The Utility of the Dementia Severity Rating Scale in Differentiating Mild Cognitive Impairment and Alzheimer’s Disease from Controls

Abstract

The current study investigated the utility of the Dementia Severity Rating Scale (DSRS) total score to identify individuals at the earliest stage of impairment (i.e., Mild Cognitive Impairment/MCI). In addition, the authors sought to investigate how well the measure correlates with an expanded battery of cognitive tests and other measures of functional abilities. Of the 320 participants included in this study, 85 were normal controls, 96 had single or multiple domain amnestic MCI, and 139 had possible or probable Alzheimer’s disease (AD). Each participant underwent a thorough cognitive, neurological, and physical examination. Results from this study indicated that the DSRS total scores differed significantly between the three groups (p < .001) and accurately identified 81% of the control group, 60% of the MCI group, and 78% of the AD group in a post-hoc discriminant analysis. When combined with a brief cognitive measure (i.e., Consortium to Establish a Registry for Alzheimer’s Disease Word List 5-minute recall test), the DSRS accurately identified 98% of the control group, 76% of the MCI group, and 82% of the AD group. Implications for clinical practice and proposed areas of future research are discussed.

All-cause dementia is defined as declines in cognitive impairment severe enough to impair a person’s ability to conduct instrumental activities of daily living (IADLS).¹ Growing consensus in the field acknowledges that there is a phase of mild cognitive impairment (MCI) that may be noticeable to the individual and family members, but do not affect the individual’s ability to conduct their regular activities, which proceeds a formal diagnosis of Alzheimer’s Disease (AD).² Although not all individuals who present with MCI will develop Alzheimer’s Disease (AD), nearly half of all people who visit their primary provider for MCI symptoms will develop dementia within four years.³

Early diagnosis of MCI and dementia is recommended because it allows earlier treatment and management. Primary care providers are thus in a unique position to detect cognitive decline over time as they provide continued care for their patients.⁴ However, studies have shown that 40-70% of physicians in primary care^5,6 fail to diagnose mild-to-moderate cognitive impairment in elderly patients. Primary care physicians may not detect mild cognitive impairment for a number of reasons, including time constraints, insurance under-reimbursement prohibiting a comprehensive evaluation, lack of valid screening tools, as well as the multidimensionality of cognitive impairment and dementia.⁷ There is, therefore, a great need for simple, valid, and brief screening tools to be available to health care professionals in order to make timely and accurate diagnoses more possible.

Instruments that have demonstrated the ability to differentiate between individuals with AD and MCI can be categorized into two groups; 1) cognitive assessments of the patient directly, and 2) informant based questionnaires where a caregiver or other knowledgeable person answers questions regarding the patient. Within the first group, the Mattis Dementia Rating Scale-Second Edition (MDRS-2) has been shown to correctly identify AD (1.0 sensitivity and 1.0 specificity) and amnestic MCI (0.8 sensitivity and 0.68 specificity), but was not specific enough to differentiate between MCI related to Parkinson’s disease and amnestic MCI.⁸ Another patient assessment, the Montreal Cognitive Assessment (MoCA), is a 10-minute cognitive screening tool, which has shown to have a high sensitivity of identifying people with MCI (0.90) and AD (1.0); and good specificity (0.87) for distinguishing between MCI or AD and normal controls (NC).⁹ Individuals with a low MoCA Memory Index Score were also more likely to convert to a diagnosis of AD within an average follow-up time of 18 months.¹⁰

With the group of informant-based questionnaires, the six-item Disability Assessment for Dementia Scale (DAD-6), was able to differentiate between MCI and mild dementia (MD) with a sensitivity of 0.83 (95% confidence interval 0.74–0.92) specificity of 0.84 (95% confidence interval 0.71-0.94). ¹¹ Another informant-based questionnaire, the 39-item Everyday Cognition scale (ECog), was able to significantly discriminate between normal controls and MCI (0.93 sensitivity, 0.80 specificity), and MCI and persons with dementia (0.75 sensitivity, 0.80 specificity).¹²

When evaluating these different scales for use in differentiating between NC, MCI, and AD, along with measuring symptomatic changes over time, a review by Weinstein and colleagues¹³ made the following recommendations: 1) linearity over the range of the scale; 2) ability to detect small changes in observation periods of less than a year; and 3) need for the administration to be brief and able to be completed in a clinic setting. In short, an instrument that easily, reliably, and consistently tracks symptom severity is essential in both the clinical management and research of MCI and AD.¹³ In this regard, the Dementia Severity Rating Scale (DSRS) was recommended as potentially being able to meet these criteria.¹³

The DSRS is an 11-item, informant-report questionnaire, assessing a variety of functional and cognitive abilities. It has a brief administration time (about 5 minutes) and has a number of advantages over other widely-used screening instruments. Clark and Ewbank¹⁴ demonstrated the ability of the DSRS to differentiate cognitively healthy individuals from those with AD in a study of 162 subjects with a diagnosis of normal (n = 24), probable AD (n = 135), and definite AD (n = 3). The authors showed the DSRS to be highly correlated with other widely-used functional measures such as the Washington University Clinical Dementia Rating Scale, and certain cognitive measures such as the language, memory, and praxis tests from the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD).¹⁵ The authors also demonstrated that the DSRS was a valid and useful measure of disease progression¹⁶ as it was able to stage the disease well into severe dementia without the constraint of a floor effect.

The DSRS also has been shown to be effective in detecting declines in function in a study of 702 patients with AD followed participants from first diagnosis until they became too impaired to return to clinic.¹⁷ The DSRS not only was able to detect gradual declines in ability in test-retest every six months, but the rate of decline remained steady throughout the disease progression, increasing an average of 4.48 points (95% confidence interval: 4.14-4.82). In contrast, the Mini Mental State scores also showed declines, but at varying rates during the disease progression, thus was not shown to be the ideal measure to track changes in symptom functioning.¹⁷

Despite the many strengths of the DSRS as a useful screening tool to differentiate normal adults from those with dementia, and track rates of decline in AD symptoms over time, its sensitivity in identifying subtle changes in cognition in individuals with MCI and mild dementia remains unexplored. The utility of the DSRS to detect early dementia is of particular interest, especially given that this early period is where treatment is most effective.¹⁸ The current study seeks to determine if the DSRS total score (a) can identify individuals at the MCI stage of impairment from those who are cognitively healthy (i.e., NC), (b) correlates with an expanded battery of cognitive tests, and (c) correlates with measures of activities of daily living (e.g., Pfeffer Functional Activities Questionnaire, Bristol Activities of Daily Living Scale).

Method

Participants

Study participants consisted of 362 community-dwelling older adults evaluated at the University of California in Irvine, Alzheimer’s Disease Research Center between January 2002 and June 2006. Normal control participants (n = 109) were volunteers enrolled in the “Successful Aging Program” and had no significant cognitive complaints. Participants (n = 24) were excluded from this group when they demonstrated either cognitive and/or functional impairments at their initial or subsequent assessments. The remaining 85 control participants showed no significant cognitive, neurological, functional, or behavioral deficits.

Participants in the MCI group (n = 107) had a subjective memory complaint and met criteria for either the single- or multiple-domain forms of amnestic MCI.¹⁹ Subjects scored at or below 1.5 standard deviations from the mean for their age on the 30-minute delayed recall measure of either the WMS-III Logical Memory test²⁰ or the CERAD Word List.¹⁵ MCI participants (n = 11) were excluded when they showed deficits in their basic activities of daily living²¹ at initial or subsequent visits.

Finally, the 146 participants in the AD group were evaluated at the clinic and diagnosed as having either probable or possible Alzheimer’s disease per the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association criteria.²² AD patients (n = 7) were removed from this group for having a change in their diagnosis on a subsequent annual visit.

Since problems with depression are fairly common in both healthy and cognitively impaired older adults and can negatively impact functional abilities, participants showing evidence of a major depressive disorder, as defined as scoring 8 or above on the Geriatric Depression Scale (GDS) ²³ were excluded. Patients with mild depression, however, were included in the analysis as excluding these individuals would have resulted in highly selective and non-representative groups of MCI and AD patients.

Of the 320 qualifying participants included in the data analyses, 85 (26.6%) were normal controls, 96 (30%) had either Amnestic MCI (n = 50) or Amnestic MCI with multiple domains (n = 46), and the remaining 139 (43.4%) had possible (n = 28) or probable (n = 111) AD. This sample was comprised of approximately equal numbers of men (n = 157) and women (n = 163). Mean scores on the Folstein Mini-Mental State Examination (MMSE) for the control and MCI participants fell in the unimpaired range and were 29.2 (SD = 1.1) and 26.7 (SD = 2.3), respectively. In comparison, the AD patients showed mild to moderately severe dementia with an average MMSE score of 19.3 (SD = 6.0). The mean ages of the control, MCI, and AD patients were 73.6 (SD = 9.4), 75.8 (SD = 7.2), and 77.5 (SD = 8.3) years, respectively. Participants in all three groups were well educated, with a mean number of years of schooling of 16 (SD = 2.2), 15.7 (SD = 2.7), and 14.4 (SD = 3.1) in the control, MCI, and AD groups, respectively.

Across the three groups, the informants providing information about the participants’ functional abilities were primarily spouses (53.5%) or adult children (33.7%). For a small number of participants, the primary informants were other relatives (1.7%), friends and/or neighbors (6.7%), or paid caregivers (0.7%). A few (n = 9) of the control participants served as their own informant in the completion of the functional measures. Given that healthy older adults have been shown to provide reliable self-reports,²⁴ these individuals’ reports were deemed valid and therefore included in the present study. None of the participants with MCI or AD served as their own informant. A majority of the informants (i.e., 92%) reported having known the participants for more than 10 years and saw them anywhere from several times a week (11.3%) to one or more times per day (66.3%). All clinical diagnoses were reached by consensus of a board-certified neurologist and licensed clinical neuropsychologist, following clinical interview and review of data. The primary outcome measures in this study were not utilized in making diagnosis determinations.

Procedures

All participants completed a two-hour neuropsychological test battery that included measures of global mental status, recent verbal memory, recent visual memory, remote (i.e., long-term) memory, attention, language, visual-spatial ability, executive functioning, and psychomotor speed. In addition, a family member, spouse, or knowledgeable informant was asked to complete questionnaires about the patient’s daily functional abilities, including the Dementia Severity Rating Scale (DSRS),¹⁴ the Pfeffer Functional Activities Questionnaire (FAQ),²⁵ and the Bristol Activities of Daily Living Scale (BADLS).²⁶ The DSRS was administered for the purposes of this study and was not used in making the clinical diagnosis so it can be used to evaluate its ability to measure group differences.

The DSRS¹⁴ is an informant-based questionnaire that provides clinicians with a brief means of assessing the severity of a patient’s dementia. Informants are asked to rate the participant’s ability in 11 areas of functioning including ADLs, IADLs, and cognitive functioning. Each category is assessed using descriptions extending from normal (i.e., independent) to severely impaired (i.e., dependent) levels of functioning. Total scores range from 0 to 51 points, with lower scores representing higher levels of functioning.

Data Analyses

Data were analyzed using the statistical software, SYSTAT- Version 11.0.²⁷ Gender differences between groups were examined using a chi-square analysis while age, education, global mental status, and functional status were compared across the three groups using an Analysis of Variance (ANOVA) test. With these tests, a Bonferroni correction was applied to account for multiple comparisons. To examine group differences on the DSRS, an ANOVA with Bonferroni corrected probabilities was conducted. Partial correlations with Bonferroni corrections were obtained to examine the relationship between the DSRS and the FAQ, BADLS, and neuropsychological measures controlling for significant demographic variables. Test scores were transformed (e.g., square roots, log) in order to make the distributions more symmetrical where appropriate.

Results

Data from 320 normal control and cognitively-impaired community dwelling older adults were analyzed. Chi-square analyses indicated a significant difference in the number of men (n = 157) and women (n = 163), χ²(1, N = 320) = 11.4, p < .01. As shown in Table 1, results indicated the AD participants were significantly older, F(2, 317) = 5.7, p < .01, and had slightly lower education, F(2, 317) = 11.3, p < .001, than the NC and MCI participants. In addition, the control group reported fewer depressive symptoms, F(2, 300) = 10.0, p < .001. As expected, the groups showed significant differences in global mental status (i.e., MMSE), F(2, 315) = 171, p < .001, and functional abilities on the DSRS, (F(2, 317) = 294.9, p < .001)

Table 1.

Demographics by Sample and Diagnostic Group

	Sample		NC		MCI		AD		Chi-Square
N	320		85		96		139
% Female	50.9		64.7		39.6		50.4		11.4*
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	F	t
Age	75.9	8.5	73.6	9.4	75.8	7.2	77.5	8.3	5.7*	AD < NC, MCI
Education	15.2	2.9	16.0	2.2	15.7	2.7	14.4	3.1	11.3**	AD < NC, MCI
MMSE	24.2	6.0	29.2	1.1	26.7	2.3	19.3	6.0	241.0**	AD < MCI < NC
DSRS	9.9	8.3	2.0	2.3	7.1	4.0	16.7	7.5	294.9**	AD < MCI < NC
FAQ	8.5	9.3	0.7	1.6	4.0	4.1	16.6	8.5	306.2**	AD < MCI < NC
BADLS	5.4	8.1	0.5	1.5	1.6	2.6	11.5	9.3	159.2**	AD < MCI < NC
GDS	2.1	2.4	1.3	1.8	2.3	2.5	2.5	2.5	10.0**	NC < MCI, AD

Note. NC = Normal Controls; MCI = Mild Cognitive Impairment; AD = Alzheimer’s Disease; MMSE = Folstein Mini-Mental State Examination; DSRS = Dementia Severity Rating Scale; FAQ = Pfeffer Functional Abilities Questionnaire; BADLS = Bristol Activities of Daily Living Scale; GDS = Geriatric Depression Scale

^*p < .01

^**p < .001

Differences in Functional Status between Groups

It was hypothesized that the DSRS total score would significantly differ between groups of NC, MCI, and AD participants. Because the asymmetrical nature of this sample distribution violated the assumptions of the parametric procedures, a square-root transformation was conducted on the DSRS total score prior to analysis. An ANOVA with Bonferroni corrected probabilities indicated significant differences in functional abilities among the groups, F(2, 316) = 292.54, p < .001. A post-hoc analysis was conducted and revealed a significant difference between all three groups (p < .001). More specifically, the AD group exhibited significantly greater functional deficits on the DSRS than both the control, t(2) = 2.90, p < .001, and the MCI group, t(2) = 1.43, p < .001. Likewise, the MCI group exhibited significantly greater impairment than the control group, t(2) = 1.46, p < .001.

To further examine the relevance of this finding, a discriminant analysis was conducted to evaluate how well subjects could be classified (diagnostically) by their DSRS total score. A jackknifed classification matrix indicated that the total score of the DSRS accurately identified 81% of the control group, 60% of the MCI group, and 78% of the AD group. Overall, the DSRS accurately classified 72% of the participants thereby misclassifying approximately 1/4 of the total sample.

DSRS and Neuropsychological Performance

It was hypothesized that the DSRS would correlate significantly to measures of memory, language, praxis, and global mental status, visual-spatial ability, and executive functioning. Gender, age, and education were controlled for in the analyses due to a significant relationship to the DSRS total score. As a result, partial correlations with Bonferroni corrected probabilities were used to examine the relationship between the DSRS total score and the neuropsychological measures.

The specific neuropsychological instruments included in this analysis can be seen in Table 2. For the Trail Making Tests (i.e., TMT-A and TMT-B), 19 scores from TMT-A and 95 scores from TMT-B were excluded from the analyses because of imputed (i.e., floored) values. Partial correlations for the entire sample indicated that the DSRS total score was significantly related to all 15 neuropsychological instruments. In an effort to examine the relationship within each diagnostic group, partial correlations were obtained for each of the three groups. Results from this analysis indicate that the relationship of the DSRS total score to neuropsychological performance was primarily due to the strength of the relationship found in the AD group. As shown in Table 2, the DSRS total score was (a) significantly related to Logical Memory 1 (a measure of verbal memory) in the control group (p < .001), (b) not significantly related to any of the cognitive measures in the MCI group, and (c) significantly related to most of the cognitive measures in the AD group (p < .001), with the exception of the CWLT-5, CWLT-30, WMS-III LM2, and TMT-B.

Table 2.

Partial Correlations between DSRS and Neuropsychological Measures by Sample and Diagnostic Group with Bonferroni Corrections

	Sample		NC		MCI		AD
Instrument	N	Coef	n	Coef	n	Coef	n	Coef
Memory
CERAD Word List- 5 min. recall	311	-.73**	81	-.19	93	-.07	137	-.23
WMS-III Logical Memory 1	317	-.74**	84	-.39**	96	-.19	137	-.32**
WMS-III Logical Memory 2	317	-.72**	84	-.32	96	-.17	137	-.10
Attention
WAIS-III Digit Span	267	-.44**	46	-.02	86	.01	135	-.41**
WAIS-III Symbol Digit Modality	261	-.63**	46	.22	86	-.14	129	-.44**
Language
Boston Naming Test	317	-.65**	84	-.22	96	.03	137	-.39**
COWAT	317	-.57**	84	-.03	96	-.11	137	-.45**
Category Fluency	317	-.66**	84	-.04	96	-.07	137	-.42**
Visual-Spatial
Constructional Praxis	317	-.45**	84	-.19	96	.00	137	-.21**
Block Design	309	-.56**	84	-.08	96	-.04	129	-.39**
Executive Functioning
Trail Making Test- A	299	.52**	85	-.06	96	.01	118	.31*
Trail Making Test- B	221	.48**	85	.05	82	.03	54	.02
WAIS-III Similarities	315	-.62**	84	-.16	96	-.03	135	-.39**
Global Mental Status
Folstein Mini-Mental State Examination	317	-.69**	84	.05	96	.04	137	-.47**

Note. NC = Normal Controls; MCI = Mild Cognitive Impairment; AD = Alzheimer’s Disease; CERAD = Consortium to Establish a Registry for Alzheimer’s Disease; WMS-III = Wechsler Memory Scale- 3^rd Edition; WAIS-III = Wechsler AdultIntelligence Scale - 3^rd Edition; COWAT = Controlled Oral Word Association Test

^*p < .01

^**p < .001

DSRS and Activities of Daily Living

In the present study, the DSRS total score was hypothesized to be significantly related to two other widely-used measures of functional status: the FAQ and the BADLS. Partial correlations were obtained and indicated that the DSRS total score was significantly related to the total scores for both the FAQ and the BADLS (p < .001). In an effort to examine the relationships between these measures within groups, partial correlations were obtained for each group controlling for gender, age, and education. As shown in Table 3, the DSRS total score was significantly related to both the FAQ and the BADLS in all three groups (p < .001).

Table 3.

Partial Correlations between DSRS and other Functional Measures by Sample and Diagnostic Group with Bonferroni Corrections

	Sample		NC		MCI		AD
Instrument	N	Coef	n	Coef	n	Coef	n	Coef
FAQ	311	.89*	83	.61*	95	.77*	133	.80*
BADLS	279	.80*	76	.34*	88	.61*	115	.81*

Note. NC = Normal Controls; MCI = Mild Cognitive Impairment; AD = Alzheimer’s Disease; FAQ = Pfeffer Functional Activities Questionnaire; BADLS = Bristol Activities of Daily Living Scale

^*p < .001

Post-Hoc Analyses

In light of previous studies reporting the benefit of combining cognitive measures with functional scales to increase diagnostic accuracy,^28,29 previous measures shown to correlate with the DSRS¹⁴ were examined in a discriminant analysis. These measures included several subtests from the CERAD Word List memory test (Note that the 30-minute delay was not used since it had potentially been used to assist in diagnosis; see Table 4). When examined alone, results indicated that the 5-minute recall test (CWLT-5 performed comparably or better than the other three CERAD measures at correctly classifying diagnostic category. The CWLT- 5 was used in conjunction with the DSRS in a discriminant analysis. Results from this analysis indicated that the combined measure was more accurate than both the DSRS and CWLT- 5 used alone. Specifically, the jackknifed classification matrix indicated the total score of the combined CWLT-5 and DSRS accurately identified 98% of the control group, 76% of the MCI group, and 82% of the AD group. For the overall sample, the CWLT-5/ DSRS combination accurately classified 85% of the participants. Results for all models are shown in Table 4.

Table 4.

Discriminant Analysis Results for all Models

	NC	MCI	AD	TOTAL
Instrument	% Correct	% Correct	% Correct	% Correct
DSRS	81	60	78	72
FAQ	82	47	85	73
BADLS	86	36	81	68
CWLT-5	96	62	79	79
CWLT- 3rd Trial	90	41	81	71
CWLT- Trials 1-3	94	57	72	73
CWLT- Composite	95	66	78	79
CWLT-5 + DSRS	98	76	82	85

Note. NC = Normal Controls; MCI = Mild Cognitive Impairment; AD = Alzheimer’s Disease; FAQ = Pfeffer Functional Activities Questionnaire; BADLS = Bristol Activities of Daily Living Scale; CWLT-5 = CERAD Word List 5-minute recall test; CWLT- Composite = Total composite score of 5- minute recall, 1^st trial, 2^nd trial, and 3^rd trial

Discussion

The purpose of the current study was to address the utility of the DSRS toward better characterizing and distinguishing individuals diagnosed with MCI and AD from those without cognitive impairment. Results indicated that the DSRS would serve as a useful screening tool, able to successfully distinguish NC subjects from both MCI and AD groups. These findings support Clark and Ewbank’s¹⁴ original research on the usefulness of the DSRS with groups of normal and demented (i.e., AD) older adults. Identifying cognitive impairment in its earliest stages (i.e., MCI) is critical, as early diagnosis and treatment have been shown to help slow or stabilize cognitive decline, reduce caregiver stress, and delay institutionalization.¹⁸

The relationship between the DSRS and neuropsychological performance was also examined in the current study. This study expanded upon the Clark and Ewbank¹⁴ study by adding measures of attention and executive functioning to the memory, language, and visual-spatial scales used in the original study. In addition, the current study examined within-group performance which included a MCI group. With the AD group, the majority of the correlations were significant. This is consistent with previous findings that demonstrate the strength of the relationship between cognitive status and functional status for AD patients.³⁰ For the MCI group, however, none of the neuropsychological measures correlated significantly with the DSRS. There was, however, a strong trend between measures of memory and DSRS total score. This indicates that, while impairment in memory is co-occurring with impairment in functional status, it is not occurring at the same rate for all participants as is more characteristic of the AD group.

One possible explanation for this is that individuals with a MCI diagnosis are not expected to be as severely impaired in as many domains as those diagnosed with AD. As a result, MCI patients may possess the ability to compensate for loss in a given cognitive domain whereas an AD patient may have multiple domains significantly impaired. For example, a patient with severe memory impairment may compensate using their executive skills by setting alarms, taking notes, and taking other creative measures to counteract the memory loss. The AD patient may lack the initiative, creativity, and executive control to compensate for the memory loss. Therefore, the AD patient will have significant cognitive impairment adversely affecting everyday living while the MCI will present with less impairment in IADLs.

Researchers²⁸ have demonstrated that combining a brief test of episodic memory (i.e., Word List recall) with a functional measure proved to be a discriminant combination in separating individuals with MCI from those in the earliest stages of AD. Likewise, the current study combined the DSRS with the CWLT- 5. Results indicated that using the two measures in conjunction as a brief screening tool correctly enhanced diagnostic accuracy, identifying 98% of the controls, 76% of the MCI group, and 82% of the AD group (see Graph 1 for comparisons).

Graph 1.

Diagnostic predictability of the DSRS and DSRS plus CWLT-5.

Note. NC = Normal Controls; MCI = Mild Cognitive Impairment; AD = Alzheimer’s Disease; CWLT-5 = CERAD Word List 5-minute recall test

The DSRS was also highly related to both the FAQ and the BADLS, demonstrating good construct reliability. Although other measures have been shown to also be able to differentiate between AD and MCI, the DSRS has a number of advantages including its brevity, comprehensiveness, sensitivity to early changes in cognition and functional impairment, and ability to track disease progression over time.¹⁷

Limitations and Future Directions

One finding that warrants further discussion is that the AD group was found to have more symptoms of depression than the NC or MCI groups. Since dementia is associated with a higher risk for depression,³¹ occurring in 11-54% of individuals with MCI^32,33 and 30-50% of early AD patients, ^34,35 this finding is not unanticipated and is most likely representative of the AD population. Additionally, while the mean GDS score of the AD patients (M=2.5, SD=2.5) was higher than that of the MCI patients (M=2.3, SD=2.5) and NC (M=1.3, SD=1.8), all three averages were quite low and fell below the generally-accepted cutoff for mild depression of 5 out of 15 points (see Table 1). As reported earlier, all individuals in the NC, MCI, and AD groups with significant levels of depression were excluded from the study. Future studies may wish to look at whether higher levels of depression negatively impact the ability of the DSRS to distinguish between AD, MCI and normal aging.

As with any study, it is important that the findings be replicated with other samples to further ascertain the generalizability of the results. While the age and gender demographics were similar to other studies, the level of education for the participants in this study exceeded both state and national averages. That is, 96.6% of the participants in this study had at least 12 years of education and 56.9% had at least 16 years of education. This is relatively high when compared to the national averages of the time which indicated that 73.1% of individuals age 65 and older had 12 or more years 18.7% had 16 or more years of education,³⁶ though the educational levels in the study were consistent with those found in the sample location (in Orange County, California, 90.1% of adults over age 65 had at least 12 years and 46.8% had at least 16 years of education).³⁷

This sample was additionally limited to individuals fluent in the English language resulting in a relatively homogenous ethnic sample. The lack of multiculturally and multilinguistically sensitive neuropsychological instruments presents a significant deficit in the field at large as opposed to being a limitation unique to this study.

It might also be useful to examine a more heterogeneous MCI diagnostic group with both amnestic and non-amnestic participants, and/or solely examine the non-amnestic subtype. This does not take away from the contribution of this study, however, given that the majority of people with an MCI diagnosis have memory impairment.³⁸ Given that both the MCI and AD groups in this study shared memory impairment, incorporating a non-amnestic diagnostic group may result in a change in some of the data. For example, adding a non-amnestic MCI group could possibly increase the overlap between NC and MCI groups on memory tasks, and potentially decrease the overlap between the MCI and AD groups, since non-amnestic MCI patients have been shown to convert at a lower rate than the amnestic subtype.³⁹

In conclusion, the brief and easy-to-administer DSRS has demonstrated a high level of accuracy in differentiating between older adults with NC, MCI, or mild AD. Since the DSRS takes only five minutes to administer, it could easily be incorporated into annual exams for older adults. . Even if there is no cognitive impairment present during the first visit, results from that testing would provide an invaluable baseline for changes in memory and functioning over time. In addition, routine evaluations of the DSRS provide a useful measure of disease progression which can aid in treatment and diagnostic decisions.¹⁷ This is particularly important given the demand for ways to quickly and effectively identify cognitive impairment among older adults in the primary care setting.