Examination of discrepancies between subjective and objective memory with latent factors of the geriatric depression scale

Abstract

Objective: The purpose of the present study was to examine whether multiple constructs of depressive symptoms explained discrepancies between subjective (SM) and objective memory (OM) in older adulthood.

Method: A sub-sample (Range: 65–98 years, N = 606) of the 2000 Wave Panel 3 of the Long Beach Longitudinal Study was used to explore underlying latent factors within the Geriatric Depression Scale (GDS).

Results: Four GDS factors were identified (Dysphoria, Vigor/Withdrawal, Cognitive Concerns, and Agitation); the factor solution was tested in a confirmatory factor analysis. Mechanisms of the GDS factors on SM and OM were tested by using structural equation modeling. Dysphoria negatively predicted OM, but not SM, suggesting that people who scored high in Dysphoria might not sense their memory failures although they obviously performed poorly in OM. Cognitive concerns negatively predicted SM, but not OM, suggesting that cognitive concerns were not influential for OM, but negatively affected one's evaluation of SM.

Conclusion: The study results have clinical implications since such relationships can be concealed while assessing depressive symptoms as a single GDS factor. Thus, the patterns of the GDS factors, SM, and OM derived from this study can be very useful for clinicians who conduct assessments on depressive symptoms and memory failures.

Keywords:

• Depressive symptoms
• factor analysis
• structural equation modeling
• subjective memory
• objective memory

View correction statement:

Corrigendum

Disclosure statement

We have no disclosures to acknowledge.

Notes

1. Among the late adulthood group, 52 people (9.2%) reported they always had difficulty walking upstairs, 39 people (6.9%) reported that they almost always need help walking, and 51 people (8.4%) reported that they suffered from stroke(s).

2. Forty-nine people (8.1%) reported they had psychiatric problems.

3. A series of one-way ANCOVAs were performed to examine whether those physical and psychiatric disorders impacted OM performance. Equality of variances for the compared groups was examined utilizing Levene's test. Persons that endorsed always needing help walking showed slightly lower scores on immediate recall I (p < .5; Partial Eta Squared = .016) compared to those without the need for constant help with walking; this effect was found while controlling for age, education, and depression (assessed by the GDS total score). Overall, the impact of physical and psychiatric disorders on OM performance was negligible.

4. Prior research showed the associations between physical health conditions (i.e. hypertension, diabetes, and heart problems) and OM performance. Most of people in the sample with those health conditions reported taking medication. Among the current sample, the number of people with hypertension was 298 (49.2%). Of these individuals, only 46 did not take hypertension medication. In our sample, 63 people (10.4%) had diabetes, and of those, 17 people did not take diabetes medication. Finally, 171 people (28.2%) had heart problems, and of those, only 47 people did not take medication for their heart problems. The effects of these health conditions on OM were tested. A series of one-way ANCOVAs were performed to examine whether physical health conditions (high blood pressure, diabetes, and heart problems) impacted OM performance. Equality of variances for the compared groups was examined by Levene's test. Persons with high blood pressure showed slightly lower scores in Text Recall II (p < .5; Partial Eta Squared=.008) than those without high blood pressure while controlling for age, education, and depression (assessed by the GDS total scores). Overall, the effects of these health conditions on OM performance were negligible.

5. A one-way ANCOVA was performed to examine whether depressive symptoms (i.e. 10 ≤ on total GDS scores ≤ 30) impacted OM performance. Equality of variances for the compared groups was examined utilizing Levene's test. There was no mean difference in OM performance between the groups when controlling for age and education.

6. The cut-off values of the GDS were not used for our analyses in this current study.

7. The GDS factor structure was explored in the middle adulthood group (age range: 40–64 years) and the late adulthood group (age range: 65–98 years old). No factor structure was identified in the middle adulthood group, while the four GDS factors were extracted in the late adulthood group. Analyses are available upon request.

8. The LBLS data does not include a measure of anxiety. Yet, of the GDS factors, the items assessing Agitation are very relevant in regards to symptoms of anxiety (e.g. ‘Are you afraid that something bad is going to happen’, ‘Do you frequently worry about the future?’, and ‘Do you worry a lot about the past?’). As shown in Figure 3, Agitation had no impact on SM and OM while controlling for the effects of the other GDS factors (Dysphoria, Vigor-Withdrawal, and Cognitive Concerns) on SM and OM.