Glossary of abbreviationsSCIP: Southern Cochlear Implant ProgrammeSCIPA: Southern Cochlear Implant Programme- Adult HINT: Hearing in Noise Test (described in introduction) HINT A: auditory cues alone HINT AV: audiovisual cues S/O: Switch- on of cochlear implant BKB sentence test: Bench, Kowal and Bamford sentence lists—open-set sentence lists (i.e. response options not prescribed in advance) similar to HINT CID sentence test: Central Institute for the Deaf everyday sentence lists- open-set sentence tests similar to HINT K-CID sentence test: Korean version of the CID test NVA speech recognition: Nederlands Vereninging Audiologie—Dutch open-set speech recognition tests consisting of monosyllabic (consonant-vowel-consonant) word lists NU-6 words: Northwestern University Auditory Test No. 6: open-set lists of monosyllabic words A cochlear implant (CI) is an electronic device which bypasses the auditory function of the outer, middle and inner ear and stimulates the cells of the spiral ganglion (and therefore the cochlear nerve) directly. Cochlear implants are indicated in severe and profound sensorineural hearing loss (SNHL) when hearing aids provide insufficient information for understanding speech. Cochlear implantation was first performed in New Zealand in 1986.1 An initially small national programme was based in Auckland. Cochlear implants were first performed in Christchurch in 1998. The Southern Cochlear Implant Programme (SCIP) was established in 2003. Located in Christchurch it serves the South Island and lower North Island. It is divided into separate Adult (SCIPA) and paediatric programmes. The SCIPA comprises two surgeons, two audiologists and a rehabilitationist. Potential candidates are referred to the programme, principally by audiologists and otolaryngologists, and undergo assessment by all three components of the CI team. Currently 80% of referrals are found to be suitable candidates. The key audiologic test performed is aided speech discrimination, details of which are provided below. This testing aims to determine how well the candidates can utilise their aided hearing to understand speech. The same testing processes are used post-implant to assess the effectiveness of the CI. Previous research has indicated that there are many factors contributing to the outcome of CI including age, duration of deafness/age at onset of deafness, anatomic issues as determined by radiology, other handicaps, speech and language abilities, functional hearing, family and social support, expectations/motivation, educational setting, communication mode, availability of support services, and the intensity of post-implant rehabilitation therapy.15,4 Many studies have shown a significant negative correlation with duration of deafness,4,7,9,11,12 and a positive correlation with age at onset of deafness, hearing aid use, and progressive hearing loss.7 In particular, hearing loss prior to development of speech and language inhibits its development and also reduces the effectiveness of cochlear implantation later on. There are differing findings in the literature regarding the time after CI that maximal benefit is obtained. Some studies indicate that this is reached as early as 3 months,10others suggest it is as late as 2 years.4 This is an important factor to be able to counsel patients about before and after their CI, so they will have some idea of their likely long-term outcome as they carry out their rehabilitation. In 2005 our Programme introduced the Nucleus® Freedom™ (Cochlear Ltd, Lane Cove, New South Wales, Australia) device which superseded the older technology from the same company of the Nucleus 22 and the later Nucleus 24 cochlear implants. Anecdotally we thought our patients were experiencing more rapid improvements in speech discrimination and slightly higher levels of peak performance, but wished to accurately investigate this. The objective of this study was to assess the outcome with respect to improvement in speech discrimination test results after CI in patients enrolled at SCIPA, and to determine how long after a CI there is a plateau in performance. We also aimed to assess the effect of duration of deafness and percentage of life spent deaf on results, and the effects of a new CI device. Methods A retrospective review of all files of patients receiving follow-up in Christchurch for CI between 1999 and Dec 2008 was undertaken. Exclusion criteria were patients under age 18 (these would be under the Paediatric Programme), and patients with no or inadequate pre-implant data. Some patients had CI performed by the Southern Cochlear Implant Programme - Adult, then moved to the Northern Cochlear Implant Programme (NCIP) for follow-up. Most of these were included, as follow-up data was sent from the NCIP. Some patients had received their CI in NCIP and then moved to the SCIPA for follow-up. These were included if we had a copy of their pre-implant data sent from NCIP. Files were obtained and the following were recorded: age, sex, type of implant: Nucleus® Freedom™ versus older devices - Nucleus 22 and Nucleus 24, age of onset of deafness, age at implant, percentage of life spent deaf (calculated from above), cause of deafness where known, and speech discrimination test results. We originally intended to measure both age at onset of hearing loss and age when hearing aids were first fitted, along with which sides were aided. Due to lack of consistent information we were only able to estimate age of onset of deafness with any degree of accuracy. Speech discrimination was measured using the hearing in noise test (HINT), which is a standardized test of speech recognition using sentences delivered in a sound field of 55 dB HL sound pressure level. In determining cochlear implant candidacy, HINT is performed without background noise, despite its name. HINT consists of 25 equivalent 10-sentence lists, two lists are used at each assessment. The tests are played on a DVD with a speaker/amplifier and are tested both with audiovisual (AV) cues (watching the presenter read the lists whilst listening) and with only auditory (A) cues (visual display turned off). The patient is scored for the percentage of words correctly identified on the lists. This is carried out pre-implant with their hearing aid on, at switch-on (s/o) of CI, then at 1 month, 3 months, 6 months, 9 months, 12 months, 18 months and 3 years post-implant where available. The following were assessed: improvement in HINT A and AV over time post-CI, overall improvement at 9, 12 and 18 months, characteristics of patients lost to follow-up or with missing data, relationship between final outcome and duration of deafness, percentage of life spent deaf and type of implant used. Statistical methods: All analyses were carried out using SAS 9.1. To investigate the time course of recovery up to 9 months after an implant, measures A and AV were analysed separately using repeated measures analysis of variance (PROC ANOVA) with time as the repeated measure (pre-implant, s/o, 1 month, 3 months, 6 months and 9 months), and a between subjects factor based on division of the patients into approximately three equal-sized groups based on their pro-operation scores. The Huynh-Feldt adjustment for non-sphericity was used for repeated measures sources of variance. Chi-square tests and t-tests were used to compare those who had complete data up to 9 months and those who did not on pre-implant variables. Prediction of long-term outcome used data from 18 months if available and otherwise the 12 month values were used. Including 12 month values in this way increased the number of A outcome observations from 40 to 46 and the number of AV observations from 37 to 44. Prediction of outcome was carried out using one way analysis of variance (PROC GLM). Because of ceiling effects, particularly for AV, Welch’s test was used if Levene’s test for the homogeneity of variance was significant with α=0.10, although this precaution mostly made little difference. Results A total of 171 patients were on the cochlear implant rehabilitation files at SCIPA. Of this total, 57 were excluded for the following reasons: 11 had been implanted too recently for analysis (2 months or less), 12 had transferred from the Paediatric Programme so did not apply to our study, 17 had been implanted prior to 1997 so speech discrimination tests other than HINT had been used for their pre-implant and post-implant assessments, nine had no follow-up notes available, four were non-users (discussed later), and four did not have pre-implant speech discrimination tests available. A total of 114 patients were then left for analysis. However, among this group there was much missing data. Only 55 had complete data for A and 53 for AV. Comparison of those with complete data and the remaining patients showed no significant differences in any of the background characteristics or of pre-implant functioning. We used complete data only for the analyses as imputation of data was deemed inappropriate due to the ceiling effect of the results. This was to ensure internal validity, i.e. comparing people across time for the same people. Pre-implant characteristics are shown in Tables 1-3, and the pre-implant scores on HINT are shown in Tables 4 and 5. Table 1. Characteristics of those with a cochlear implant between 1999 and 2008 Variables Number assessed Mean (SD) Range Age (years) Age (years) of onset of deafness Duration of deafness (years) Percentage of life deaf HINT pre-implant A HINT pre-implant AV 113 111 111 111 112 111 50 (15.3) 21 (19.4) 30 (14.5) 64% (30.5) 15 (16.8) 67 (23.9) 20-83 1-63 1-63 2-99 0-65 0-100 Table 2. Mean HINT AV scores (% correct) at pre-implant, switch-on (s/o), and 1-9 months post-implant: 53 patients Variables Mean HINT AV score SD Range Pre-implant s/o 1 month 3 months 6 months 9 months 66 82 93 96 96 98 24 17 10 9 6 4 0-100 32-100 67-100 57-100 74-100 79-100 Table 3. Mean HINT A scores (% correct) among those with complete data up to 9 months by grouping based on pre-implant scores (55 patients) HINT A scores pre-implant Number Mean pre-implant score (SD) Mean s/o score (SD) Mean score at 1 month (SD) Mean score at 3 months (SD) Mean score at 6 months (SD) Mean score at 9 months (SD) 0 1-15 16+ 18 17 20 0 (0) 6 (4) 34 (12) 33 (27) 23 (28) 46 (29) 60 (27) 44 (31) 69 (29) 71 (29) 53 (33) 79 (22) 77 (27) 65 (31) 82 (24) 83 (21) 69 (29) 83 (25) Total 55 14 (17) 34 (29) 58 (30) 68 (30) 75 (28) 79 (25) Table 4. Mean HINT AV scores (% correct) among those with complete data up to 9 months by grouping based on pre-implant scores (53 patients) HINT AV scores pre-implant Number Mean pre-implant score (SD) Mean s/o score (SD) Mean score at 1 month (SD) Mean score at 3 months (SD) Mean score at 6 months (SD) Mean score at 9 months (SD) 0-60 61-78 79+

Glossary of abbreviationsSCIP: Southern Cochlear Implant ProgrammeSCIPA: Southern Cochlear Implant Programme- Adult HINT: Hearing in Noise Test (described in introduction) HINT A: auditory cues alone HINT AV: audiovisual cues S/O: Switch- on of cochlear implant BKB sentence test: Bench, Kowal and Bamford sentence lists—open-set sentence lists (i.e. response options not prescribed in advance) similar to HINT CID sentence test: Central Institute for the Deaf everyday sentence lists- open-set sentence tests similar to HINT K-CID sentence test: Korean version of the CID test NVA speech recognition: Nederlands Vereninging Audiologie—Dutch open-set speech recognition tests consisting of monosyllabic (consonant-vowel-consonant) word lists NU-6 words: Northwestern University Auditory Test No. 6: open-set lists of monosyllabic words A cochlear implant (CI) is an electronic device which bypasses the auditory function of the outer, middle and inner ear and stimulates the cells of the spiral ganglion (and therefore the cochlear nerve) directly. Cochlear implants are indicated in severe and profound sensorineural hearing loss (SNHL) when hearing aids provide insufficient information for understanding speech. Cochlear implantation was first performed in New Zealand in 1986.1 An initially small national programme was based in Auckland. Cochlear implants were first performed in Christchurch in 1998. The Southern Cochlear Implant Programme (SCIP) was established in 2003. Located in Christchurch it serves the South Island and lower North Island. It is divided into separate Adult (SCIPA) and paediatric programmes. The SCIPA comprises two surgeons, two audiologists and a rehabilitationist. Potential candidates are referred to the programme, principally by audiologists and otolaryngologists, and undergo assessment by all three components of the CI team. Currently 80% of referrals are found to be suitable candidates. The key audiologic test performed is aided speech discrimination, details of which are provided below. This testing aims to determine how well the candidates can utilise their aided hearing to understand speech. The same testing processes are used post-implant to assess the effectiveness of the CI. Previous research has indicated that there are many factors contributing to the outcome of CI including age, duration of deafness/age at onset of deafness, anatomic issues as determined by radiology, other handicaps, speech and language abilities, functional hearing, family and social support, expectations/motivation, educational setting, communication mode, availability of support services, and the intensity of post-implant rehabilitation therapy.15,4 Many studies have shown a significant negative correlation with duration of deafness,4,7,9,11,12 and a positive correlation with age at onset of deafness, hearing aid use, and progressive hearing loss.7 In particular, hearing loss prior to development of speech and language inhibits its development and also reduces the effectiveness of cochlear implantation later on. There are differing findings in the literature regarding the time after CI that maximal benefit is obtained. Some studies indicate that this is reached as early as 3 months,10others suggest it is as late as 2 years.4 This is an important factor to be able to counsel patients about before and after their CI, so they will have some idea of their likely long-term outcome as they carry out their rehabilitation. In 2005 our Programme introduced the Nucleus® Freedom™ (Cochlear Ltd, Lane Cove, New South Wales, Australia) device which superseded the older technology from the same company of the Nucleus 22 and the later Nucleus 24 cochlear implants. Anecdotally we thought our patients were experiencing more rapid improvements in speech discrimination and slightly higher levels of peak performance, but wished to accurately investigate this. The objective of this study was to assess the outcome with respect to improvement in speech discrimination test results after CI in patients enrolled at SCIPA, and to determine how long after a CI there is a plateau in performance. We also aimed to assess the effect of duration of deafness and percentage of life spent deaf on results, and the effects of a new CI device. Methods A retrospective review of all files of patients receiving follow-up in Christchurch for CI between 1999 and Dec 2008 was undertaken. Exclusion criteria were patients under age 18 (these would be under the Paediatric Programme), and patients with no or inadequate pre-implant data. Some patients had CI performed by the Southern Cochlear Implant Programme - Adult, then moved to the Northern Cochlear Implant Programme (NCIP) for follow-up. Most of these were included, as follow-up data was sent from the NCIP. Some patients had received their CI in NCIP and then moved to the SCIPA for follow-up. These were included if we had a copy of their pre-implant data sent from NCIP. Files were obtained and the following were recorded: age, sex, type of implant: Nucleus® Freedom™ versus older devices - Nucleus 22 and Nucleus 24, age of onset of deafness, age at implant, percentage of life spent deaf (calculated from above), cause of deafness where known, and speech discrimination test results. We originally intended to measure both age at onset of hearing loss and age when hearing aids were first fitted, along with which sides were aided. Due to lack of consistent information we were only able to estimate age of onset of deafness with any degree of accuracy. Speech discrimination was measured using the hearing in noise test (HINT), which is a standardized test of speech recognition using sentences delivered in a sound field of 55 dB HL sound pressure level. In determining cochlear implant candidacy, HINT is performed without background noise, despite its name. HINT consists of 25 equivalent 10-sentence lists, two lists are used at each assessment. The tests are played on a DVD with a speaker/amplifier and are tested both with audiovisual (AV) cues (watching the presenter read the lists whilst listening) and with only auditory (A) cues (visual display turned off). The patient is scored for the percentage of words correctly identified on the lists. This is carried out pre-implant with their hearing aid on, at switch-on (s/o) of CI, then at 1 month, 3 months, 6 months, 9 months, 12 months, 18 months and 3 years post-implant where available. The following were assessed: improvement in HINT A and AV over time post-CI, overall improvement at 9, 12 and 18 months, characteristics of patients lost to follow-up or with missing data, relationship between final outcome and duration of deafness, percentage of life spent deaf and type of implant used. Statistical methods: All analyses were carried out using SAS 9.1. To investigate the time course of recovery up to 9 months after an implant, measures A and AV were analysed separately using repeated measures analysis of variance (PROC ANOVA) with time as the repeated measure (pre-implant, s/o, 1 month, 3 months, 6 months and 9 months), and a between subjects factor based on division of the patients into approximately three equal-sized groups based on their pro-operation scores. The Huynh-Feldt adjustment for non-sphericity was used for repeated measures sources of variance. Chi-square tests and t-tests were used to compare those who had complete data up to 9 months and those who did not on pre-implant variables. Prediction of long-term outcome used data from 18 months if available and otherwise the 12 month values were used. Including 12 month values in this way increased the number of A outcome observations from 40 to 46 and the number of AV observations from 37 to 44. Prediction of outcome was carried out using one way analysis of variance (PROC GLM). Because of ceiling effects, particularly for AV, Welch’s test was used if Levene’s test for the homogeneity of variance was significant with α=0.10, although this precaution mostly made little difference. Results A total of 171 patients were on the cochlear implant rehabilitation files at SCIPA. Of this total, 57 were excluded for the following reasons: 11 had been implanted too recently for analysis (2 months or less), 12 had transferred from the Paediatric Programme so did not apply to our study, 17 had been implanted prior to 1997 so speech discrimination tests other than HINT had been used for their pre-implant and post-implant assessments, nine had no follow-up notes available, four were non-users (discussed later), and four did not have pre-implant speech discrimination tests available. A total of 114 patients were then left for analysis. However, among this group there was much missing data. Only 55 had complete data for A and 53 for AV. Comparison of those with complete data and the remaining patients showed no significant differences in any of the background characteristics or of pre-implant functioning. We used complete data only for the analyses as imputation of data was deemed inappropriate due to the ceiling effect of the results. This was to ensure internal validity, i.e. comparing people across time for the same people. Pre-implant characteristics are shown in Tables 1-3, and the pre-implant scores on HINT are shown in Tables 4 and 5. Table 1. Characteristics of those with a cochlear implant between 1999 and 2008 Variables Number assessed Mean (SD) Range Age (years) Age (years) of onset of deafness Duration of deafness (years) Percentage of life deaf HINT pre-implant A HINT pre-implant AV 113 111 111 111 112 111 50 (15.3) 21 (19.4) 30 (14.5) 64% (30.5) 15 (16.8) 67 (23.9) 20-83 1-63 1-63 2-99 0-65 0-100 Table 2. Mean HINT AV scores (% correct) at pre-implant, switch-on (s/o), and 1-9 months post-implant: 53 patients Variables Mean HINT AV score SD Range Pre-implant s/o 1 month 3 months 6 months 9 months 66 82 93 96 96 98 24 17 10 9 6 4 0-100 32-100 67-100 57-100 74-100 79-100 Table 3. Mean HINT A scores (% correct) among those with complete data up to 9 months by grouping based on pre-implant scores (55 patients) HINT A scores pre-implant Number Mean pre-implant score (SD) Mean s/o score (SD) Mean score at 1 month (SD) Mean score at 3 months (SD) Mean score at 6 months (SD) Mean score at 9 months (SD) 0 1-15 16+ 18 17 20 0 (0) 6 (4) 34 (12) 33 (27) 23 (28) 46 (29) 60 (27) 44 (31) 69 (29) 71 (29) 53 (33) 79 (22) 77 (27) 65 (31) 82 (24) 83 (21) 69 (29) 83 (25) Total 55 14 (17) 34 (29) 58 (30) 68 (30) 75 (28) 79 (25) Table 4. Mean HINT AV scores (% correct) among those with complete data up to 9 months by grouping based on pre-implant scores (53 patients) HINT AV scores pre-implant Number Mean pre-implant score (SD) Mean s/o score (SD) Mean score at 1 month (SD) Mean score at 3 months (SD) Mean score at 6 months (SD) Mean score at 9 months (SD) 0-60 61-78 79+