit, all participants completed a battery of neuropsychological tests that included the RBANS and either the WRAT-3 or WRAT-4 Reading subtest. The RBANS is an individually administered set of twelve subtests that are used to calculate age-corrected scaled scores (M = 100, SD = 15) on the Indexes of Immediate Memory, Visuospatial/Constructional, Language, Attention, Delayed Memory, and Total Scale. The WRAT-3 or -4 Reading subtest is a list of 42 or 55 irregular words, respectively, used to estimate premorbid intellectual functioning, which also yields an age-corrected scaled score. Multiple studies have identified WRAT Reading tests as adequate estimates of premorbid intellect (Ahl et al., 2013; Berg, Durant, Banks, & Miller, 2016; Casaletto et al., 2014; O’Rourke et al., 2011). All tests were administered and scored as defined in their respective manuals by a trained research assistant. Data Analyses To develop the premorbid estimates, in the cognitively normal sample, a separate stepwise linear regression model was calculated for each Index score on the RBANS (criterion variable) using demographic variables (age, education, gender) and an estimate of premorbid intellect (WRAT-3 Reading) as the predictor variables. As an internal validation, the results of these six regression models were then applied to the cognitively normal sample to generate estimates of premorbid RBANS functioning in these participants. Dependent t-tests were used to compare the estimated premorbid RBANS scores to the observed current RBANS scores, with the expectation that premorbid and current RBANS scores should be comparable. To externally validate the premorbid estimates, in the MCI sample, the six regression models in Table 2 were applied to the MCI sample to generate estimates of premorbid RBANS functioning in these participants. Dependent t-tests were used to compare the estimated premorbid RBANS scores to the observed current RBANS scores, with the expectation that premorbid RBANS scores would be significantly higher than their observed current RBANS scores. An alpha value of p < .05 was used throughout. Results Development and Internal Validation of Premorbid Estimates The observed current RBANS Index scores for this cognitively intact sample are presented in Table 1. Not surprisingly, the mean scores for this sample were in the average range. Before conducting the regression analyses, data were checked for outliers (univariate and multivariate), normality, linearity, homoscedasticity, and multicollinearity, and no concerns were identified. In the six regression models, the WRAT-3 Reading significantly predicted all of the RBANS Indexes (see Table 2). The only demographic variable that added to any of these models was gender, which only added to the Language Index. The R2 values and standard error of the estimate are also included in Table 2. Table 1. Observed current and predicted premorbid RBANS scores in the intact sample RBANS Index Observed current Predicted premorbid t (df), p Immediate memory 110.24 (13.47) 110.23 (3.80) 0.0 (142), .99 78–152 97–119 Visuospatial/Constructional 105.17 (13.76) 104.88 (2.71) 0.3 (142), .80 78–131 95–111 Language 105.08 (10.94) 105.25 (3.01) −0.2 (142), .84 82–137 95–111 Attention 105.85 (14.11) 106.33 (3.95) −0.4 (142), .68 75–138 93–116 Delayed memory 108.85 (9.30) 108.84 (3.18) 0.0 (142), .98 84–131 98–116 Total scale 109.96 (12.09) 110.27 (4.88) −0.3 (142), .74 86–146 93–122 RBANS Index Observed current Predicted premorbid t (df), p Immediate memory 110.24 (13.47) 110.23 (3.80) 0.0 (142), .99 78–152 97–119 Visuospatial/Constructional 105.17 (13.76) 104.88 (2.71) 0.3 (142), .80 78–131 95–111 Language 105.08 (10.94) 105.25 (3.01) −0.2 (142), .84 82–137 95–111 Attention 105.85 (14.11) 106.33 (3.95) −0.4 (142), .68 75–138 93–116 Delayed memory 108.85 (9.30) 108.84 (3.18) 0.0 (142), .98 84–131 98–116 Total scale 109.96 (12.09) 110.27 (4.88) −0.3 (142), .74 86–146 93–122 Note. In Columns 2 and 3, means, standard deviations (in parentheses), and ranges are presented. Table 1. Observed current and predicted premorbid RBANS scores in the intact sample RBANS Index Observed current Predicted premorbid t (df), p Immediate memory 110.24 (13.47) 110.23 (3.80) 0.0 (142), .99 78–152 97–119 Visuospatial/Constructional 105.17 (13.76) 104.88 (2.71) 0.3 (142), .80 78–131 95–111 Language 105.08 (10.94) 105.25 (3.01) −0.2 (142), .84 82–137 95–111 Attention 105.85 (14.11) 106.33 (3.95) −0.4 (142), .68 75–138 93–116 Delayed memory 108.85 (9.30) 108.84 (3.18) 0.0 (142), .98 84–131 98–116 Total scale 109.96 (12.09) 110.27 (4.88) −0.3 (142), .74 86–146 93–122 RBANS Index Observed current Predicted premorbid t (df), p Immediate memory 110.24 (13.47) 110.23 (3.80) 0.0 (142), .99 78–152 97–119 Visuospatial/Constructional 105.17 (13.76) 104.88 (2.71) 0.3 (142), .80 78–131 95–111 Language 105.08 (10.94) 105.25 (3.01) −0.2 (142), .84 82–137 95–111 Attention 105.85 (14.11) 106.33 (3.95) −0.4 (142), .68 75–138 93–116 Delayed memory 108.85 (9.30) 108.84 (3.18) 0.0 (142), .98 84–131 98–116 Total scale 109.96 (12.09) 110.27 (4.88) −0.3 (142), .74 86–146 93–122 Note. In Columns 2 and 3, means, standard deviations (in parentheses), and ranges are presented. Table 2. Regression models in the intact sample RBANS Index R2 F (df), p SEE Prediction model Immediate memory 0.08 12.17 (142), .001 12.97 57.35 + (WRAT * 0.49) Visuospatial/Constructional 0.04 5.78 (142), .02 13.54 67.11 + (WRAT * 0.35) Language (Step 1) 0.04 6.56 (142), .01 10.74 100.08 + (Gender * 6.06) Language (Step 2) 0.08 5.68 (142), .004 10.60 73.72 + (WRAT * 0.25) + (Gender * 5.52) Attention 0.08 11.77 (142), .001 13.60 51.29 + (WRAT * 0.51) Delayed memory 0.12 18.64 (142), <.001 8.77 64.59 + (WRAT * 0.41) Total scale 0.16 27.13 (142), <.001 11.11 42.28 + (WRAT * 0.63) RBANS Index R2 F (df), p SEE Prediction model Immediate memory 0.08 12.17 (142), .001 12.97 57.35 + (WRAT * 0.49) Visuospatial/Constructional 0.04 5.78 (142), .02 13.54 67.11 + (WRAT * 0.35) Language (Step 1) 0.04 6.56 (142), .01 10.74 100.08 + (Gender * 6.06) Language (Step 2) 0.08 5.68 (142), .004 10.60 73.72 + (WRAT * 0.25) + (Gender * 5.52) Attention 0.08 11.77 (142), .001 13.60 51.29 + (WRAT * 0.51) Delayed memory 0.12 18.64 (142), <.001 8.77 64.59 + (WRAT * 0.41) Total scale 0.16 27.13 (142), <.001 11.11 42.28 + (WRAT * 0.63) Note: WRAT = standard score on the Reading subtest of the Wide Range Achievement Test; Gender: male = 0, female = 1; SEE = standard error of the estimate of the regression model. Table 2. Regression models in the intact sample RBANS Index R2 F (df), p SEE Prediction model Immediate memory 0.08 12.17 (142), .001 12.97 57.35 + (WRAT * 0.49) Visuospatial/Constructional 0.04 5.78 (142), .02 13.54 67.11 + (WRAT * 0.35) Language (Step 1) 0.04 6.56 (142), .01 10.74 100.08 + (Gender * 6.06) Language (Step 2) 0.08 5.68 (142), .004 10.60 73.72 + (WRAT * 0.25) + (Gender * 5.52) Attention 0.08 11.77 (142), .001 13.60 51.29 + (WRAT * 0.51) Delayed memory 0.12 18.64 (142), <.001 8.77 64.59 + (WRAT * 0.41) Total scale 0.16 27.13 (142), <.001 11.11 42.28 + (WRAT * 0.63) RBANS Index R2 F (df), p SEE Prediction model Immediate memory 0.08 12.17 (142), .001 12.97 57.35 + (WRAT * 0.49) Visuospatial/Constructional 0.04 5.78 (142), .02 13.54 67.11 + (WRAT * 0.35) Language (Step 1) 0.04 6.56 (142), .01 10.74 100.08 + (Gender * 6.06) Language (Step 2) 0.08 5.68 (142), .004 10.60 73.72 + (WRAT * 0.25) + (Gender * 5.52) Attention 0.08 11.77 (142), .001 13.60 51.29 + (WRAT * 0.51) Delayed memory 0.12 18.64 (142), <.001 8.77 64.59 + (WRAT * 0.41) Total scale 0.16 27.13 (142), <.001 11.11 42.28 + (WRAT * 0.63) Note: WRAT = standard score on the Reading subtest of the Wide Range Achievement Test; Gender: male = 0, female = 1; SEE = standard error of the estimate of the regression model. Using the results of these six regression models, premorbid RBANS estimates were calculated for these intact participants (see Table 1). When these predicted premorbid RBANS scores were compared to observed current RBANS scores, differences tended to be relatively small in this cognitively intact sample (mean differences <1 scaled score point), and none were statistically different (all dependent t-tests ps > .05). External Validation of Premorbid Estimates The observed current RBANS Index scores for the MCI sample are presented in Table 3. The premorbid RBANS estimates for these participants are also presented in Table 3. When these predicted premorbid RBANS scores were compared with observed current RBANS scores, significant differences were observed on all Indexes, with premorbid estimate scores being significantly higher than observed current scores (see Table 3). Table 3. Observed current and predicted premorbid RBANS scores in the MCI sample RBANS Index Observed current Predicted premorbid t (df), p, d Immediate memory 81.53 (17.13) 110.52 (4.19) −18.68 (121), <.001 44–114 99–128 2.4 Visuospatial/Constructional 97.48 (16.04) 105.09 (2.99) −5.15 (121), <.001 62–136 97–118 0.7 Language 91.45 (10.60) 103.38 (3.73) −12.63 (121), <.001 54–122 97–115 1.5 Attention 96.84 (15.19) 106.63 (4.36) −7.44 (121), <.001 64–132 95–125 0.9 Delayed memory 76.80 (21.57) 109.08 (3.50) −16.78 (121), <.001 40–117 99–124 2.1 Total scale 85.06 (13.35) 110.64 (5.38) −21.43 (121), <.001 50–121 96–134 2.5 RBANS Index Observed current Predicted premorbid t (df), p, d Immediate memory 81.53 (17.13) 110.52 (4.19) −18.68 (121), <.001 44–114 99–128 2.4 Visuospatial/Constructional 97.48 (16.04) 105.09 (2.99) −5.15 (121), <.001 62–136 97–118 0.7 Language 91.45 (10.60) 103.38 (3.73) −12.63 (121), <.001 54–122 97–115 1.5 Attention 96.84 (15.19) 106.63 (4.36) −7.44 (121), <.001 64–132 95–125 0.9 Delayed memory 76.80 (21.57) 109.08 (3.50) −16.78 (121), <.001 40–117 99–124 2.1 Total scale 85.06 (13.35) 110.64 (5.38) −21.43 (121), <.001 50–121 96–134 2.5 Note: In Columns 2 and 3, means, standard deviations (in parentheses), and ranges are presented. Table 3. Observed current and predicted premorbid RBANS scores in the MCI sample RBANS Index Observed current Predicted premorbid t (df), p, d Immediate memory 81.53 (17.13) 110.52 (4.19) −18.68 (121), <.001 44–114 99–128 2.4 Visuospatial/Constructional 97.48 (16.04) 105.09 (2.99) −5.15 (121), <.001 62–136 97–118 0.7 Language 91.45 (10.60) 103.38 (3.73) −12.63 (121), <.001 54–122 97–115 1.5 Attention 96.84 (15.19) 106.63 (4.36) −7.44 (121), <.001 64–132 95–125 0.9 Delayed memory 76.80 (21.57) 109.08 (3.50) −16.78 (121), <.001 40–117 99–124 2.1 Total scale 85.06 (13.35) 110.64 (5.38) −21.43 (121), <.001 50–121 96–134 2.5 RBANS Index Observed current Predicted premorbid t (df), p, d Immediate memory 81.53 (17.13) 110.52 (4.19) −18.68 (121), <.001 44–114 99–128 2.4 Visuospatial/Constructional 97.48 (16.04) 105.09 (2.99) −5.15 (121), <.001 62–136 97–118 0.7 Language 91.45 (10.60) 103.38 (3.73) −12.63 (121), <.001 54–122 97–115 1.5 Attention 96.84 (15.19) 106.63 (4.36) −7.44 (121), <.001 64–132 95–125 0.9 Delayed memory 76.80 (21.57) 109.08 (3.50) −16.78 (121), <.001 40–117 99–124 2.1 Total scale 85.06 (13.35) 110.64 (5.38) −21.43 (121), <.001 50–121 96–134 2.5 Note: In Columns 2 and 3, means, standard deviations (in parentheses), and ranges are presented. Discussion The current study sought to add to the limited literature on predicting premorbid cognitive abilities as a method to determine “decline” with a single neuropsychological assessment. Initially, in cognitively intact older adults, an estimate of premorbid intellect predicted all of the current scores on the RBANS. Additionally, gender significantly added to one of the prediction equations. These prediction formulae are consistent with previous attempts to predict premorbid memory (Duff, 2010; Duff et al., 2011; Gladsjo et al., 1999; Hilsabeck & Sutker, 2009; Isella et al., 2005; Schretlen et al., 2005), which have also found that demographic variables and premorbid intellect estimates can predict memory functioning. The current prediction formulae accounted for 4–16% of the variance in the actual RBANS Index scores, which suggests that there are other variables that need to be considered to better predict these Indexes. As an initial attempt to validate these prediction equations, they were applied to this same cohort. For all five Indexes and the Total Scale score of the RBANS, the predicted scores were statistically similar to the observed scores, with mean differences of <1 standard scale point. However, such small differences were expected, as they were developed on this same sample. Therefore, these equations need to be validated in “impaired” samples before they can be considered clinically useful. When the RBANS prediction equations from the intact sample were applied to the subjects with MCI, there were statistically significant differences for all Indexes, with the predicted premorbid scores being higher than the observed current scores. The largest differences (approximately two standard deviations, effect sizes >2) occurred on the immediate and delayed memory scales and the global cognitive composite. Smaller differences (less than one standard deviation, effect sizes ≤1.5) were observed on the scores tapping visuospatial and construction, language, and attention. This result would seem to suggest cognitive decline across multiple domains in these individuals with possible prodromal Alzheimer’s disease. In this way, this second set of results validated the models of predicted premorbid scores on the RBANS from the first study. Although further validation is needed, there is growing support for these estimates of premorbid cognitive abilities, which may aid clinicians and researchers in determining change across time in older patients. As noted earlier, two of the largest discrepancies between observed and predicted scores in the MCI sample were on the Immediate Memory Index and the Delayed Memory Index of the RBANS. This may not be that surprising in that these participants were all judged to meet a psychometric definition of amnestic MCI. However, this is also consistent with prior attempts to predict premorbid abilities using other memory tests. For example, we have previously developed premorbid estimates for the Hopkins Verbal Learning Test—Revised and Brief Visuospatial Memory Test—Revised (Duff, 2010). In this prior study, demographic variables (age, education, gender) contributed to all of the equations, whereas in the current study, demographic variables were not predictive of these two memory indexes on the RBANS. In the prior study, 9–17% of the variance was explained, whereas 8–12% were explained on these two RBANS Indexes. Similar to the current study, significant differences were observed between observed and predicted memory scores in an impaired sample in Duff (2010). The current results are also largely in line with other attempts to predict premorbid memory functioning (Gladsjo et al., 1999; Hilsabeck & Sutker, 2009; Isella et al., 2005; Schretlen et al., 2005). Besides adding to the limited literature on predicting premorbid memory functioning, the current study provided premorbid estimates for other cognitive domains, including visuospatial perception and construction abilities, language, and attention. Only a few other studies have developed premorbid estimates of abilities beyond intelligence and memory (Crawford et al., 1992; Schretlen et al., 2005). In their study, Crawford and colleagues (1992) predicted premorbid phonemic fluency in a sample of healthy individuals using the National Adult Reading Test, another commonly used estimate of premorbid intellect. In a neurologically compromised sample, Crawford et al. reported that the observed fluency scores were significantly below their predicted premorbid fluency scores. The results from this prior sample are similar to those seen in our current MCI sample, in which the discrepancies between observed and predicted scores for the non-memory Indexes were less dramatic (although still statistically significant). The current study also expands the work in this area to a commonly used screening measure in neuropsychology, the RBANS. For those who are less familiar with these prediction equations, a case example might be helpful. A 72-year-old male with 16 years of education presents for a neuropsychological evaluation due to recent complaints of memory problems. There are no prior evaluations with which to compare his current test scores. His age-corrected standard score on the WRAT-IV Reading subtest is 112, putting his premorbid intellect at the 79th percentile (i.e., “high average”). His observed current RBANS Index scores are presented in Table 4. For example, his observed current Delayed Memory Index was 92, which is at the 30th percentile (i.e., “average”). Although his current delayed memory score is lower than his premorbid intellect, it might not be enough to consider this a significant decline. However, if one calculates his premorbid predicted scores based on the equation in Table 2, then a different conclusion might be reached. For example, his premorbid predicted Delayed Memory Index was approximately 110 (i.e., [64.59 + {WRAT * 0.41}] = [64.59 + {112 * 0.41}] = [64.59 + 45.92] = 110.51). This reflects an 18-point discrepancy between observed current and premorbid predicted scores (i.e., 92–110). When this discrepancy is divided by the standard error of the estimate of the regression equation, this equates to a z-score of −2.11 (i.e., −18.51/8.77 = −2.11), which reflects a decline at the second percentile. Clinically, a provider would likely be more comfortable calling this a significant decline with the use of the premorbid predicted score. Table 4 shows a smaller decline on the Immediate Memory Index (z = −1.33) and Total Scale (z = −0.70), and no discernible declines on the other Indexes. An Excel spreadsheet that performs these calculations can be requested from the first author. Table 4. Case example RBANS Index Observed current Predicted premorbid Current–predicted z Immediate Memory 95 112.23 −17.23 −1.32 Visuospatial/Constructional 104 106.31 −2.31 −0.17 Language 100 101.72 −1.72 −0.16 Attention 106 108.41 −2.41 −0.18 Delayed Memory 92 110.51 −18.51 −2.11 Total Scale 105 112.84 −7.84 −0.70 RBANS Index Observed current Predicted premorbid Current–predicted z Immediate Memory 95 112.23 −17.23 −1.32 Visuospatial/Constructional 104 106.31 −2.31 −0.17 Language 100 101.72 −1.72 −0.16 Attention 106 108.41 −2.41 −0.18 Delayed Memory 92 110.51 −18.51 −2.11 Total Scale 105 112.84 −7.84 −0.70 Note: z = observed current–predicted premorbid/see from Table 2. Table 4. Case example RBANS Index Observed current Predicted premorbid Current–predicted z Immediate Memory 95 112.23 −17.23 −1.32 Visuospatial/Constructional 104 106.31 −2.31 −0.17 Language 100 101.72 −1.72 −0.16 Attention 106 108.41 −2.41 −0.18 Delayed Memory 92 110.51 −18.51 −2.11 Total Scale 105 112.84 −7.84 −0.70 RBANS Index Observed current Predicted premorbid Current–predicted z Immediate Memory 95 112.23 −17.23 −1.32 Visuospatial/Constructional 104 106.31 −2.31 −0.17 Language 100 101.72 −1.72 −0.16 Attention 106 108.41 −2.41 −0.18 Delayed Memory 92 110.51 −18.51 −2.11 Total Scale 105 112.84 −7.84 −0.70 Note: z = observed current–predicted premorbid/see from Table 2. Without these regression equations, a clinician might need to use the WRAT Reading score as an estimate of premorbid abilities across all RBANS Indexes. However, this practice could be problematic. First, as seen in this case example, although the WRAT Reading score was close to the predicted RBANS Indexes, they were not uniformly so. For example, the predicted Delayed Memory Index was nearly identical to the WRAT Reading score in this case (110.51 vs. 112), but larger differences were present for the Language (101.72) and Visuospatial/Constructional (106.31) Indexes. Second, and perhaps more importantly, if only the WRAT Reading score was available, then the clinician would need to use the standard deviation of the WRAT Reading score as the denominator in determining a discrepancy z-score. The population standard deviation of the WRAT Reading score is 15, which is larger than many of the standard error of the estimates from the regression equations (range: 8.77–13.60, see middle column of Table 2), which suggests less “decline” than if using the prediction equations. For example, using the WRAT Reading score, the discrepancy z-score for the Delayed Memory Index would be −1.3 (i.e., 92–112 = −20, −20/15 = z = −1.3) compared with the discrepancy z-score from the regression equation (i.e., 92–110.51 = −18.51, −18.51/8.77 = z = −2.11). However, it is worth reiterating that there can be notable discrepancies between intellect (including premorbid estimates of intellect) and memory and other cognitive abilities (Binder et al., 2009; Glass et al., 2009; Hawkins & Tulsky, 2001). These discrepancies may be more problematic in those with higher baseline intellectual abilities, like this case example and many in the current sample. Nonetheless, more precise estimates of predicted cognitive abilities and smaller standard error of the estimates seem to be clear advantages of the current method. Although the current results have some promise for clinical and research applications, some limitations should be noted. First, as indicated earlier, the amount of variance covered in the regression equations is relatively small (4–16%). However, this is consistent with prior studies that have made similar prediction formulae. Nonetheless, there is a need to identify other factors that improve these estimates. Second, these prediction formulae should be used very cautiously with individuals who do not match the demographic data of the intact sample from the first study. That is, these equations may be less appropriate if applied to individuals who are younger than 65 or older than 96, individuals with education outside the range of 8–22 years, or individuals with WRAT Reading scores below 81 or over 122 (however, this range account for about 82% of the area under a normal curve). Even though these demographic variables were not part of the prediction equations, they likely limit the generalizability of the equations. The current development and validation samples were also exclusively Caucasian. As such, it is doubtful that these prediction equations would be appropriate for other racial/ethnic groups. Additional work with a more diverse sample is clearly needed. Third, the current study used two different versions of the WRAT Reading subtest as its estimate of premorbid intellect, version 3 in the development sample and version 4 in the validation sample. Although it would have been ideal to use the same version of the WRAT Reading subtest, word reading premorbid estimates are highly related (Berg et al., 2016; Bright & van der Linde, 2018; Griffin, Mindt, Rankin, Ritchie, & Scott, 2002). It is also not known if other estimates (e.g., Test of Premorbid Functioning, National Adult Reading Test) would be as effective as predicting premorbid RBANS scores. Fourth, the current study did not use a consensus group to diagnose/classify participants as intact or MCI. A largely psychometric definition was used to classify individuals as “intact,” and a combination of a clinical diagnosis and a psychometric approach was used to classify individuals as MCI. Ideally, a consensus diagnosis for all participants would have been preferred. Fifth, additional validation of these prediction models in other older samples (e.g., patients with cancer, heart disease, stroke, depression, etc.) is warranted. Finally, this regression-based approach for estimating premorbid abilities is just one method of determining change. It has its own unique limitations, including poorly estimating outliers, needing access to prediction formulae like those developed here, and accounting for relatively little variance. Ultimately, some combination of demographic formulae, “hold” tests, and academic achievement records might be needed to most accurately estimate premorbid abilities in the individual patient. Funding The project described was supported by research grants from the National Institutes on Aging: R01AG045163. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute on Aging or the National Institutes of Health. Conflict of interest None declared. References Ahl , R. E. , Beiser , A. , Seshadri , S. , Auerbach , S. , Wolf , P. A. , & Au , R. ( 2013 ). Defining MCI in the Framingham Heart Study Offspring: Education versus WRAT-based norms . Alzheimer Disease and Associated Disorders , 27 , 330 – 336 . doi:10.1097/WAD.0b013e31827bde32 . Google Scholar CrossRef Search ADS Albert , M. S. , DeKosky , S. T. , Dickson , D. , Dubois , B. , Feldman , H. H. , Fox , N. C. , et al. . ( 2011 ). The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease . Alzheimer’s & Dementia : the Journal of the Alzheimer’s Association , 7 , 270 – 279 . Google Scholar CrossRef Search ADS Baade , L. E. , & Schoenberg , M. R. ( 2004 ). A proposed method to estimate premorbid intelligence utilizing group achievement measures from school records . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 19 , 227 – 243 . Google Scholar CrossRef Search ADS Babcock , H. ( 1930 ). An experiment in the measurement of mental deterioration . Archives of Psychology , 18 , 5 – 10 . Barona , A. , Reynolds , C. R. , & Chastain , R. ( 1984 ). A demographically based index of premorbid intelligence for the WAIS-R . Journal of Consulting and Clinical Psychology , 52 , 885 – 887 . Google Scholar CrossRef Search ADS Berg , J. L. , Durant , J. , Banks , S. J. , & Miller , J. B. ( 2016 ). Estimates of premorbid ability in a neurodegenerative disease clinic population: Comparing the Test of Premorbid Functioning and the Wide Range Achievement Test, 4th Edition . The Clinical Neuropsychologist , 30 , 547 – 557 . doi:10.1080/13854046.2016.1186224 . Google Scholar CrossRef Search ADS Binder , L. M. , Iverson , G. L. , & Brooks , B. L. ( 2009 ). To err is human: “Abnormal” neuropsychological scores and variability are common in healthy adults . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 24 , 31 – 46 . doi:10.1093/arclin/acn001 . Google Scholar CrossRef Search ADS Blair , J. , & Spreen , O. ( 1989 ). Predicting premorbid IQ: A revision of the National Adult Reading Test . The Clinical Neuropsychologist , 3 , 129 – 136 . Google Scholar CrossRef Search ADS Bright , P. , & van der Linde , I. ( 2018 ). Comparison of methods for estimating premorbid intelligence . Neuropsychological Rehabilitation , 1 – 14 . doi:10.1080/09602011.2018.1445650 . Casaletto , K. B. , Cattie , J. , Franklin , D. R. , Moore , D. J. , Woods , S. P. , Grant , I. , et al. . ( 2014 ). The Wide Range Achievement Test-4 Reading subtest “holds” in HIV-infected individuals . Journal of Clinical and Experimental Neuropsychology , 36 , 992 – 1001 . doi:10.1080/13803395.2014.960370 . Google Scholar CrossRef Search ADS Crawford , J. R. , Moore , J. W. , & Cameron , I. M. ( 1992 ). Verbal fluency: A NART-based equation for the estimation of premorbid performance . The British Journal of Clinical Psychology / the British Psychological Society , 31 , 327 – 329 . Google Scholar CrossRef Search ADS Crawford , J. R. , Nelson , H. E. , Blackmore , L. , Cochrane , R. H. B. , & Allan , K. M. ( 1990 ). Estimating premorbid intelligence by combining NART and demographic variables: An examination of the NART standardisation sample and supplementary equations . Personality and Individual Differences , 11 , 1153 – 1157 . Google Scholar CrossRef Search ADS Duff , K. ( 2010 ). Predicting premorbid memory functioning in older adults . Applied Neuropsychology , 17 , 278 – 282 . Google Scholar CrossRef Search ADS Duff , K. , Chelune , G. J. , & Dennett , K. ( 2011 ). Predicting estimates of premorbid memory functioning: validation in a dementia sample . Archives of Clinical Neuropsychology : the Official Journal of the National Academy of Neuropsychologists , 26 , 701 – 705 . Google Scholar CrossRef Search ADS Gladsjo , J. A. , Heaton , R. K. , Palmer , B. W. , Taylor , M. J. , & Jeste , D. V. ( 1999 ). Use of oral reading to estimate premorbid intellectual and neuropsychological functioning . Journal of the International Neuropsychological Society : JINS , 5 , 247 – 254 . Google Scholar CrossRef Search ADS Glass , L. A. , Bartels , J. M. , & Ryan , J. J. ( 2009 ). WAIS-III FSIQ and GAI in ability-memory discrepancy analysis . Applied Neuropsychology , 16 , 19 – 22 . doi:10.1080/09084280802644102 . Google Scholar CrossRef Search ADS Griffin , S. L. , Mindt , M. R. , Rankin , E. J. , Ritchie , A. J. , & Scott , J. G. ( 2002 ). Estimating premorbid intelligence: Comparison of traditional and contemporary methods across the intelligence continuum . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 17 , 497 – 507 . Google Scholar CrossRef Search ADS Hawkins , K. A. , & Tulsky , D. S. ( 2001 ). The influence of IQ stratification on WAIS-III/WMS-III FSIQ-general memory index discrepancy base-rates in the standardization sample . Journal of the International Neuropsychological Society : JINS , 7 , 875 – 880 . Hilsabeck , R. C. , & Sutker , P. B. ( 2009 ). Using an implicit memory task to estimate premorbid memory . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 24 , 179 – 191 . Google Scholar CrossRef Search ADS Isella , V. , Villa , M. L. , Forapani , E. , Piamarta , F. , Russo , A. , & Appollonio , I. M. ( 2005 ). Ineffectiveness of an Italian NART-equivalent for the estimation of verbal learning ability in normal elderly . Journal of Clinical and Experimental Neuropsychology , 27 , 618 – 623 . Google Scholar CrossRef Search ADS Krull , K. R. , Scott , J. G. , & Sherer , M. ( 1995 ). Estimation of premorbid intelligence from combined performance and demongraphic variables . The Clinical Neuropsychologist , 9 , 83 – 88 . Google Scholar CrossRef Search ADS Lange , R. T. , Schoenberg , M. R. , Chelune , G. J. , Scott , J. G. , & Adams , R. L. ( 2005 ). Development of the WAIS-III general ability index estimate (GAI-E) . The Clinical Neuropsychologist , 19 , 73 – 86 . Google Scholar CrossRef Search ADS Lezak , M. D. , Howieson , D. B. , & Loring , D. W. ( 2004 ). Neuropsychological assessment ( 4th ed. ). New York : Oxford University Press . McFie , J. ( 1975 ). Assessment of organic intellectual impairment . New York : Academic Press . Nadolne , M. J. , Adams , R. L. , Scott , J. G. , Hoffman , R. G. , & Tremont , G. ( 1997 ). Clinical utility of Wechsler Memory Scale-Revised and predicted IQ discrepancies in closed head injury . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 12 , 757 – 762 . Google Scholar CrossRef Search ADS Nelson , H. E. ( 1982 ). National adult reading test (NART): Test manual . Windsor, England : NFER-Nelson . Nelson , H. E. , & O’Connell , A. ( 1978 ). Dementia: The estimation of premorbid intelligence levels using the New Adult Reading Test . Cortex; A Journal Devoted to the Study of the Nervous System and Behavior , 14 , 234 – 244 . Google Scholar CrossRef Search ADS O’Rourke , J. J. , Adams , W. H. , Duff , K. , Byars , J. , Nopoulos , P. , Paulsen , J. S. , et al. . ( 2011 ). Estimating premorbid functioning in Huntington’s disease: The relationship between disease progression and the wide range achievement test reading subtest . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 26 , 59 – 66 . Google Scholar CrossRef Search ADS Petersen , R. C. ( 2004 ). Mild cognitive impairment as a diagnostic entity . Journal of Internal Medicine , 256 , 183 – 194 . Google Scholar CrossRef Search ADS Powell , B. D. , Brossart , D. F. , & Reynolds , C. R. ( 2003 ). Evaluation of the accuracy of two regression-based methods for estimating premorbid IQ . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 18 , 277 – 292 . Google Scholar CrossRef Search ADS Psychological Corporation . ( 2001 ). Wechsler Test of Adult Reading Manual . San Antonio : Author . Randolph , C. ( 2012 ). Repeatable battery for the assessment of neuropsychological status . Bloomington, MN : The Psychological Corporation . Reynolds , C. , & Gutkin , T. B. ( 1979 ). Predicting the premorbid intellectual status of children using demographic data . The Clinical Neuropsychologist , 1 , 36 – 38 . Schoenberg , M. R. , Lange , R. T. , Brickell , T. A. , & Saklofske , D. H. ( 2007 ). Estimating premorbid general cognitive functioning for children and adolescents using the American Wechsler Intelligence Scale for Children-Fourth Edition: Demographic and current performance approaches . Journal of Child Neurology , 22 , 379 – 388 . Google Scholar CrossRef Search ADS Schoenberg , M. R. , Scott , J. G. , Duff , K. , & Adams , R. L. ( 2002 ). Estimation of WAIS-III intelligence from combined performance and demographic variables: Development of the OPIE-3 . The Clinical Neuropsychologist , 16 , 426 – 437 . Google Scholar CrossRef Search ADS Schoenberg , M. R. , Scott , J. G. , Ruwe , W. , Patton , D. , & Adams , R. L. ( 2004 ). Assumptions that underlie predicting premorbid IQ: A comment on the “evaluation of the accuracy of two regression-based methods for estimating premorbid IQ” . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 19 , 1103 – 1106 . Google Scholar CrossRef Search ADS Schretlen , D. J. , Buffington , A. L. , Meyer , S. M. , & Pearlson , G. D. ( 2005 ). The use of word-reading to estimate “premorbid” ability in cognitive domains other than intelligence . Journal of the International Neuropsychological Society : JINS , 11 , 784 – 787 . Google Scholar CrossRef Search ADS Skeel , R. L. , Sitzer , D. , Fogal , T. , Wells , J. , & Johnstone , B. ( 2004 ). Comparison of Predicted-difference, Simple-difference, and Premorbid-estimation methodologies for evaluating IQ and memory score discrepancies . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 19 , 363 – 374 . doi:10.1016/S0887-6177(03)00072-6 . Google Scholar CrossRef Search ADS Vanderploeg , R. D. , & Schinka , J. A. ( 1995 ). Predicting WAIS-R IQ premorbid ability: combining subtest performance and demographic variable predictors . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 10 , 225 – 239 . Google Scholar CrossRef Search ADS Vanderploeg , R. D. , Schinka , J. A. , Baum , K. M. , Tremont , G. , & Mittenberg , W. ( 1998 ). WISC-III premorbid prediction strategies: Demographic and best performance approaches . Psychological Assessment , 10 , 277 – 284 . Google Scholar CrossRef Search ADS Wechsler , D. ( 1958 ). The measurement and appraisal of adult intelligence ( 4th ed. ). Baltimore : Williams & Wilkins . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Archives of Clinical Neuropsychology Oxford University Press

Predicting Premorbid Scores on the Repeatable Battery for the Assessment of Neuropsychological Status and their Validation in an Elderly Sample

Loading next page...
 
/lp/ou_press/predicting-premorbid-scores-on-the-repeatable-battery-for-the-UphNGTD2ZG
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0887-6177
eISSN
1873-5843
D.O.I.
10.1093/arclin/acy050
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective Assessing cognitive change during a single visit requires the comparison of estimated premorbid abilities and current neuropsychological functioning. Although premorbid intellect has been widely examined, premorbid expectations for other cognitive abilities have received less attention. The current study sought to develop and validate premorbid estimates for the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). Method Using demographic variables and an estimate of premorbid intellect, premorbid performance on the RBANS was predicted in a sample of 143 community-dwelling, cognitively intact older adults. Results On all six Indexes of the RBANS, premorbid intellect was the best predictor of current cognitive functioning, with gender adding to one of the prediction models (R2 = 0.04–0.16, ps < .02). These prediction formulae were then applied to a sample of 122 individuals with amnestic Mild Cognitive Impairment to look for discrepancies between premorbid and current RBANS scores. Despite minimal differences between premorbid and current RBANS scores in the intact sample, large, and statistically significant differences were observed in the impaired sample, especially on the Immediate Memory Index (discrepancy = −29.00, p < .001), Delayed Memory Index (discrepancy = −32.28, p < .001), and Total Scale score (discrepancy = −25.58, p < .001). Conclusion Although validation in larger samples is needed, the current estimates of premorbid RBANS abilities may aid clinicians in determining change across time. Learning and memory, Mild cognitive impairment, Elderly/geriatrics/aging Introduction Assessing cognitive change within a single evaluation requires the clinician to infer cognitive decline based on a discrepancy between current neuropsychological functioning and estimates of premorbid abilities (Babcock, 1930; Lezak, Howieson, & Loring, 2004; McFie, 1975; Wechsler, 1958). Multiple techniques have been developed to estimate premorbid general cognitive functioning. For example, algorithms have been developed based on various demographic variables (Barona, Reynolds, & Chastain, 1984; Psychological Corporation, 2001; Reynolds & Gutkin, 1979; Schoenberg, Lange, Brickell, & Saklofske, 2007; Vanderploeg, Schinka, Baum, Tremont, & Mittenberg, 1998). “Hold” tests of current abilities, including single word reading tests or measures of vocabulary, can also estimate premorbid intellect (Blair & Spreen, 1989; Nelson & O’Connell, 1978; Nelson, 1982; Vanderploeg & Schinka, 1995). Academic records and achievement scores can indicate one’s prior cognitive abilities (Baade & Schoenberg, 2004). Finally, these methods have been combined (e.g., demographics and word reading tests) to shed light on this clinical dilemma (Crawford, Nelson, Blackmore, Cochrane, & Allan, 1990; Krull, Scott, & Sherer, 1995; Lange, Schoenberg, Chelune, Scott, & Adams, 2005; Schoenberg, Scott, Duff, & Adams, 2002; Vanderploeg & Schinka, 1995). At this time, no single method appears to be the “gold standard” (Powell, Brossart, & Reynolds, 2003; Schoenberg, Scott, Ruwe, Patton, & Adams, 2004). Although most estimates of premorbid abilities have focused on intellectual abilities, there is a growing body of work on predicting premorbid memory abilities. Williams (1997) reviewed estimates of premorbid memory, and noted that large discrepancies (approximately 23 standard score points) are needed between premorbid and current memory scores to infer “change.” However, it was also indicated that demographic variables and other performance measures might improve the prediction of prior memory abilities. Because this time, there have been a few attempts to predict premorbid memory (Duff, 2010; Duff, Chelune, & Dennett, 2011; Gladsjo, Heaton, Palmer, Taylor, & Jeste, 1999; Hilsabeck & Sutker, 2009; Isella et al., 2005; Schretlen, Buffington, Meyer, & Pearlson, 2005), which have found mixed relationships between demographic variables, premorbid intellect estimates, and memory functioning. Even fewer studies have attempted to predict non-memory premorbid estimates (Crawford, Moore, & Cameron, 1992; Schretlen et al., 2005). Some of this is due to the complicated relationship between intellect and memory and other cognitive abilities. For example, Binder, Iverson, and Brooks (2009) found that is relatively common for individuals in normative samples to show large discrepancies between at least two test scores, and that this was just as common in those with above average intellect. In both normative and patient samples, intellect and memory (and other cognitive scores) can be discrepant (Glass, Bartels, & Ryan, 2009; Hawkins & Tulsky, 2001; Nadolne, Adams, Scott, Hoffman, & Tremont, 1997; Skeel, Sitzer, Fogal, Wells, & Johnstone, 2004). In the current study, we sought to use demographic variables and an estimate of premorbid intellect to predict premorbid cognitive abilities on the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), a popular brief battery that examines multiple cognitive domains (Randolph, 2012), in cognitively intact older adults. Next, these prediction formulae were applied to a sample of individuals with amnestic Mild Cognitive Impairment (MCI) to look for discrepancies between premorbid and current RBANS scores. It was hypothesized that some combination of demographic variables and premorbid intellect scores would predict current RBANS scores in the intact sample. It was also expected that significant discrepancies would be found between premorbid and current RBANS scores in the impaired sample. Methods Participants Two samples of participants were used in this study. The first sample included one hundred forty-three community-dwelling older adults, who were classified as cognitively intact (see the following). These individuals were recruited from senior centers and independent living facilities to prospectively study practice effects in older adults. Their mean age was 75.3 (SD = 7.2, range = 65–96) years and their mean education was 15.6 (SD = 2.7, range = 8–22) years. Most were female (83%) and all were Caucasian. Premorbid intellect at baseline was average, as assessed by the Wide Range Achievement Test – third edition (WRAT-3) Reading subtest (standard score: M = 107.9, SD = 7.7, range = 81–126). Depression was minimal (Geriatric Depression Scale 30-item raw score: M = 4.02, SD = 3.7, range = 0–14). The second sample included one hundred twenty-two community-dwelling older adults, who were classified as either single or multidomain amnestic MCI (see the following). These individuals were recruited both from cognitive disorders clinics and community settings (e.g., senior centers and independent living facilities) to prospectively study practice effects in older adults. Their mean age was 75.8 (SD = 5.8, range = 65–89) years and their mean education was 16.4 (SD = 2.9, range = 12–25) years. Slightly more than half were male (54%) and all were Caucasian. Premorbid intellect at baseline was average, as assessed by the Wide Range Achievement Test -fourth edition (WRAT-4) Reading subtest (standard score: M = 108.5, SD = 8.5, range = 85–145). Depression was minimal (Geriatric Depression Scale 30-item raw score: M = 4.02, SD = 3.7, range = 0–14). Using results from a research interview with a participant and a knowledgeable informant and a baseline cognitive evaluation, these individuals were classified as either cognitively intact or amnestic MCI based on criteria by Albert et al. (2011) and Petersen (2004). These criteria include objective memory impairment, memory complaint, and largely intact functional activities. Objective memory and non-memory cognitive performances were dichotomized as intact or impaired using the seventh percentile as the cutoff (i.e., −1.5 SD, a common demarcation point for MCI). A self-reported memory complaint was required to be classified as MCI, but could be present or absent to be classified as intact. Functional status (e.g., driving, managing medications, handling finances, completing household chores) needed to be reported to be normal/typical/intact for all participants (intact and MCI). All cognitive data and classifications for both samples were reviewed by a neuropsychologist (KD). General inclusion criteria for both samples included: age 65 years or older and functionally independent (by report of the participant and a knowledgeable informant). General exclusion criteria for both samples included: neurological condition likely to adversely affect cognition (e.g., stroke, head injury with loss of consciousness of >30 min, epilepsy, Parkinson’s disease), dementia, major psychiatric condition (e.g., bipolar disorder, schizophrenia, uncontrolled major depression), current severe depression (i.e., score of >19 on 30-item Geriatric Depression Scale), substance abuse, taking anticonvulsant or antipsychotic medications, and residing in a nursing home or other skilled living facility. Procedures All participants provided informed consent prior to participation, and all procedures were approved by the local Institutional Review Board. During a baseline visit, all participants completed a battery of neuropsychological tests that included the RBANS and either the WRAT-3 or WRAT-4 Reading subtest. The RBANS is an individually administered set of twelve subtests that are used to calculate age-corrected scaled scores (M = 100, SD = 15) on the Indexes of Immediate Memory, Visuospatial/Constructional, Language, Attention, Delayed Memory, and Total Scale. The WRAT-3 or -4 Reading subtest is a list of 42 or 55 irregular words, respectively, used to estimate premorbid intellectual functioning, which also yields an age-corrected scaled score. Multiple studies have identified WRAT Reading tests as adequate estimates of premorbid intellect (Ahl et al., 2013; Berg, Durant, Banks, & Miller, 2016; Casaletto et al., 2014; O’Rourke et al., 2011). All tests were administered and scored as defined in their respective manuals by a trained research assistant. Data Analyses To develop the premorbid estimates, in the cognitively normal sample, a separate stepwise linear regression model was calculated for each Index score on the RBANS (criterion variable) using demographic variables (age, education, gender) and an estimate of premorbid intellect (WRAT-3 Reading) as the predictor variables. As an internal validation, the results of these six regression models were then applied to the cognitively normal sample to generate estimates of premorbid RBANS functioning in these participants. Dependent t-tests were used to compare the estimated premorbid RBANS scores to the observed current RBANS scores, with the expectation that premorbid and current RBANS scores should be comparable. To externally validate the premorbid estimates, in the MCI sample, the six regression models in Table 2 were applied to the MCI sample to generate estimates of premorbid RBANS functioning in these participants. Dependent t-tests were used to compare the estimated premorbid RBANS scores to the observed current RBANS scores, with the expectation that premorbid RBANS scores would be significantly higher than their observed current RBANS scores. An alpha value of p < .05 was used throughout. Results Development and Internal Validation of Premorbid Estimates The observed current RBANS Index scores for this cognitively intact sample are presented in Table 1. Not surprisingly, the mean scores for this sample were in the average range. Before conducting the regression analyses, data were checked for outliers (univariate and multivariate), normality, linearity, homoscedasticity, and multicollinearity, and no concerns were identified. In the six regression models, the WRAT-3 Reading significantly predicted all of the RBANS Indexes (see Table 2). The only demographic variable that added to any of these models was gender, which only added to the Language Index. The R2 values and standard error of the estimate are also included in Table 2. Table 1. Observed current and predicted premorbid RBANS scores in the intact sample RBANS Index Observed current Predicted premorbid t (df), p Immediate memory 110.24 (13.47) 110.23 (3.80) 0.0 (142), .99 78–152 97–119 Visuospatial/Constructional 105.17 (13.76) 104.88 (2.71) 0.3 (142), .80 78–131 95–111 Language 105.08 (10.94) 105.25 (3.01) −0.2 (142), .84 82–137 95–111 Attention 105.85 (14.11) 106.33 (3.95) −0.4 (142), .68 75–138 93–116 Delayed memory 108.85 (9.30) 108.84 (3.18) 0.0 (142), .98 84–131 98–116 Total scale 109.96 (12.09) 110.27 (4.88) −0.3 (142), .74 86–146 93–122 RBANS Index Observed current Predicted premorbid t (df), p Immediate memory 110.24 (13.47) 110.23 (3.80) 0.0 (142), .99 78–152 97–119 Visuospatial/Constructional 105.17 (13.76) 104.88 (2.71) 0.3 (142), .80 78–131 95–111 Language 105.08 (10.94) 105.25 (3.01) −0.2 (142), .84 82–137 95–111 Attention 105.85 (14.11) 106.33 (3.95) −0.4 (142), .68 75–138 93–116 Delayed memory 108.85 (9.30) 108.84 (3.18) 0.0 (142), .98 84–131 98–116 Total scale 109.96 (12.09) 110.27 (4.88) −0.3 (142), .74 86–146 93–122 Note. In Columns 2 and 3, means, standard deviations (in parentheses), and ranges are presented. Table 1. Observed current and predicted premorbid RBANS scores in the intact sample RBANS Index Observed current Predicted premorbid t (df), p Immediate memory 110.24 (13.47) 110.23 (3.80) 0.0 (142), .99 78–152 97–119 Visuospatial/Constructional 105.17 (13.76) 104.88 (2.71) 0.3 (142), .80 78–131 95–111 Language 105.08 (10.94) 105.25 (3.01) −0.2 (142), .84 82–137 95–111 Attention 105.85 (14.11) 106.33 (3.95) −0.4 (142), .68 75–138 93–116 Delayed memory 108.85 (9.30) 108.84 (3.18) 0.0 (142), .98 84–131 98–116 Total scale 109.96 (12.09) 110.27 (4.88) −0.3 (142), .74 86–146 93–122 RBANS Index Observed current Predicted premorbid t (df), p Immediate memory 110.24 (13.47) 110.23 (3.80) 0.0 (142), .99 78–152 97–119 Visuospatial/Constructional 105.17 (13.76) 104.88 (2.71) 0.3 (142), .80 78–131 95–111 Language 105.08 (10.94) 105.25 (3.01) −0.2 (142), .84 82–137 95–111 Attention 105.85 (14.11) 106.33 (3.95) −0.4 (142), .68 75–138 93–116 Delayed memory 108.85 (9.30) 108.84 (3.18) 0.0 (142), .98 84–131 98–116 Total scale 109.96 (12.09) 110.27 (4.88) −0.3 (142), .74 86–146 93–122 Note. In Columns 2 and 3, means, standard deviations (in parentheses), and ranges are presented. Table 2. Regression models in the intact sample RBANS Index R2 F (df), p SEE Prediction model Immediate memory 0.08 12.17 (142), .001 12.97 57.35 + (WRAT * 0.49) Visuospatial/Constructional 0.04 5.78 (142), .02 13.54 67.11 + (WRAT * 0.35) Language (Step 1) 0.04 6.56 (142), .01 10.74 100.08 + (Gender * 6.06) Language (Step 2) 0.08 5.68 (142), .004 10.60 73.72 + (WRAT * 0.25) + (Gender * 5.52) Attention 0.08 11.77 (142), .001 13.60 51.29 + (WRAT * 0.51) Delayed memory 0.12 18.64 (142), <.001 8.77 64.59 + (WRAT * 0.41) Total scale 0.16 27.13 (142), <.001 11.11 42.28 + (WRAT * 0.63) RBANS Index R2 F (df), p SEE Prediction model Immediate memory 0.08 12.17 (142), .001 12.97 57.35 + (WRAT * 0.49) Visuospatial/Constructional 0.04 5.78 (142), .02 13.54 67.11 + (WRAT * 0.35) Language (Step 1) 0.04 6.56 (142), .01 10.74 100.08 + (Gender * 6.06) Language (Step 2) 0.08 5.68 (142), .004 10.60 73.72 + (WRAT * 0.25) + (Gender * 5.52) Attention 0.08 11.77 (142), .001 13.60 51.29 + (WRAT * 0.51) Delayed memory 0.12 18.64 (142), <.001 8.77 64.59 + (WRAT * 0.41) Total scale 0.16 27.13 (142), <.001 11.11 42.28 + (WRAT * 0.63) Note: WRAT = standard score on the Reading subtest of the Wide Range Achievement Test; Gender: male = 0, female = 1; SEE = standard error of the estimate of the regression model. Table 2. Regression models in the intact sample RBANS Index R2 F (df), p SEE Prediction model Immediate memory 0.08 12.17 (142), .001 12.97 57.35 + (WRAT * 0.49) Visuospatial/Constructional 0.04 5.78 (142), .02 13.54 67.11 + (WRAT * 0.35) Language (Step 1) 0.04 6.56 (142), .01 10.74 100.08 + (Gender * 6.06) Language (Step 2) 0.08 5.68 (142), .004 10.60 73.72 + (WRAT * 0.25) + (Gender * 5.52) Attention 0.08 11.77 (142), .001 13.60 51.29 + (WRAT * 0.51) Delayed memory 0.12 18.64 (142), <.001 8.77 64.59 + (WRAT * 0.41) Total scale 0.16 27.13 (142), <.001 11.11 42.28 + (WRAT * 0.63) RBANS Index R2 F (df), p SEE Prediction model Immediate memory 0.08 12.17 (142), .001 12.97 57.35 + (WRAT * 0.49) Visuospatial/Constructional 0.04 5.78 (142), .02 13.54 67.11 + (WRAT * 0.35) Language (Step 1) 0.04 6.56 (142), .01 10.74 100.08 + (Gender * 6.06) Language (Step 2) 0.08 5.68 (142), .004 10.60 73.72 + (WRAT * 0.25) + (Gender * 5.52) Attention 0.08 11.77 (142), .001 13.60 51.29 + (WRAT * 0.51) Delayed memory 0.12 18.64 (142), <.001 8.77 64.59 + (WRAT * 0.41) Total scale 0.16 27.13 (142), <.001 11.11 42.28 + (WRAT * 0.63) Note: WRAT = standard score on the Reading subtest of the Wide Range Achievement Test; Gender: male = 0, female = 1; SEE = standard error of the estimate of the regression model. Using the results of these six regression models, premorbid RBANS estimates were calculated for these intact participants (see Table 1). When these predicted premorbid RBANS scores were compared to observed current RBANS scores, differences tended to be relatively small in this cognitively intact sample (mean differences <1 scaled score point), and none were statistically different (all dependent t-tests ps > .05). External Validation of Premorbid Estimates The observed current RBANS Index scores for the MCI sample are presented in Table 3. The premorbid RBANS estimates for these participants are also presented in Table 3. When these predicted premorbid RBANS scores were compared with observed current RBANS scores, significant differences were observed on all Indexes, with premorbid estimate scores being significantly higher than observed current scores (see Table 3). Table 3. Observed current and predicted premorbid RBANS scores in the MCI sample RBANS Index Observed current Predicted premorbid t (df), p, d Immediate memory 81.53 (17.13) 110.52 (4.19) −18.68 (121), <.001 44–114 99–128 2.4 Visuospatial/Constructional 97.48 (16.04) 105.09 (2.99) −5.15 (121), <.001 62–136 97–118 0.7 Language 91.45 (10.60) 103.38 (3.73) −12.63 (121), <.001 54–122 97–115 1.5 Attention 96.84 (15.19) 106.63 (4.36) −7.44 (121), <.001 64–132 95–125 0.9 Delayed memory 76.80 (21.57) 109.08 (3.50) −16.78 (121), <.001 40–117 99–124 2.1 Total scale 85.06 (13.35) 110.64 (5.38) −21.43 (121), <.001 50–121 96–134 2.5 RBANS Index Observed current Predicted premorbid t (df), p, d Immediate memory 81.53 (17.13) 110.52 (4.19) −18.68 (121), <.001 44–114 99–128 2.4 Visuospatial/Constructional 97.48 (16.04) 105.09 (2.99) −5.15 (121), <.001 62–136 97–118 0.7 Language 91.45 (10.60) 103.38 (3.73) −12.63 (121), <.001 54–122 97–115 1.5 Attention 96.84 (15.19) 106.63 (4.36) −7.44 (121), <.001 64–132 95–125 0.9 Delayed memory 76.80 (21.57) 109.08 (3.50) −16.78 (121), <.001 40–117 99–124 2.1 Total scale 85.06 (13.35) 110.64 (5.38) −21.43 (121), <.001 50–121 96–134 2.5 Note: In Columns 2 and 3, means, standard deviations (in parentheses), and ranges are presented. Table 3. Observed current and predicted premorbid RBANS scores in the MCI sample RBANS Index Observed current Predicted premorbid t (df), p, d Immediate memory 81.53 (17.13) 110.52 (4.19) −18.68 (121), <.001 44–114 99–128 2.4 Visuospatial/Constructional 97.48 (16.04) 105.09 (2.99) −5.15 (121), <.001 62–136 97–118 0.7 Language 91.45 (10.60) 103.38 (3.73) −12.63 (121), <.001 54–122 97–115 1.5 Attention 96.84 (15.19) 106.63 (4.36) −7.44 (121), <.001 64–132 95–125 0.9 Delayed memory 76.80 (21.57) 109.08 (3.50) −16.78 (121), <.001 40–117 99–124 2.1 Total scale 85.06 (13.35) 110.64 (5.38) −21.43 (121), <.001 50–121 96–134 2.5 RBANS Index Observed current Predicted premorbid t (df), p, d Immediate memory 81.53 (17.13) 110.52 (4.19) −18.68 (121), <.001 44–114 99–128 2.4 Visuospatial/Constructional 97.48 (16.04) 105.09 (2.99) −5.15 (121), <.001 62–136 97–118 0.7 Language 91.45 (10.60) 103.38 (3.73) −12.63 (121), <.001 54–122 97–115 1.5 Attention 96.84 (15.19) 106.63 (4.36) −7.44 (121), <.001 64–132 95–125 0.9 Delayed memory 76.80 (21.57) 109.08 (3.50) −16.78 (121), <.001 40–117 99–124 2.1 Total scale 85.06 (13.35) 110.64 (5.38) −21.43 (121), <.001 50–121 96–134 2.5 Note: In Columns 2 and 3, means, standard deviations (in parentheses), and ranges are presented. Discussion The current study sought to add to the limited literature on predicting premorbid cognitive abilities as a method to determine “decline” with a single neuropsychological assessment. Initially, in cognitively intact older adults, an estimate of premorbid intellect predicted all of the current scores on the RBANS. Additionally, gender significantly added to one of the prediction equations. These prediction formulae are consistent with previous attempts to predict premorbid memory (Duff, 2010; Duff et al., 2011; Gladsjo et al., 1999; Hilsabeck & Sutker, 2009; Isella et al., 2005; Schretlen et al., 2005), which have also found that demographic variables and premorbid intellect estimates can predict memory functioning. The current prediction formulae accounted for 4–16% of the variance in the actual RBANS Index scores, which suggests that there are other variables that need to be considered to better predict these Indexes. As an initial attempt to validate these prediction equations, they were applied to this same cohort. For all five Indexes and the Total Scale score of the RBANS, the predicted scores were statistically similar to the observed scores, with mean differences of <1 standard scale point. However, such small differences were expected, as they were developed on this same sample. Therefore, these equations need to be validated in “impaired” samples before they can be considered clinically useful. When the RBANS prediction equations from the intact sample were applied to the subjects with MCI, there were statistically significant differences for all Indexes, with the predicted premorbid scores being higher than the observed current scores. The largest differences (approximately two standard deviations, effect sizes >2) occurred on the immediate and delayed memory scales and the global cognitive composite. Smaller differences (less than one standard deviation, effect sizes ≤1.5) were observed on the scores tapping visuospatial and construction, language, and attention. This result would seem to suggest cognitive decline across multiple domains in these individuals with possible prodromal Alzheimer’s disease. In this way, this second set of results validated the models of predicted premorbid scores on the RBANS from the first study. Although further validation is needed, there is growing support for these estimates of premorbid cognitive abilities, which may aid clinicians and researchers in determining change across time in older patients. As noted earlier, two of the largest discrepancies between observed and predicted scores in the MCI sample were on the Immediate Memory Index and the Delayed Memory Index of the RBANS. This may not be that surprising in that these participants were all judged to meet a psychometric definition of amnestic MCI. However, this is also consistent with prior attempts to predict premorbid abilities using other memory tests. For example, we have previously developed premorbid estimates for the Hopkins Verbal Learning Test—Revised and Brief Visuospatial Memory Test—Revised (Duff, 2010). In this prior study, demographic variables (age, education, gender) contributed to all of the equations, whereas in the current study, demographic variables were not predictive of these two memory indexes on the RBANS. In the prior study, 9–17% of the variance was explained, whereas 8–12% were explained on these two RBANS Indexes. Similar to the current study, significant differences were observed between observed and predicted memory scores in an impaired sample in Duff (2010). The current results are also largely in line with other attempts to predict premorbid memory functioning (Gladsjo et al., 1999; Hilsabeck & Sutker, 2009; Isella et al., 2005; Schretlen et al., 2005). Besides adding to the limited literature on predicting premorbid memory functioning, the current study provided premorbid estimates for other cognitive domains, including visuospatial perception and construction abilities, language, and attention. Only a few other studies have developed premorbid estimates of abilities beyond intelligence and memory (Crawford et al., 1992; Schretlen et al., 2005). In their study, Crawford and colleagues (1992) predicted premorbid phonemic fluency in a sample of healthy individuals using the National Adult Reading Test, another commonly used estimate of premorbid intellect. In a neurologically compromised sample, Crawford et al. reported that the observed fluency scores were significantly below their predicted premorbid fluency scores. The results from this prior sample are similar to those seen in our current MCI sample, in which the discrepancies between observed and predicted scores for the non-memory Indexes were less dramatic (although still statistically significant). The current study also expands the work in this area to a commonly used screening measure in neuropsychology, the RBANS. For those who are less familiar with these prediction equations, a case example might be helpful. A 72-year-old male with 16 years of education presents for a neuropsychological evaluation due to recent complaints of memory problems. There are no prior evaluations with which to compare his current test scores. His age-corrected standard score on the WRAT-IV Reading subtest is 112, putting his premorbid intellect at the 79th percentile (i.e., “high average”). His observed current RBANS Index scores are presented in Table 4. For example, his observed current Delayed Memory Index was 92, which is at the 30th percentile (i.e., “average”). Although his current delayed memory score is lower than his premorbid intellect, it might not be enough to consider this a significant decline. However, if one calculates his premorbid predicted scores based on the equation in Table 2, then a different conclusion might be reached. For example, his premorbid predicted Delayed Memory Index was approximately 110 (i.e., [64.59 + {WRAT * 0.41}] = [64.59 + {112 * 0.41}] = [64.59 + 45.92] = 110.51). This reflects an 18-point discrepancy between observed current and premorbid predicted scores (i.e., 92–110). When this discrepancy is divided by the standard error of the estimate of the regression equation, this equates to a z-score of −2.11 (i.e., −18.51/8.77 = −2.11), which reflects a decline at the second percentile. Clinically, a provider would likely be more comfortable calling this a significant decline with the use of the premorbid predicted score. Table 4 shows a smaller decline on the Immediate Memory Index (z = −1.33) and Total Scale (z = −0.70), and no discernible declines on the other Indexes. An Excel spreadsheet that performs these calculations can be requested from the first author. Table 4. Case example RBANS Index Observed current Predicted premorbid Current–predicted z Immediate Memory 95 112.23 −17.23 −1.32 Visuospatial/Constructional 104 106.31 −2.31 −0.17 Language 100 101.72 −1.72 −0.16 Attention 106 108.41 −2.41 −0.18 Delayed Memory 92 110.51 −18.51 −2.11 Total Scale 105 112.84 −7.84 −0.70 RBANS Index Observed current Predicted premorbid Current–predicted z Immediate Memory 95 112.23 −17.23 −1.32 Visuospatial/Constructional 104 106.31 −2.31 −0.17 Language 100 101.72 −1.72 −0.16 Attention 106 108.41 −2.41 −0.18 Delayed Memory 92 110.51 −18.51 −2.11 Total Scale 105 112.84 −7.84 −0.70 Note: z = observed current–predicted premorbid/see from Table 2. Table 4. Case example RBANS Index Observed current Predicted premorbid Current–predicted z Immediate Memory 95 112.23 −17.23 −1.32 Visuospatial/Constructional 104 106.31 −2.31 −0.17 Language 100 101.72 −1.72 −0.16 Attention 106 108.41 −2.41 −0.18 Delayed Memory 92 110.51 −18.51 −2.11 Total Scale 105 112.84 −7.84 −0.70 RBANS Index Observed current Predicted premorbid Current–predicted z Immediate Memory 95 112.23 −17.23 −1.32 Visuospatial/Constructional 104 106.31 −2.31 −0.17 Language 100 101.72 −1.72 −0.16 Attention 106 108.41 −2.41 −0.18 Delayed Memory 92 110.51 −18.51 −2.11 Total Scale 105 112.84 −7.84 −0.70 Note: z = observed current–predicted premorbid/see from Table 2. Without these regression equations, a clinician might need to use the WRAT Reading score as an estimate of premorbid abilities across all RBANS Indexes. However, this practice could be problematic. First, as seen in this case example, although the WRAT Reading score was close to the predicted RBANS Indexes, they were not uniformly so. For example, the predicted Delayed Memory Index was nearly identical to the WRAT Reading score in this case (110.51 vs. 112), but larger differences were present for the Language (101.72) and Visuospatial/Constructional (106.31) Indexes. Second, and perhaps more importantly, if only the WRAT Reading score was available, then the clinician would need to use the standard deviation of the WRAT Reading score as the denominator in determining a discrepancy z-score. The population standard deviation of the WRAT Reading score is 15, which is larger than many of the standard error of the estimates from the regression equations (range: 8.77–13.60, see middle column of Table 2), which suggests less “decline” than if using the prediction equations. For example, using the WRAT Reading score, the discrepancy z-score for the Delayed Memory Index would be −1.3 (i.e., 92–112 = −20, −20/15 = z = −1.3) compared with the discrepancy z-score from the regression equation (i.e., 92–110.51 = −18.51, −18.51/8.77 = z = −2.11). However, it is worth reiterating that there can be notable discrepancies between intellect (including premorbid estimates of intellect) and memory and other cognitive abilities (Binder et al., 2009; Glass et al., 2009; Hawkins & Tulsky, 2001). These discrepancies may be more problematic in those with higher baseline intellectual abilities, like this case example and many in the current sample. Nonetheless, more precise estimates of predicted cognitive abilities and smaller standard error of the estimates seem to be clear advantages of the current method. Although the current results have some promise for clinical and research applications, some limitations should be noted. First, as indicated earlier, the amount of variance covered in the regression equations is relatively small (4–16%). However, this is consistent with prior studies that have made similar prediction formulae. Nonetheless, there is a need to identify other factors that improve these estimates. Second, these prediction formulae should be used very cautiously with individuals who do not match the demographic data of the intact sample from the first study. That is, these equations may be less appropriate if applied to individuals who are younger than 65 or older than 96, individuals with education outside the range of 8–22 years, or individuals with WRAT Reading scores below 81 or over 122 (however, this range account for about 82% of the area under a normal curve). Even though these demographic variables were not part of the prediction equations, they likely limit the generalizability of the equations. The current development and validation samples were also exclusively Caucasian. As such, it is doubtful that these prediction equations would be appropriate for other racial/ethnic groups. Additional work with a more diverse sample is clearly needed. Third, the current study used two different versions of the WRAT Reading subtest as its estimate of premorbid intellect, version 3 in the development sample and version 4 in the validation sample. Although it would have been ideal to use the same version of the WRAT Reading subtest, word reading premorbid estimates are highly related (Berg et al., 2016; Bright & van der Linde, 2018; Griffin, Mindt, Rankin, Ritchie, & Scott, 2002). It is also not known if other estimates (e.g., Test of Premorbid Functioning, National Adult Reading Test) would be as effective as predicting premorbid RBANS scores. Fourth, the current study did not use a consensus group to diagnose/classify participants as intact or MCI. A largely psychometric definition was used to classify individuals as “intact,” and a combination of a clinical diagnosis and a psychometric approach was used to classify individuals as MCI. Ideally, a consensus diagnosis for all participants would have been preferred. Fifth, additional validation of these prediction models in other older samples (e.g., patients with cancer, heart disease, stroke, depression, etc.) is warranted. Finally, this regression-based approach for estimating premorbid abilities is just one method of determining change. It has its own unique limitations, including poorly estimating outliers, needing access to prediction formulae like those developed here, and accounting for relatively little variance. Ultimately, some combination of demographic formulae, “hold” tests, and academic achievement records might be needed to most accurately estimate premorbid abilities in the individual patient. Funding The project described was supported by research grants from the National Institutes on Aging: R01AG045163. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute on Aging or the National Institutes of Health. Conflict of interest None declared. References Ahl , R. E. , Beiser , A. , Seshadri , S. , Auerbach , S. , Wolf , P. A. , & Au , R. ( 2013 ). Defining MCI in the Framingham Heart Study Offspring: Education versus WRAT-based norms . Alzheimer Disease and Associated Disorders , 27 , 330 – 336 . doi:10.1097/WAD.0b013e31827bde32 . Google Scholar CrossRef Search ADS Albert , M. S. , DeKosky , S. T. , Dickson , D. , Dubois , B. , Feldman , H. H. , Fox , N. C. , et al. . ( 2011 ). The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease . Alzheimer’s & Dementia : the Journal of the Alzheimer’s Association , 7 , 270 – 279 . Google Scholar CrossRef Search ADS Baade , L. E. , & Schoenberg , M. R. ( 2004 ). A proposed method to estimate premorbid intelligence utilizing group achievement measures from school records . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 19 , 227 – 243 . Google Scholar CrossRef Search ADS Babcock , H. ( 1930 ). An experiment in the measurement of mental deterioration . Archives of Psychology , 18 , 5 – 10 . Barona , A. , Reynolds , C. R. , & Chastain , R. ( 1984 ). A demographically based index of premorbid intelligence for the WAIS-R . Journal of Consulting and Clinical Psychology , 52 , 885 – 887 . Google Scholar CrossRef Search ADS Berg , J. L. , Durant , J. , Banks , S. J. , & Miller , J. B. ( 2016 ). Estimates of premorbid ability in a neurodegenerative disease clinic population: Comparing the Test of Premorbid Functioning and the Wide Range Achievement Test, 4th Edition . The Clinical Neuropsychologist , 30 , 547 – 557 . doi:10.1080/13854046.2016.1186224 . Google Scholar CrossRef Search ADS Binder , L. M. , Iverson , G. L. , & Brooks , B. L. ( 2009 ). To err is human: “Abnormal” neuropsychological scores and variability are common in healthy adults . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 24 , 31 – 46 . doi:10.1093/arclin/acn001 . Google Scholar CrossRef Search ADS Blair , J. , & Spreen , O. ( 1989 ). Predicting premorbid IQ: A revision of the National Adult Reading Test . The Clinical Neuropsychologist , 3 , 129 – 136 . Google Scholar CrossRef Search ADS Bright , P. , & van der Linde , I. ( 2018 ). Comparison of methods for estimating premorbid intelligence . Neuropsychological Rehabilitation , 1 – 14 . doi:10.1080/09602011.2018.1445650 . Casaletto , K. B. , Cattie , J. , Franklin , D. R. , Moore , D. J. , Woods , S. P. , Grant , I. , et al. . ( 2014 ). The Wide Range Achievement Test-4 Reading subtest “holds” in HIV-infected individuals . Journal of Clinical and Experimental Neuropsychology , 36 , 992 – 1001 . doi:10.1080/13803395.2014.960370 . Google Scholar CrossRef Search ADS Crawford , J. R. , Moore , J. W. , & Cameron , I. M. ( 1992 ). Verbal fluency: A NART-based equation for the estimation of premorbid performance . The British Journal of Clinical Psychology / the British Psychological Society , 31 , 327 – 329 . Google Scholar CrossRef Search ADS Crawford , J. R. , Nelson , H. E. , Blackmore , L. , Cochrane , R. H. B. , & Allan , K. M. ( 1990 ). Estimating premorbid intelligence by combining NART and demographic variables: An examination of the NART standardisation sample and supplementary equations . Personality and Individual Differences , 11 , 1153 – 1157 . Google Scholar CrossRef Search ADS Duff , K. ( 2010 ). Predicting premorbid memory functioning in older adults . Applied Neuropsychology , 17 , 278 – 282 . Google Scholar CrossRef Search ADS Duff , K. , Chelune , G. J. , & Dennett , K. ( 2011 ). Predicting estimates of premorbid memory functioning: validation in a dementia sample . Archives of Clinical Neuropsychology : the Official Journal of the National Academy of Neuropsychologists , 26 , 701 – 705 . Google Scholar CrossRef Search ADS Gladsjo , J. A. , Heaton , R. K. , Palmer , B. W. , Taylor , M. J. , & Jeste , D. V. ( 1999 ). Use of oral reading to estimate premorbid intellectual and neuropsychological functioning . Journal of the International Neuropsychological Society : JINS , 5 , 247 – 254 . Google Scholar CrossRef Search ADS Glass , L. A. , Bartels , J. M. , & Ryan , J. J. ( 2009 ). WAIS-III FSIQ and GAI in ability-memory discrepancy analysis . Applied Neuropsychology , 16 , 19 – 22 . doi:10.1080/09084280802644102 . Google Scholar CrossRef Search ADS Griffin , S. L. , Mindt , M. R. , Rankin , E. J. , Ritchie , A. J. , & Scott , J. G. ( 2002 ). Estimating premorbid intelligence: Comparison of traditional and contemporary methods across the intelligence continuum . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 17 , 497 – 507 . Google Scholar CrossRef Search ADS Hawkins , K. A. , & Tulsky , D. S. ( 2001 ). The influence of IQ stratification on WAIS-III/WMS-III FSIQ-general memory index discrepancy base-rates in the standardization sample . Journal of the International Neuropsychological Society : JINS , 7 , 875 – 880 . Hilsabeck , R. C. , & Sutker , P. B. ( 2009 ). Using an implicit memory task to estimate premorbid memory . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 24 , 179 – 191 . Google Scholar CrossRef Search ADS Isella , V. , Villa , M. L. , Forapani , E. , Piamarta , F. , Russo , A. , & Appollonio , I. M. ( 2005 ). Ineffectiveness of an Italian NART-equivalent for the estimation of verbal learning ability in normal elderly . Journal of Clinical and Experimental Neuropsychology , 27 , 618 – 623 . Google Scholar CrossRef Search ADS Krull , K. R. , Scott , J. G. , & Sherer , M. ( 1995 ). Estimation of premorbid intelligence from combined performance and demongraphic variables . The Clinical Neuropsychologist , 9 , 83 – 88 . Google Scholar CrossRef Search ADS Lange , R. T. , Schoenberg , M. R. , Chelune , G. J. , Scott , J. G. , & Adams , R. L. ( 2005 ). Development of the WAIS-III general ability index estimate (GAI-E) . The Clinical Neuropsychologist , 19 , 73 – 86 . Google Scholar CrossRef Search ADS Lezak , M. D. , Howieson , D. B. , & Loring , D. W. ( 2004 ). Neuropsychological assessment ( 4th ed. ). New York : Oxford University Press . McFie , J. ( 1975 ). Assessment of organic intellectual impairment . New York : Academic Press . Nadolne , M. J. , Adams , R. L. , Scott , J. G. , Hoffman , R. G. , & Tremont , G. ( 1997 ). Clinical utility of Wechsler Memory Scale-Revised and predicted IQ discrepancies in closed head injury . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 12 , 757 – 762 . Google Scholar CrossRef Search ADS Nelson , H. E. ( 1982 ). National adult reading test (NART): Test manual . Windsor, England : NFER-Nelson . Nelson , H. E. , & O’Connell , A. ( 1978 ). Dementia: The estimation of premorbid intelligence levels using the New Adult Reading Test . Cortex; A Journal Devoted to the Study of the Nervous System and Behavior , 14 , 234 – 244 . Google Scholar CrossRef Search ADS O’Rourke , J. J. , Adams , W. H. , Duff , K. , Byars , J. , Nopoulos , P. , Paulsen , J. S. , et al. . ( 2011 ). Estimating premorbid functioning in Huntington’s disease: The relationship between disease progression and the wide range achievement test reading subtest . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 26 , 59 – 66 . Google Scholar CrossRef Search ADS Petersen , R. C. ( 2004 ). Mild cognitive impairment as a diagnostic entity . Journal of Internal Medicine , 256 , 183 – 194 . Google Scholar CrossRef Search ADS Powell , B. D. , Brossart , D. F. , & Reynolds , C. R. ( 2003 ). Evaluation of the accuracy of two regression-based methods for estimating premorbid IQ . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 18 , 277 – 292 . Google Scholar CrossRef Search ADS Psychological Corporation . ( 2001 ). Wechsler Test of Adult Reading Manual . San Antonio : Author . Randolph , C. ( 2012 ). Repeatable battery for the assessment of neuropsychological status . Bloomington, MN : The Psychological Corporation . Reynolds , C. , & Gutkin , T. B. ( 1979 ). Predicting the premorbid intellectual status of children using demographic data . The Clinical Neuropsychologist , 1 , 36 – 38 . Schoenberg , M. R. , Lange , R. T. , Brickell , T. A. , & Saklofske , D. H. ( 2007 ). Estimating premorbid general cognitive functioning for children and adolescents using the American Wechsler Intelligence Scale for Children-Fourth Edition: Demographic and current performance approaches . Journal of Child Neurology , 22 , 379 – 388 . Google Scholar CrossRef Search ADS Schoenberg , M. R. , Scott , J. G. , Duff , K. , & Adams , R. L. ( 2002 ). Estimation of WAIS-III intelligence from combined performance and demographic variables: Development of the OPIE-3 . The Clinical Neuropsychologist , 16 , 426 – 437 . Google Scholar CrossRef Search ADS Schoenberg , M. R. , Scott , J. G. , Ruwe , W. , Patton , D. , & Adams , R. L. ( 2004 ). Assumptions that underlie predicting premorbid IQ: A comment on the “evaluation of the accuracy of two regression-based methods for estimating premorbid IQ” . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 19 , 1103 – 1106 . Google Scholar CrossRef Search ADS Schretlen , D. J. , Buffington , A. L. , Meyer , S. M. , & Pearlson , G. D. ( 2005 ). The use of word-reading to estimate “premorbid” ability in cognitive domains other than intelligence . Journal of the International Neuropsychological Society : JINS , 11 , 784 – 787 . Google Scholar CrossRef Search ADS Skeel , R. L. , Sitzer , D. , Fogal , T. , Wells , J. , & Johnstone , B. ( 2004 ). Comparison of Predicted-difference, Simple-difference, and Premorbid-estimation methodologies for evaluating IQ and memory score discrepancies . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 19 , 363 – 374 . doi:10.1016/S0887-6177(03)00072-6 . Google Scholar CrossRef Search ADS Vanderploeg , R. D. , & Schinka , J. A. ( 1995 ). Predicting WAIS-R IQ premorbid ability: combining subtest performance and demographic variable predictors . Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists , 10 , 225 – 239 . Google Scholar CrossRef Search ADS Vanderploeg , R. D. , Schinka , J. A. , Baum , K. M. , Tremont , G. , & Mittenberg , W. ( 1998 ). WISC-III premorbid prediction strategies: Demographic and best performance approaches . Psychological Assessment , 10 , 277 – 284 . Google Scholar CrossRef Search ADS Wechsler , D. ( 1958 ). The measurement and appraisal of adult intelligence ( 4th ed. ). Baltimore : Williams & Wilkins . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Archives of Clinical NeuropsychologyOxford University Press

Published: Jun 6, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off