This study had many strengths. First, the tri-ethnic sample was recruited in the community, and volunteer participants were highly motivated and maintained their interest in the study, as evidenced by a 79% retention rate over the 26-month study. The ACTIVE trials reported 80% retention at 24-month follow-up (Ball et al., 2002
). Further, the range of outcomes was comprehensive and administered at each measurement occasion. Finally, we explored differences in racial and ethnic group responses to the memory training. Study limitations included the lack of a no-treatment control group. Further, the generalizability of the findings from the minority sample is not clear.
The participants in both the memory and health promotion groups improved their use of external and internal memory strategies. This is a unique finding for an intervention that did not place any emphasis on memory strategies in the health training (Lachman et al., 1992
; Lachman, Andreoletti, & Pearman, 2006
). However, both groups received mental stimulation and social engagement, in part that may explain the benefits to memory. Participants may have engaged in self-study to improve memory strategies as a result of their participation in the study. Health promotion training is not a normative intervention in people’s daily lives. When this study was launched in 2001, evidence of the effectiveness of memory training with older adults was known (Floyd & Scogin, 1997
; Verhaeghen et al., 1992
), but, the benefits of health promotion as mental stimulation were less well known. Nevertheless, the health promotion group’s maintenance of cognitive functioning is consistent with growing scientific evidence that health promotion interventions, including physical activity, may prevent decline (Abbott, White, Ross, Masaki, Curb, & Petrovitch, 2004
; Boron, Willis, & Schaie, 2007
; Salthouse, 2006
; Salthouse, Berish, & Miles, 2002
; Verghese et al., 2003
; Weuve et al., 2004
In the analyses, race and ethnicity were covariates, which suggests that memory performance may change differently over time based on demographic characteristics. Both Hispanics and Blacks tended to make greater gains than Whites on visual memory; Blacks also performed better over time on instrumental functional abilities, though Hispanics performed worse on the memory measure in the short run. These results raise the question: Are minorities more likely to gain from health promotion or memory training interventions? Because pretest scores were included in the model as covariates, the differential changes over time for Black and Hispanic participants as compared to Whites were not due to lower pretest scores.
The participants improved on the DAFS performance measure from baseline to end of classes, but there was no significant change in scores at the end of the study. Transfer of learning or practice effects may have occurred as a result of taking the DAFS repeatedly. Even though our participants were screened out for dementia, we found that the DAFS (even excluding the eating and dressing/grooming skills subscales) had a non-normal distribution and a ceiling effect despite the fact that two-thirds of these same respondents scored in the poor or impaired ranges of a test of everyday memory performance. This pattern of results suggested that the DAFS might not be sensitive to the early deficits in functioning exhibited by older adults who are experiencing poor memory performance but are still living independently in the community.
Also, our study included four required 90-minute booster sessions at 3 months post-training for all participants. These additional sessions may have increased learning and retention through practice effects with everyday activities. The four booster sessions for the health promotion group may have increased the mental stimulation for this group (Austin-Wells, McDougall, & Becker, 2006
Less educated, Black, and Hispanic participants had lower scores than Whites on verbal memory at baseline. In our HLM analyses, we examined the verbal memory delayed recall T-score as an outcome at Times 2 and 5. Controlling for ethnicity, age, education, and verbal memory scores at Time 1, there were no treatment effects. However, testing with Hispanic elders was often done over many days to accommodate their low literacy skills and level of frustration with the complex battery of measures. Practice effects may have been greater for Blacks and Hispanics, or the interventions may have been more effective for minority participants (McDougall et al., 2008
). Since different patterns were observed for Blacks and Hispanics, future research should examine ethnic/racial groups separately and consider the moderator effects of race and ethnicity on memory training.
Following the training, we observed an improvement in global cognition; however, there was no change from baseline to end of study. The initial improvement may reflect a practice effect induced by repetition because the period between baseline testing and post-class was approximately 8 weeks. The MMSE has a documented selective bias toward episodic memory and visuospatial skills in normal elderly adults without dementia (Hill & Backman, 1995
). Also, in a longitudinal analysis, Rabbitt, Diggle, Holland, and McInnes (2004)
found that practice improvements were greatest between the first and second experiences with a cognitive task.
The memory group showed no greater gains in everyday memory performance, than the health promotion group. These findings contrast with the ACTIVE trial findings, in which the participants in each intervention group improved targeted cognitive abilities by an amount approximately equal to the cognitive decline that would naturally occur in older adults without dementia, and maintained these gains for 5 years (Ball et al., 2002
; Willis et al., 2006
The participants in our study did not have dementia. However, 169 (64%) individuals had measurable problems in one or more domains of everyday function, and 46 (17%) had problems that would be of serious enough concern for further neuropsychological referral and would potentially meet the criteria for mild cognitive impairment (MCI). The MCI prevalence rate in a nationally representative sample was 22% (Plassman et al., 2008
). In recent studies, individuals with MCI might have received differential benefits from cognitive training (Belleville, 2008
; Unverzagt et al., 2007
). The issue of memory complaints deserves further discussion because these subjective concerns are often predictive of a future diagnosis of MCI (Pearman & Storandt, 2004
; Purser, Fillenbaum, & Wallace, 2006
; Winblad et al., 2004
; Zandi, 2004
; Zeintl et al., 2006
). We found a reduction in memory complaints for participants in the memory group. At 26 months, they continued to report fewer complaints.
Whites made relatively greater gains than Blacks did in memory self-efficacy. We expected memory self-efficacy to improve based on evidence from previous studies with diverse older adults (Dittmann-Kohli et al., 1990; Lachman et al., 1992
; McDougall, 1998
; Rebok, & Balcerak, 1989
; West et al., 2008
). However, all previous intervention studies used different memory efficacy questionnaires. Difficulty in comprehending the MSEQ was reflected in our findings. Future studies will require further adjustment of this measure, and possibly other measures, to accommodate the literacy level of participants (Institute of Medicine, 2004
; Scott et al., 2002).
In a recent study, Algase, Souder, Roberts, & Beattie (2006)
elevated the urgency to prepare advanced practice geropsychiatric nurses to care for the projected 15 million older adults by 2030 with major psychiatric illness. The Cognitive Behavioral Model of Everyday Memory provides a framework for intervening with the everyday memory concerns of older adults striving for independence. Nevertheless, the findings that Black and Hispanic elderly made greater gains than Whites on some memory performance measures merits further study.