Statistical Methods for Health Care Research

Mandarin Chinese speech recognition by pediatric cochlear implant users

Objective: Development of the Mandarin Chinese matrix (CMNmatrix) sentence test for speech intelligibility measurements in noise according to the international standard procedure.Design: A 50-word ...

Construction and evaluation of the Mandarin Chinese matrix (CMNmatrix) sentence test for the assessment of speech recognition in noise.

To investigate the outcomes of cochlear implantation in patients with neurofibromatosis type 2 (NF2), and to discuss the current management strategy for NF2 patients. The medical records of NF2 patients who received cochlear implants (CI) at our center between 2012 and 2016 were retrospectively reviewed. Pre-operative hearing status, tumor status, treatment of tumors, and auditory outcomes post-implantation were evaluated. Twelve patients were included in the study. Five were implanted with the tumor in situ; two of them received radiotherapy pre-implantation, and three were implanted without any previous treatment. Four patients were implanted simultaneously with tumor removal. Three patients were implanted as second-stage after failed hearing preservation surgery. The mean pure tone audiometry with the implant was 44 dB (range 25–80 dB) and the mean sentence recognition score (SRS) in a quiet environment without lip reading was 63% (range 0–97%). A poorer objective auditory outcome was identified in one patient who showed no response to electrical promontory stimulation (EPS), but the sound perception was still helpful. In total, 11 of 12 (91%) patients were daily users, and the other patient used the implant as a “sleeper” device due to its interference with contralateral hearing provided by a hearing aid. CI is an effective option in auditory rehabilitation and should be considered primarily for NF2 patients with intact cochlear nerve. EPS might be a predictor for cochlear implant performance. Good contralateral hearing may present a barrier to daily use.

Impact of cochlear implantation on the management strategy of patients with neurofibromatosis type 2

Importance The bone-anchored hearing device (BAHD) was not introduced in China until 2010. To our knowledge, this is the first study to assess the efficacy of Chinese Mandarin-speaking patients with bilateral aural atresia. Objective To evaluate the speech recognition of Chinese Mandarin-speaking patients with BAHDs as well as patients’ satisfaction using 2 questionnaires. Design, Setting, and Participants A retrospective case review of 16 patients with bilateral aural atresia conducted at a tertiary referral center. Intervention A BAHD was implanted during auricle reconstruction surgery or after the auricle was rebuilt. A surgical method to combine the BAHD implantation with the second stage of ear reconstruction was introduced. Main Outcomes and Measures Speech audiometry test and mean pure-tone threshold results were compared among patients with unaided hearing and those with BAHDs. Scores from the BAHD user questionnaire and Glasgow Children’s Benefit Inventory (GCBI) were used to measure patients’ satisfaction and subjective health benefit. Results The mean (SD) speech discrimination scores measured in a sound field with a presentation level of 45 dB HL (hearing level) were 6.7% (7.4%) unaided and 86.5% (4.4%) with a BAHD. Scores with a presentation level of 65 dB HL were 56.5% (7.4%) unaided and 90.1% (3.4%) with a BAHD. The speech reception threshold was 60.6 (7.5) dB HL unaided and 24.7 (5.0) dB HL with a BAHD. The mean (SD) pure-tone threshold of the patients was 61.6 (7.8) dB HL unaided and 23.8 (5.9) dB HL with a BAHD. The BAHD application questionnaire demonstrated excellent patient satisfaction. The mean (SD) benefit score of GCBI was 45.6 (14.4). Conclusions and Relevance For aural atresia, the BAHD has been one of the most reliable methods of auditory rehabilitation. It can improve the patient’s word recognition performance and quality of life. The technique of BAHD implantation combined with auricular reconstruction in a 2-stages-in-1 surgery and the modified incision of patients with reconstructed auricle proved to be safe and effective.

https://jamanetwork.com/journals/jamaotolaryngology/articlepdf/1816004/ooi130128.pdf

The Efficacy of Unilateral Bone-Anchored Hearing Devices in Chinese Mandarin-Speaking Patients With Bilateral Aural Atresia

Objective
To edit the sentence list for mandarin speech test materials （MSTM） and to test their equivalence in order to use the list as routine clinical speech tests for evaluation of cochlear implants and hearing aids in different stages of the intervention and rehabilitation. 


Methods
Firstly, two groups of professionals were set up ： one for establishing the principles of the edit, another for experiments. The criteria to guide the selection of spoken sentences were as following：①suitable for everyday speech audiometry, esp. clinically usage;② a normal sampling of everyday spoken sentences in mandarin. Totally, 300 sentences were selected in 30 lists for the experiment. The record cassette was used by a male speaker. The college students （54 persons, 54 ears） with normal hearing participated in the equivalence test. Two statistic methods were used for equivalence analysis. 


Results
Through speech tests and statistic analysis, 2 sets with 29 lists （290 sentences） were selected as our equivalent lists for further evaluation and clinical trial. 


Conclusions
By setting up the strict criteria of selecting sentences, speech test method and statistic analysis, 2 sets of 29 lists （290） sentences were chosen as our speech evaluation materials, but it was necessary for further analysis in different laboratories and clinics. The sentence lists should also get used for hearing impaired subjects, hearing aid users and cochlear implant recipients.


Key words: 
Speech audiometry; Chinese; Mandarin; Equivalence

Edit and evaluation of Mandarin sentence materials for Chinese speech audiometry

Objective
To edit the spondee （disyllable） word lists as a part of mandarin speech test materials （MSTM）. These will be basic speech materials for routine tests in clinics and laboratories. 


Methods
Two groups of professionals （audiologists, Chinese and Mandarin scientists, linguistician and statistician） were set up at first. The editing principles were established after 3 round table meetings. Ten spondee lists, each with 50 words, were edited and recorded into cassettes. All lists were phonemically balanced （3-dimensions： vowels, consonants and Chinese tones）. Seventy-three normal hearing college students were tested. The speech was presented by earphone monaurally. Three statistic methods were used for equivalent analysis. 


Results
Related analysis showed that all lists were much related, except List 5. Cluster analysis showed that all ten lists could be classified as two groups. But Kappa test showed that the lists＇ homogeneity were not well. 


Conclusions
Spondee lists are one of the most routine speech test materials. Their editing, recording and equivalent evaluation are affected by many factors. This also needs multi-discipline cooperation. All lists edited in present study need future modification in recording and testing in order to be used clinically and in research. The phonemic balance should be kept.


Key words: 
Audiometry, speech; Cochlear implantation; Linguistics; Vocabulary; Rehabilitation of speech and language disorders

Development and equivalence evaluation of spondee lists of mandarin speech test materials

Objective After preliminary edit and equivalent tests, Mandarin speech test materials (MSTM) were digitally recorded and further tested in order to get multi-center analysis and clinical application. Method The contents of the lists were modified and the trial lists were added. All materials were digitally recorded into 3 CDs with international criterions. Each CD contained a calibration pure-tone, a trial list and the speech test lists. MSTM contained 10 lists of monosyllabic words with 50 words in each list, 10 lists of spondees with 50 disyllabic words (100 Chinese characters) in each lists, and 30 lists of sentences with 10 sentences in each list. Sixty normal-hearing college students were tested. The tests were conducted in the sound booths with earphones. The presentation levels of the monosyllables, disyllables and sentences were different. The presentation levels considered as the scores were about 50% correct,in which condition the speech intelligibility was most sensitive. Result Statistic analysis showed that two sets (9 lists in each set) of monosyllable lists, 9 disyllable lists and 27 sentence lists were equivalent in difficulty. If the clinicians do not need so many lists to test the patients, they had better use the most equivalent ones. Conclusion The equivalent Mandarin speech lists have been made into CDs and ready to be used clinically. But the multi-center evaluation is needed.

Digital recording and equivalent evaluation of Mandarin speech test materials

Objective To evaluate the equivalence of difficulty of disyllabic Mandarin Speech Test Materials (MSTM) in order to facilitate wider use of speech audiometry in Chinese audiology clinics. Method The Mandarin disyllabic materials, including 10 lists with 50 words in each, have been recorded and converted digitally into CD format. Sixty-five subjects (the age-range 18-25 years) with normal hearing who speak Mandarin well in their daily lives participated in this study. Five of them joined the pilot study, which aimed to identify a presentation level which would be used into the formal test. The other sixty subjects participated in the formal speech perception test. The results were analyzed statistically with Statistical Package for the Social Sciences Software Version 11.0 (SPSS 11.0). Result Nine of the 10 lists were found to have equal difficulty with application of the ANOVA and SNK-Q statistical methods. Conclusion Nine lists of 50 words of these Mandarin disyllabic materials can be considered to have equal difficulty and could be used in a clinical setting.

Equivalence analysis of disyllabic Mandarin speech test materials

Objective: To develop and evaluate the Mandarin Quick Speech-in-Noise (M-Quick SIN) Test materials in mainland China.
Design: Four parts were included in the experiment to (1) develop sentence materials and select equivalent sentences, (2) evaluate the reliability of the lists we grouped afterwards, (3) discuss the formula of SNR loss fitted for M-Quick SIN, and (4) quantify the classification of SNR loss among normal-hearing and hearing-impaired people. 132 normal-hearing and 30 hearing-impaired subjects were participated in the experiment.
Results: A 300 sentence corpus was established and 78 sentences with better homogeneity were selected from it. After the equivalence and the test-retest reliability was established for the group materials, 11 equivalent lists for research and clinical use were chosen. The SNR-50 value for these sentences was -2 dB for normal-hearing people, and the formula was defined as “SNR loss=24.5-correct words”. The classification of SNR loss was preliminarily quantified as: normal (≤ -2 dB), mild (-2 to 10 dB), moderate (10 to 20 dB) and severe (≥ 20 dB).
Conclusions: The M-Quick SIN test provided us 11 equivalent test lists (each list had 6 sentences and 30 key words) and 2 practice lists for testing normal-hearing and hearing-impaired people. The normal value of SNR-50 was -2 dB SNR, and the 6 SNRs: 20, 15, 10, 5, 0, -5 dB SNR were determined to test SNR loss for the M-Quick SIN.

https://www.longdom.org/open-access/development-and-evaluation-of-the-mandarin-quick-speechinnoise-testmaterials-in-mainland-china-2471-9455-1000124.pdf

Jing Chen

Papers

Edit and evaluation of Mandarin sentence materials for Chinese speech audiometry

Development and equivalence evaluation of spondee lists of mandarin speech test materials

Digital recording and equivalent evaluation of Mandarin speech test materials

Equivalence analysis of disyllabic Mandarin speech test materials

Development and Evaluation of the Mandarin Quick Speech-in-Noise Test Materials in Mainland China