Pica (disorder)

Normal elderly and mildly aphasie individuals may exhibit similar impairments in comprehension and expression. The discriminative validity between normal and aphasie performance on most standardized measures of aphasia has not been reported. The authors compared the performance of 18 aphasie and 18 normal adults to determine the discriminative validity of 2 general language measures-the Porch Index of Communicative Ability (B. E. Porch, 1967) and the Western Aphasia Battery (A. Kertesz, 1982)-and 2 functional communication measures-the Communication Activities of Daily Living-Second Edition (A. L. Holland, C. Frattali, & D. Fromm, 1999) and the American Speech-Language-Hearing Association's Functional Assessment of Communication Skills for Adults (C. Frattali, C. K. Thompson, A. L. Holland, C. B. Wohl, & M. K. Ferketic, 1995). all between-groups comparisons of summary scores for each measure showed significant mean differences. Expressive language ability and efficiency of performance best differentiated between the aphasie and normal groups. However, group performance ranges overlapped by at least 10% on each measure. To enhance the differential diagnosis of aphasia, supplementing formal test results with additional subjective and objective evidence is recommended.

Key Words: aphasia, diagnosis, language impairment, activity limitation, validity

Normal elderly and mildly aphasie individuals may exhibit similar impairments in comprehension and expression (Chapman & Ulatowska, 1991, 1994). Differentiating normal from aphasie performance is essential for providing a diagnosis and, subsequently, for providing a prognosis, focusing treatment, and justifying reimbursement for services (Biddle, Watson, Hooper, Lohr, & Sutton, 2002). Systematic and comprehensive collection of data permits comparison of behavioral profiles between normal and aphasie populations. Formal appraisal, using standardized tests of general language or functional communication, facilitates valid, reliable sampling of behaviors relevant to aphasia while minimizing the influence of extraneous factors (Davis, 2000). However, summary scores derived from these measures may not distinguish well between normal and mildly aphasie behavior. Supplemental indices of performance (e.g., index of determination and degree of overlap) may allow a more accurate distinction between aphasie and normal performance and thus improve the clinical usefulness of formal tests.

Both normal elderly and aphasie individuals may show difficulty comprehending complex syntax and inference in paragraph-length contexts (Chapman & Ulatowska, 1991 ; Cohen, 1979; Feier & Gerstman, 1980; Obier & Albert, 1984; Peach, 1987). Similarly, naming deficits common in aphasia-lexical access, object naming, naming latency, and word fluency-may also result from increasing age (respectively, Chapman & Ulatowska, 1991; Bowles & Poon, 1985;Obler& Albert, 1981; Schow, Christensen, Hutchinson, & Nerbonne, 1978). Both populations may also demonstrate changes in the flow of speech, manifested as hyper- or hypofluency (Chapman & Ulatowska, 1991 ; Critchley, 1984; Olber & Albert, 1981). And, both populations may show reduced content of speech, characterized by increased use of indefinite terms, reduction in verb forms, decreased syntactic complexity, and diminished linguistic clarity (Chapman & Ulatowska, 1991, 1994; Kemper, 1988; North, Ulatowska, Macaluso-Haynes, & Bell, 1986; Obler & Albert, 1984).

Similar impairments may be precipitated or maintained by different processes (Chapey, 1994). For example, the primary language comprehension deficit in aphasie individuals is thought to be linguistic, whereas the comprehension deficits in elderly populations may be caused by cognitive impairments in attention, memory, and reasoning (Chapman & Ulatowska, 1991; Darley, 1979). Moreover, factors coexist. In older aphasie individuals, deficits may result from normal aging processes, from aphasia, or from a combination of normal and pathological processes. A primary goal of assessment is to identify which factors disrupt language and communication to determine if they can be reduced or eliminated (Chapey, 1994). Thus, correct differential diagnosis-normal versus abnormal-is essential to stating a prognosis and, if appropriate, focusing treatment.

Comparison between normal elderly and aphasie performance on language and communication tasks is crucial to the differentiation of the confounding effects of aging (Chapman & Ulatowska, 1991). Formal appraisal, using standardized tests of general language or functional communication, facilitates systematic, comprehensive, consistent sampling of behaviors relevant to differential diagnosis of aphasia while minimizing the influence of extraneous factors such as educational level, cultural variation, and reasoning skill (Davis, 2000). The Porch Index of Communicative Ability (PICA; Porch, 1967) and the Western Aphasia Battery (WAB; Kertesz, 1982) are standardized measures recommended by the Agency for Healthcare Research and Quality for formal appraisal of general language (Biddle et al., 2002). The Communication Activities of Daily Living-Second Edition (CADL-2; Holland, Frattali, & Fromm, 1999) and the American Speech-Language-Hearing Association's Functional Assessment of Communication Skills for Adults (ASHA FACS; Frattali, Thompson, Holland, Wohl, & Ferketic, 1995) are standardized measures more recently developed for formal appraisal of functional communication.

Although these measures have been shown to be valid for identifying behaviors common in aphasia, their summary scores may not distinguish well between normal and mildly aphasie behavior. First, adequate between-group comparisons of summary scores have not been reported. Porch (1967) recommended the PICA for comparison of abilities among groups. Yet, only one heterogeneous group of brain-injured participants was included in the test's standardization. Kertesz (1979) included both aphasie and normal groups in the first standardization of the WAB Aphasia Quotient (AQ). However, individuals in the "normal" group suffered from "spinal cord disease, peripheral neuropathy, blackouts, tics, vertigo, ataxia, etc." (Kertesz, 1979, p. 56). Thus, the influence of central nervous system damage on performance in Kertesz's "normal" sample cannot be ruled out. Holland (1980) stated that a clinician's task, when using the original CADL, "is to determine how closely a given aphasie individual's CADL score approximates normal functional communication" (p. 32). Although the CADL-2 shares the same theoretical motivation (Holland et al., 1999), neither a normal nor an aphasie participant group was included in its standardization. The authors later collected data regarding the performance of 30 normal adults. However, direct comparison of normal and aphasie performance is not reported. Similarly, the authors of the ASHA FACS did not include normal adult participants in its standardization. Moreover, information regarding the performance of a combined group of aphasie and cognitive-communication-impaired participants is not provided.

second, the use of cutoff scores for differentiating normal from aphasie performance is problematic. For example, Kertesz (1979) specified an AQ of 93.8 as a cutoff for discriminating normal from aphasie performance. This score, however, represents the mean AQ performance of a third group (individuals with diffuse/ subcortical brain damage) and is unrelated to his "normal" or aphasie groups' performance. Porch (1967) reported that normal, literate adults usually average almost 15.0 on the PICA (the highest possible score is 16.0). However, because normal participants were not included in its standardization, no data were provided to support this position. An assumed ceiling effect for normal performance on such measures of language or communication may not be correct (Biddle et al., 2002), and only one study has examined the effect (Goodglass & Kaplan, 1983, pp. 27-28). Normal performance is addressed by the authors of the CADL-2 in that, "an individual who does not have a neurogenic disorder will probably score higher than most of this normative [175-participant, disordered] group (i.e., he or she will probably have a percentile rank above 96)" (Holland et al., 1999, p. 19). A cutoff score for aphasie performance is otherwise unspecified. Finally, cutoff scores are not provided for differentiating normal from impaired performance on the ASHA FACS.

Third, previous reports suggest that the test performance of aphasie and normal groups may overlap. For example, Duffy and Keith (1980) compared the PICA performance of 131 normal adults to that of Porch's (1967) heterogeneous sample of 150 patients. They found that, on five of the PICA'S 18 subtests, the degree of overlap between normal and aphasie performance exceeded 30%. Based on the groups' overall PICA scores, they concluded that approximately 8% of left-brain-injured patients might not be distinguishable from normal adults. This is consistent with Schuell, Jenkins, and Jimenez-Pabon's (1964) observation of an 8% overlap between the performance of non-brain-injured and aphasie adults on the Minnesota Test for the Differential Diagnosis of Aphasia (Schuell, 1965). Kertesz (1979) administered the Ravens Coloured Progressive Matrices (RCPM), a nonverbal test of logical reasoning and visuospatial ability (p. 255) that is included as part of the WAB Cortical Quotient, to 111 aphasie and 19 non-brain-damaged, age-matched adults. he determined the portion of aphasie participants performing normally on the RCPM by defining normal performance as "the same or better than the mean minus one standard deviation from the mean of normal controls" (p. 259). Forty-two percent of the aphasie individuals performed above this level. Finally, Holland and colleagues (1999) suggested that neurogenically impaired and normal adults' functional communication, as measured by the CADL-2, may overlap by at least 4%.

Supplemental indices of performance, such as degree of overlap and index of determination, may allow a more accurate distinction between aphasie and normal performance and thus improve the clinical usefulness of formal tests. The degree of overlap represents the percentage of aphasie participants scoring at or above the minimum expected score for 95% of normal volunteers. The index of determination (Young, 1976) uses regression analysis concepts to measure the degree to which being classified, a priori, as "aphasie" or "normal" predicts performance on a given measure. A small degree of overlap or a large index of determination is associated with good discriminatory power in distinguishing impaired from normal performance (Duffy & Keith, 1980). Using these analyses, Duffy and Keith (1980) found the single best discriminator between left-brain-injured and normal PICA performance to be the amount of time required to complete the test. Of the PICA subtests, the best discriminators between left-brain-injured and normal performance were Graphic subtests A through E. Of the modality scores, the Graphic modality score best distinguished between groups' communicative ability.

Summary and Statement of Purpose

Normal elderly and mildly aphasie individuals may exhibit similar impairments in comprehension and expression. Accurate differential diagnosis permits provision of an appropriate prognosis and focused treatment. Although the aforementioned tests-PICA, WAB, CADL-2, and ASHA FACS-are used to identify language and communication deficits characteristic of aphasia, summary scores derived from these measures may not distinguish well between normal and mildly aphasie behavior. Supplemental indices of performance may allow a more accurate distinction between aphasie and normal performance and thus improve the clinical usefulness of formal tests.

The current study was designed to assess the discriminative validity of the four general language and functional communication measures discussed above. Rationale for selecting these measures included demonstrated evidence of their validity for identifying behaviors common in aphasia, in the absence of empirical evidence of their validity for differentiating normal from aphasie performance. The primary research question was, Do general language and functional communication measures differentiate normal from aphasie adults? The secondary research question was, Which subtests within each measure best distinguish between the groups?

Method

Participants

all participants met the following selection criteria-age: between 40 and 80 years; education: premorbidly literate in English; hearing: no worse than an estimated 40 decibel (dB) speech recognition threshold (SRT) in the better ear as determined by the Carhart method (i.e., the average of pure-tone thresholds at 500 and 1000 Hz, minus 2 dB; Carhart, 1971); corrected visual acuity: no worse than 20/100 in the better eye as determined by a pocket-sized Snellen chart; motor ability: one upper extremity sufficiently intact to point, gesture, and write; and, medical and psychological status: no coexisting major medical or psychological disorders that would interfere with participation in the study, including no active treatment for substance abuse. Aphasie participants had a history of one or more strokes, brain damage confined to the left hemisphere, no history of other disease that would affect communication ability, and, a diagnosis of aphasia provided by the referring speech-language pathologist and/or physician. Also, each participant verified that his or her communication ability was impaired subsequent to a stroke and that the deficits persisted at the time of participation in the study.

all participants' oral-verbal expression, auditory comprehension, reading comprehension, and writing abilities; self-reported performance in everyday communication activities (e.g., reading the mail or understanding the radio); and self-reported social participation (e.g., maintaining social relationships or participating in community activities) were documented in a structured interview. The presence of aphasia was confirmed in all aphasie participants and excluded in all normal participants using the following operational definition:

Aphasia is an impairment, due to acquired... damage of the central nervous system, of the ability to comprehend and formulate language. It is a multi-modality disorder represented by a variety of impairments in auditory comprehension, reading, oral-expressive language, and writing. The disrupted language may be influenced by physiological inefficiency or impaired recognition, but it cannot be explained by dementia, sensory loss, or motor dysfunction. (Rosenbek, LaPointe, & Wertz, 1989, p. 53)

The presence or absence of aphasia was also based on observations from the structured interview, social and medical history, and the selection criteria detailed earlier. Because it is clinically important and often difficult to differentiate normal adults from mildly or "recovered" aphasie individuals, the aphasie adults in this study were at least 6 months poststroke.

Normal adult participants had no history of brain injury or other disease that would affect communication ability. Absence of brain damage in normal adults was based on self-reported history, whereas the presence of brain damage in aphasie participants was based on neuroradiological data.

Eighteen aphasie and 18 normal adults met selection criteria and participated in the study. Table 1 shows demographic information regarding age and education for both groups and regarding months poststroke for the aphasie group. Thirteen men (72%) and 5 women (28%) composed each group. Neither age nor educational level differed significantly between groups (p > .05). Twelve (67%) of the aphasic participants suffered ischemic strokes, 4 (22%) suffered hemorrhagic strokes, and 2(11%) suffered a ruptured aneurysm. Subsequent administration of the WAB classified 5 (28%) aphasic participants as having anomic aphasia, 4 (22%) as having conduction aphasia, 3 (17%) as having Broca's aphasia, and 2 (11%) as having Wernicke's aphasia. The scores of 4 (22%) aphasic participants exceeded Kertesz's (1979) cutoff for aphasic performance (i.e., AQ of 93.8); thus, they were classified as "not aphasic." These individuals exhibited persistent, mild impairment in verbal expression and auditory comprehension and more significant impairment in reading and/or writing. The WAB, however, uses only two of four language modalities-auditory comprehension and verbal expression-to diagnose and classify aphasia. The WAB classified all (100%) of the normal participants as not aphasic.

Procedures

All participants were administered two general language tests-the PICA and the WAB-and one functional communication assessment-the CADL-2. A second functional communication measure-the ASHA FACS-was completed for each participant. In order to allow a maximum number of interactions with participants before completing the ASHA FACS, tests were administered in the following order: WAB, PICA, CADL-2, and ASHA FACS.

Statistical Analyses and Purposes

To determine whether the mean performance of aphasic adults on summary measures of general language and functional communication is significantly different from that of normal adults, Wilcoxon's ranked-sum tests were performed, comparing aphasic and normal groups' mean performance on the following scores-(a) PICA: overall mean raw score; (b) WAB: AQ and Cortical Quotient (CQ); (c) CADL-2: total score; and, (d) ASHA FACS: Overall Communication Independence Domains mean score and Overall Communication Dimensions mean score. To correct for possible family wise error, an alpha level of p

To determine whether the performance ranges of normal and aphasic adult groups overlap, group ranges of performance were compared and 95% confidence intervals were calculated to establish limits on the true difference between each pair of mean test scores (Sackett, Richardson, Rosenberg, & Haynes, 1997). To determine which subtests best differentiated normal from aphasic adults, supplementary analyses involving the degree of overlap in performance and indices of determination (Duffy & Keith, 1980; Young, 1976) were used.

Results

Comparison of Summary Scores

Table 2 contrasts the mean performance of normal and aphasic adults on standardized assessments of general language and functional communication. The mean performance of aphasic adults for all summary scores was significantly lower than that of normal adults. However, 95% confidence intervals calculated for the true difference between the group means show that, depending on the unknown degree of measurement error present in the test scores, actual differences may have been very small. For example, the true difference between the mean performance of aphasic and normal adults on the ASHA FACS Communication Independence Domains mean overall score may have been as small as 0.21 of 1 point, out of 7 possible points. Table 2 also shows that, for all summary scores, group ranges in performance overlapped.

Supplemental Analyses: Index of Determination and Degree of Overlap

Tables 3 through 6 provide additional data relevant to differentiating aphasic from normal adults. The degree of overlap represents the percentage of aphasic participants scoring at or above the minimum expected score for 95% of normal volunteers. The index of determination (Young, 1976) uses regression analysis concepts to measure the degree to which being classified, a priori, as aphasic or normal predicts performance on a given measure. Duffy and Keith ( 1980) suggested that a small percentage of overlap or a large index of determination is associated with good discriminatory power in distinguishing impaired from normal performance. To determine which subtests best differentiate normal and aphasic performance, obtained indices of determination and degrees of overlap were ranked according to quartiles.

Table 3 shows data regarding differential PICA performance. By applying the above criteria, the PICA Pantomime Modality subscore, overall raw score, and Verbal Modality subscore best differentiated between the aphasic and normal groups included in this investigation. Table 4 shows data regarding differential WAB performance. The Reading and Writing subscore, CQ (which summarizes overall test performance), and Naming subscore best differentiated between our aphasic and normal groups. Table 5 shows data regarding differential CADL-2 performance. The time required for test completion best differentiated between groups. Table 6 shows data regarding differential ASHA FACS performance. By applying the above criteria, the Communication Dimensions Adequacy score, Communication Dimensions Overall mean score, and Communication Dimensions Promptness score best differentiated between the aphasic and normal groups included in this investigation.

Discussion

The results of significance testing indicate that the summary scores of four standardized general language and functional communication assessments differentiate normal from aphasic performance. This conclusion supports previous reports by Porch (1967), Kertesz (1979), Brauer, McNeil, Duffy, Keith, and Collins (1989), and Holland et al. (1999). The mean performance of our 18 aphasic adults was significantly lower than that of our 18 normal adults on all summary scores. However, 95% confidence intervals calculated for the true difference between means showed that, depending on the unknown degree of measurement error present in test scores, actual mean differences between the groups' summary scores may be very small. In addition, for every summary score, group performance ranges overlapped. The exact degree of overlap, calculated as the percentage of aphasic participants scoring at or above the minimum expected score for 95% of normal volunteers, exceeded 10% on each measure and, on some measures, was as high as 61%. These results support previous reports by Schuell et al. (1964), Kertesz (1979), Duffy and Keith (1980), and Holland et al. (1999).

We used degrees of overlap and indices of determination to determine which subscores best differentiated aphasie from normal performance. On each test, a subscore was better able to distinguish between groups' performance than was the summary score or scores. For the PICA, the Pantomime Modality subscore, overall raw score, and Verbal Modality subscore best differentiated the groups. Our results support Brauer and colleagues' (1989) observation regarding the importance of PICA verbal subscores. Performance on the Writing Modality and test completion time were the fourth and fifth best measures in distinguishing between our groups' PICA performance. This finding supports the observations of Duffy and Keith (1980) and Brauer et al. regarding the importance of graphic performance and test completion time in differentiating left-brain-injured from normal performance. For the WAB, the Reading and Writing subscore, CQ, and Naming subscore differentiated our groups best. Our results support Kertesz's (1979) position regarding the importance of Naming ability for differentiating normal from aphasic performance. For the CADL-2, time required for test completion distinguished best between our groups. For the ASHA FACS, the Communication Dimensions Adequacy score, Communication Dimensions Overall mean score, and Communication Dimensions Promptness score distinguished best between groups. Overall, the results of the current investigation suggest that chronically aphasic persons may continue to experience general difficulty with expressive language and functional difficulty with efficiency of performance.

Of note, the ability of gestural performance to differentiate between groups conflicts, depending on test selection. Although the Pantomime Modality subscore is particularly useful for discriminating normal from aphasic PICA performance, the Praxis subtest is the least useful measure for differentiating between-groups' WAB performance. The subtests differ in two aspects. First, required tasks differ by test. The WAB praxis contains a variety of tasks-requiring the use of the upper limbs and face at instrumental and complex levels-whereas the PICA pantomime tasks are more uniform-primarily requiring the use of upper limbs at an instrumental level of complexity. Second, scoring systems differ by test. The PICA uses a 16-point multidimensional score, and the WAB uses a 0-3-point scale. It is likely that the PICA scoring system is more sensitive to differences between normal and aphasic performance. That the aphasic group's praxis performance did not differ significantly from that of normal adults suggests that our stroke survivors did not suffer from significant limb apraxia, paralysis, or paresis. However, the contribution of non-language-related variables to significant between-groups differences on the four measures examined cannot be determined.

We compared the performance of two groups-normal and aphasic adults-on four measures-the PICA, WAB, CADL-2, and ASHA FACS. Although comparisons of summary scores showed significant mean differences between groups, our groups' performance overlapped by at least 10% on each measure. Because these tests may not distinguish abnormal from normal performance when an individual's score exceeds the 90th percentile for aphasic individuals, their contribution to differentiating "mild" or "recovered" aphasic behavior from that of normally aging adults may be limited. Moreover, the inherent design of these general language and functional communication tests limits their ability to determine whether linguistic deficits, other potential contributing factors, or a combination of processes underlie an individual's communication deficits (Chapey, 1994; Chapman & Ulatowska, 1994). For example, standardized language batteries are generally constructed based on empirical evidence of aphasic deficits (Chapman & Ulatowska, 1994). Their validity for accurately measuring normal language and communication behavior is undetermined. Although our normal volunteers generally performed well and similarly on these measures, the small sample size also prohibits a conclusive, generalizable statement of ceiling effects. Finally, because standardized language batteries sample primarily linguistic behavior, their validity for accurately differentiating linguistic and cognitive factors underlying comprehension deficits is questionable (Chapman & Ulatowska, 1991).

To enhance the accuracy of differential diagnosis, clinicians must combine formal test results with additional subjective and objective evidence. Our results show that supplemental analyses, such as indices of determination and degree of overlap, may enhance existing tests' ability to distinguish between aphasic and normal behavior. Supplemental language tests designed to capture the upper range of normal performance-for example, standardized reading tests-may enhance the identification of abnormal behavior in mildly aphasic people. And, supplemental cognitive tasks may delineate the role of underlying mechanisms in precipitating and maintaining communication deficits. Chapman and Ulatowska (1991, 1994) suggested that discourse tasks, which assess both organization of information and linguistic facility, may be most appropriate for differentiating linguistically and cognitively related disorders. To our knowledge, the discriminative validity of discourse tasks for differentiating normal from mildly aphasic performance has not been reported.

It may be argued that tests for aphasia are not designed to explore the range of normal performance, and ceiling effects may make it difficult to differentiate normal from aphasic performance. Certainly, none of the measures we examined were designed to explore the range of normal performance, and some of our normal volunteers did perform at ceiling on the measures. Nevertheless, authors of some of the measures provide a criterion for normal performance, for example, around 15 on the PICA (Porch, 1967) and 93.8 and above on the WAB AQ (Kertesz, 1982). Our results indicate that both are inadequate for differentiating normal from mildly aphasic performance. Cronbach (1949) cautioned that a test that helps make one decision might have no value for making another. Thus, asking the general question, "Is this test valid?" is inappropriate. The question to ask is, "How valid is this test for the decision that I want to make?" If the clinician's question is, "Do the PICA, WAB, CADL-2 and ASHA FACS have high discriminative validity for differentiating normal performance from mildly aphasic performance?" our results suggest that the answer is no.

Acknowledgments

This study was conducted at the VA Tennessee Valley Healthcare System, Nashville, TN, and was supported, in part, by a predoctoral fellowship from the Department of Veterans Affairs and by the Department of Hearing and Speech Sciences at Vanderbilt University School of Medicine. A summary of the results was presented in November 2000 as a poster session at the annual convention of the American Speech-Language-Hearing Association, Washington, DC. We thank Daniel Ashmead, Fred Bess, Frank Freemon, and Ralph Ohde for valuable discussion.

References

Biddle. A., Watson, L., Hooper, C., Lohr, K., & Sutton, S. (2002). Criteria for determining disability in speech-language disorders. Evidence Report/Technology Assessment No. 52 (Prepared by the University of North Carolina Evidence-Based Practice Center under Contract No. 290-97-0011). AHRQ Publication No. 02-E010. Rockville, MD: Agency for Healthcare Research and Quality.

Bowles, N., & Poon, L. (1985). Aging and retrieval of words in semantic memory. Journal of Gerontology, 40, 71-77.

Brauer, D., McNeil, M. R., Duffy, J. R., Keith, R. L., & Collins, M. J. (1989). The differentiation of normal from aphasic performance using PICA discriminant function scores. In T. E. Prescott (Ed.), Clinical aphasiology (Vol. 18, pp. 117-129). Boston: College-Hill Press.

Carhart, R. (1971). Observations on relations between thresholds for pure tones and for speech. Journal of Speech and Hearing Disorders, 36, 476-483.

Chapey, R. (1994). Assessment of language disorders. In R. Chapey (Ed.), Language intervention strategies in adult aphasia (3rd ed., pp. 80-120). Baltimore: Williams & Wilkins.

Chapman, S. B., & Ulatowska, H. K. (1991). Aphasia and aging. In D. Ripich (Ed.), Handbook of geriatric communication disorders (pp. 241-254). Austin, TX: Pro-Ed.

Chapman, S. B., & Ulatowska, H. K. (1994). Differential diagnosis in aphasia. In R. Chapey (Ed.), Language intervention strategies in adult aphasia (3rd ed., pp. 121-131). Baltimore: Williams & Wilkins.

Cohen, G. (1979). Language comprehension in old age. Cognitive Psychology, 11, 412-429.

Critchley, M. (1984). And all the daughters of music shall be brought low: Language function in the elderly. Archives of Neurology, 41, 1135-1139.

Cronbach, L. J. (1949). Essentials of psychological testing (2nd ed.). New York: Harper & Row.

Darley, F. L. (1979). The differential diagnosis of aphasia. In R. H. Brookshire (Ed.), Clinical Aphasiology Conference proceedings (pp. 23-29). Minneapolis, MN: BRK Publishers.

Davis, G. A. (2000). Clinical assessment and diagnosis. In Aphasiology: Disorders and clinical practice (pp. 43-66). Boston: Allyn and Bacon.

Duffy, J. R., & Keith, R. C. (1980). Performance of non-brain injured adults on the PICA: Descriptive data and a comparison to patients with aphasia. Aphasia-Apraxia-Agnosia, 2, 1-30.

Feier, C., & Gerstman, L. (1980). Sentence comprehension abilities throughout the adult lifespan. Journal of Gerontology, 35, 729-735.

Frattali, C., Thompson, C. K., Holland, A. L., Wohl, C. B., & Ferketic, M. K. (1995). American Speech-Language-Hearing Association Functional Assessment of Communication Skills for Adults. Rockville, MD: American Speech-Language-Hearing Association.

Goodglass, H., & Kaplan, E. (1983). The assessment of aphasia and related disorders (pp. 27-28). Media, PA: Williams and Wilkins.

Holland, A. L. (1980). CADL: A test of functional communication for aphasic adults. Baltimore: University Park Press.

Holland, A. L., Frattali, C., & Fromm, D. (1999). Communication Activities of Daily Living-Second Edition. Austin: Pro-Ed.

Kemper, S. (1988). Aging and syntactic limitations on language. In L. L. Light & D. M. Burke (Eds.), Language, memory, and aging (pp. 58-76). New York: Cambridge University Press.

Kertesz, A. (1979). Aphasia and associated disorders: Taxonomy, localization, and recovery. New York: Grune & Stratton.

Kertesz, A. (1982). Western Aphasia Battery. New York: Grune & Stratton.

North, A. J., Ulatowska, H. K., Macaluso-Haynes, S., & Bell, H. (1986). Discourse performance in older adults. International Journal of Aging and Human Development, 23, 267-283.

Obler, L. K., & Albert, M. (1981). Language in the elderly aphasic and in the dementing patient. New York: Academic Press.

Obler, L. K., & Albert, M. (1984). Language in aging. In M. L. Albert (Ed.), Clinical neurology of aging (pp. 245-253). New York: Oxford University Press.

Peach, R. (1987). Language functioning. In H. Mueller & V. Geoffrey (Eds.), Communication disorders in aging (pp. 238-270). Washington, DC: Gallaudet University Press.

Porch, B. E. (1967). Porch Index of Communicative Ability. Palo Alto, CA: Consulting Psychologists Press.

Rosenbek, J. C., LaPointe, L. L., & Wertz, R. T. (1989). Aphasia: A clinical approach (p. 53). Boston: College-Hill Press.

Sackett, D. L., Richardson, W. S., Rosenberg, W., & Haynes, R. B. (1997). Evidence-based medicine: How to practice & teach EBM. New York: Churchill Livingstone.

Schow, R. L., Christensen, J. M., Hutchinson, J. M., & Nerbonne, M. A. (1978). Communication disorders of the aged: A guide for health professionals. Baltimore: University Park Press.

Schuell, H. (1965). The Minnesota Test for Differential Diagnosis of Aphasia. Minneapolis: University of Minnesota Press.

Schuell, H. Jenkins, J. J., & Jimenez-Pabon, E. (1964). Aphasia in adults: Diagnosis, prognosis, and treatment. New York: Harper and Row.

Young, M. (1976). Application of regression analysis concepts to retrospective research in speech pathology. Journal of Speech and Hearing Research, 19, 5-18.

Received January 23, 2002

Accepted December 17, 2002

DOI: 10.1044/1058-0360(2003/077)

Katherine B. Ross

Carl T. Hayden VA Medical Center, Phoenix, AZ

Robert T. Wertz

Vanderbilt University School of Medicine, and

VA Tennessee Valley Healthcare System, Nashville

Contact author: Katherine B. Ross, PhD, Carl T. Hayden Medical Center, 650 E. Indian School Road, CS/126, Phoenix, AZ 85012. E-mail: katherine.ross3@med.va.gov

Copyright American Speech-Language-Hearing Association Aug 2003
Provided by ProQuest Information and Learning Company. All rights Reserved