J Am Acad Audiol 3 : 303-307 (1992) SPIN Test Performance of Elderly Hearing- Impaired Listeners Donald J. Schum* Lois J. Matthews t Abstract Normative information is available on the expected performance of hearing-impaired individuals on the revised version of the Speech Perception in Noise (SPIN) test. This normative sample includes individuals with sensorineural hearing loss drawn from an age range of 19 to 69 years, with 79 percent of the sample being less than 65 years of age. In the present study, SPIN scores were obtained from both ears of 98 elderly, hearingimpaired listeners (age range: 60 to 78 years). When the percent-correct scores on the High-Predictability (PH) items and on the Low-Predictability (PL) items are viewed as two separate measures, mean performance of the current "elderly-only" sample is similar to the performance of the original normative sample. However, for a certain percentage (18%) of the aged ears, the contextual cues provided in the PH sentences were not used as effectively as expected based on the normative guidelines. Key Words : Word recognition, elderly, hearing loss he Speech Perception in Noise (SPIN) test (Kalikow et al, 1977) was designed T to assess word recognition skills under two controlled levels of contextual information. The listener's task is to listen to a test sentence presented in a background of speech babble and then repeat the last word of the sentence. High- Predictability (PH) sentences provide several syntactic and semantic cues as to the identity of the final word (e.g., "I have a cold and a sore throat"). Low-Predictability (PL) sentences provide few if any such cues (e.g., "He is considering the throat"). PL items are thus similar to classic word recognition test items (monosyllabic words presented in a carrier phrase ; e.g., "Say the word throat"). Therefore, performance on PL items reflects basic word recognition skills in noise. The better performance typically seen on the PH items compared to the PL items reflects the listener's ability to use contextual information when listening to speech. *Department of Otolaryngology, University of Iowa Hospital, Iowa City, Iowa t Otolaryngology and Communicative Sciences, Medical University of South Carolina, Charleston, South Carolina Reprint requests : Donald J. Schum, Department of Otolaryngology, University of Iowa Hospitals, Iowa City, IA 52242 Bilger et al (1984) collected normative data on the performance of hearing-impaired listeners on the SPIN test. This normative study led to a revision of the SPIN test lists in order to attain eight lists with similar difficulty and reliability. Bilger et al (1984) noted that the difference in performance between the PL and PH items varied in magnitude across the range of PL performance. Based on these data, Bilger (1984) described the (essentially) 95 percent confidence intervals for the expected relationship between PL and PH scores. For example, for a PL score of 24 percent, the PH score is expected to fall between 80 and 88 percent; whereas for a PL score of 60 percent, the PH score should fall between 92 and 100 percent. Cases in which the PH score is higher than expected based upon a given PL score reflect a situation in which the listener appears to be receiving a greater than expected amount of benefit from the contextual cues. When the PH score is lower than expected based upon a given PL score, the listener apparently uses the available contextual cues less effectively than most hearing-impaired listeners. The normative sample of Bilger et al (1984) included 128 ears from 128 listeners aged 19 to 69 years, with 79 percent ofthe sample less than 65 years old. Since most hearing-impaired indi- 303
Journal of the American Academy of Audiology/Volume 3, Number 5, September 1992 viduals are elderly (Catlin, 1986), it is of interest to evaluate the normative guidelines for the SPIN test specifically for elderly listeners. Several investigators (Marshall, 1981 ; Hayes,1985 ; Kausler,1988) have noted that elderly listeners may perform poorly on certain central auditory processing or cognitive skills tests. Therefore, the purpose of this investigation is to compare the performance on the SPIN test of a large number of elderly, hearing-impaired listeners to the normative guidelines presented by Bilger (1984), with emphasis on the elderly listener's ability to make full use of the linguistic cues in the PH items. Subjects METHOD 0 20 40 x m ' 60 80 100 "--F'- BILGER at al., 1984 CURRENT SUBJECTS, RIGHT CURRENT SUBJECTS, LEFT 100 1000 10000 FREQUENCY (Hz) Figure 1 Mean (±1 standard deviation) audiometric thresholds for the right and left ears for the subjects in the current investigation and for the test ears of the subjects from Bilger et al (1984). Testing was performed on both ears from 98 sensorineurally hearing-impaired listeners ranging in age from 60 to 78 years. Subjects were recruited from patient files of a medical school audiology clinic and are included in a long-term, multidisciplinary study of presbycusis. In order to be included in this portion of the study, the patient needed to demonstrate at least one audiometric threshold (from 250 through 6000 Hz) of greater than 25 db HL, with bone-conduction thresholds within 10 db of airconduction thresholds, bilaterally. Also, the estimated threshold for speech babble, following the procedures described by Bilger (1984), must have been equal to or less than 40 db HL, bilaterally. (This criterion allowed the SPIN test to be presented at 50 db SL re : the estimated babble threshold as called for by Bilger, 1984, yet still be presented at or below 90 db HL.) Figure 1 provides the mean (±1 standard deviation) hearing thresholds for the right and left ears of the current subjects. Also included in Figure 1 are the mean audiometric thresholds for the test ears of the subjects from Bilger et al (1984). The subjects from Bilger et al (1984) were tested monaurally, with an equal number of right and left ears. Figure 2 provides the distribution of ages of the current subjects and those of Bilger et al (1984). via earphones (TDH-39). The speech stimuli were presented at 8 db above the competing speech babble. The presentation level of the speech stimuli was set at 50 db above the estimated threshold for the speech babble (as per Bilger, 1984). Procedure Each subject was tested as part of a larger series of clinical and laboratory studies of presbycusis. Included in this test series were speech reception threshold (SRT) testing and monosyllabic word recognition testing (using NU-6 materials) at 30 db SL re : the SRT. Before SPIN testing, pure-tone audiometrics had been obtained for the test ear. All test ears were determined to exhibit sensorineural hearing Speech Stimuli Each subject, for each ear tested, provided verbal responses to one 50-item list from the Revised SPIN test. Each list contained 25 PL items and 25 PH items. The stimuli were taperecorded (provided by R. Bilger) and presented AGE CATEGORY Figure 2 The distribution of ages for subjects in the current investigation and from Bilger et al (1984). 304
SPIN Test Performance/Schum and Matthews Table 1 Performance on PH and PL Items Percent Correct (SD) Current Sample Right ear 87.4 (16.4) 44.2 (24.3) Left ear 85.7 (18.8) 45.8 (24.4) Bilger et al (1984) 87.3 (18.9) 45.9 (21.2) loss and to be negative for any relevant otoneurologic pathology. All audiometric testing was performed in a double-walled soundattenuating booth. The order in which the ears were tested and the choice of SPIN list number were varied from subject to subject. PH RESULTS T able 1 compares the data obtained on the PH and PL items in this study with those of Bilger et al (1984). The paired t-test was nonsignificant (p >.05) for the PH score in the right compared to the left ear. The paired t-test was also nonsignificant (p >.05) for the PL score in the right compared to the left ear. T-testing for independent samples was carried out in order to compare the results for the current subjects with the results from Bilger et al (1984). Separate analyses were carried out in order to compare right ear PH scores with Bilger et al (1984) PH scores, right ear PL scores with Bilger et al (1984) PL scores, left ear PH scores with Bilger et al (1984) PH scores, and left ear PL scores with Bilger et al (1984) PL scores. All t-test results were nonsignificant (p >.05). Figure 3 is a scatterplot of the relation between the PH and PL scores observed in this study. Right ear scores are represented by filled triangles and left ear scores are represented by unfilled circles. Heavy lines indicate the 95 percent confidence interval for the expected relation between PH and PL scores reported by Bilger (1984). Data points that fall above the confidence region indicate PH scores higher than expected based upon the PL score. Data points that fall below the confidence region indicate PH scores lower than expected based upon the PL score. For the right ear, 14 percent (n = 14) of the scores fell above the 95 percent confidence region and 14 percent (n = 14) fell below the confidence region. For the left ear, 8 percent (n = 8) of the scores fell above the confidence region and 22 percent (n = 22) fell below the confidence region. For the 36 data points that fell below the confidence region, 28 came from subjects whose PL other ear PH score was not depressed (below the 95% confidence region) and 8 came from subjects whose PH score was depressed for both ears. For the 28 subjects with a depressed PH score in only one ear, 10 showed the depressed score in the right ear and 18 showed the depressed score in the left ear. The proportion.36 (10 of 28) is not significantly different (p >.05) from chance (.5). Table 2 compares those ears with and without depressed PH scores on a variety of demographic and audiometric measures. Statistical comparisons were carried out on these variables for ears with depressed versus nondepressed PH scores using t-testing for independent samples (except for sex, which was evaluated via binomial sampling theory). In the right ear, depressed PH performance was associated with significantly (p <.05) higher average thresholds, higher SRTs, and lower word recognition scores. In the left ear, depressed PH performance was associated with significantly (p <.05) higher average audiometric thresholds, higher SRTs, lower word recognition scores, and lower PL scores. DISCUSSION n general, the listeners in this elderlyonly sample demonstrated performance on the SPIN test similar to that of the generally younger hearing-impaired subjects from the original normative sample. Mean performance 1 3 5 7 9 11 13 15 17 19 21 23 25 LOW NUMBER CORRECT Figure 3 Scatterplot of PH versus PL scores for the right and left ears of the 98 subjects. Right ear scores are represented by filled triangles and left ear scores are represented by unfilled circles. Heavy lines indicate the 95 percent confidence interval for the expected relation between PH and PL scores from Bilger (1984).
11111 *11,, + l ' i ti Y 1l li+l k~ 1411!111- NIN Nl11iiIIIIN1~If~1lP+I NI Journal of the American Academy of Audiology/Volume 3, Number 5, September 1992 Table 2 Demographic and Audiometric Variables Depressed PH Scores Nondepressed Right ear (n = 14) (n = 84) Age (years) 67.9 66.1 Sex (% female) 35.7 33.3 Average hearing level (db HL) 44.9 31.4 (mean at.5, 1, 2, and 4 khz) Audiometric slope (db/octave) 14.3 13.5 (mean from.5 to 4 khz) Speech reception threshold (db HL) 32.1 22.4 Word recognition score (% correct) 72.9 85.9 (at 30 db SL re'. SRT) PL score (% correct) 34.9 45.6 Left ear (n = 22) (n = 76) Age (years) 67.7 66.3 Sex (% female) 31.8 34.2 Average hearing level (db HL) 42.0 30.9 (mean at.5, 1, 2, and 4 khz) Audiometric slope (db/octave) 15.0 13.9 (mean from.5 to 4 khz) Speech reception threshold (db HL) 28.9 22.1 Word recognition score (% correct) 70.0 87.4 (at 30 db SL re : SRT) PL score (% correct) 28.9 50.7 on both the PH items and the PL items was not significantly different than that of the Bilger et al (1984) subjects. However, the normative guidelines proposed by Bilger (1984) for the SPIN test may not fully reflect the performance characteristics of elderly listeners. Bilger et al (1984) note that the magnitude of the difference between the PH and the PL score will vary across the range of PL scores. Therefore, the PH and PL scores can not be combined in any simple arithmetic manner to yield one performance score. The 95 percent confidence intervals were established in order to describe the expected relation between these measures. Thirty percent (n = 58) of the scores fell outside the proposed 95 percent confidence intervals. Although a significant number (n = 22) of the scores fell above the target region (reflecting greater than expected benefit of linguistic cues), most of these scores were only one position away from the target region (that is, one less PH word correct would put the score within the 95 percent confidence region). In contrast, an even greater number of scores (n = 36) fell below the target region, and most of these scores fell more than one position away from that region. Given that the upper bound of the confidence region is near 100 percent for a significant portion of the PL performance range, it may be argued that scores are limited from extending above the confidence region due to a ceiling effect. However, inspection of the data at and below eight PL items correct-a region in which there is adequate room above the confidence region-continues to reveal a greater spread in the scores below the confidence region than above. Therefore, the present data do not appear to demonstrate simply greater variability as compared to the Bilger (1984) norms. Rather, there seems to be a true tendency for the PH score to be lower than expected given the PL score. Bilger (1984) suggests that scores that fall outside the confidence interval "probably reflect non-auditory factors" (p. 34). Given a relatively large literature documenting declines in cognitive and information processing skills in the elderly (see Cohen and Wu,1980 ; Woodruff, 1983 ; Welford, 1985 ; and Kausler, 1988 for reviews), it is reasonable to assume that older listeners will be at greater risk for certain perceptual difficulties that extend beyond the peripheral hearing loss. The results from Table 2 suggest that as the peripheral hearing loss becomes more severe, the likelihood of elderly listeners being able to make full use oflinguistic cues declines. One interpretation of these results is that, compared to younger listeners, elderly listeners are not as efficient at compensating for an impoverished input from the peripheral auditory system. However, since 28 of the 36 depressed PH scores were from listeners 306
SPIN Test Performance/Schum and Matthews who showed nondepressed PH scores in the other ear, the current results cannot be attributed to a general decline in linguistic skills. It is difficult to specify precisely which changes in the nervous system account for this performance decrement. Acknowledgment. Drs. Lisa Hellstrom and John Gibson provided important comments on an earlier version of this manuscript. Portions of this paper were presented at the Annual Midwinter Meeting of the Association for Research in Otolaryngology, St. Petersburg Beach, February, 1990. This work is supported by NIH grant P50-NS25039 to the Medical University of South Carolina. REFERENCES Bilger RC. (1984) Manual for the Clinical Use of the Speech Perception in Noise (SPIN) Test. Champaign- Urbana : University of Illinois Press. Bilger RC, Nuetzel JM, Rabinowitz WM, Rzeczkowski C. (1984) Standardization of a test of speech perception in noise. J Speech Hear Res 27 :32-48. Catlin Fl. (1986) Noise-induced hearing loss. Am J Otol 7:141-149. Cohen D, Wu S. (1980) Language and cognition. Ann Rev Gerontol Geriatr 1:71-96. Hayes D. (1985) Aging and speech understanding. Semin Hear 6:147-159. Kalikow DN, Stevens KN, Elliott LL. (1977) Development of a test of speech intelligibility in noise using sentence materials with controlled word predictability. J Acoust Soc Am 61 :1337-1351. Kausler SH. (1988) Cognition and aging. In: Shadden BB, ed. Communication Behavior and Aging : A Sourcebook for Clinicians. Baltimore : Williams & Wilkins, 79-106. Marshall L. (1981) Auditory processing in aging listeners. J Speech Hear Disord 46 :226-240. Welford AT. (1985) Changes of performance with age : An overview. In : Chamess N, ed. Aging and Human Performance. New York : John Wiley & Sons, 333-369. Woodruff DS. (1983) A review of aging and cognitive processes. Res Aging 5:139-153.