Voice Recordings Research Articles

Digital voice-based biomarker for monitoring respiratory quality of life: findings from the colive voice study

Regular monitoring of respiratory quality of life (RQoL) is essential in respiratory healthcare, facilitating prompt diagnosis and tailored treatment for chronic respiratory diseases. Voice alterations resulting from respiratory conditions create unique audio signatures that can potentially be utilized for disease screening or monitoring. Analyzing data from 1908 participants from the Colive Voice study, which collects standardized voice recordings alongside comprehensive demographic, epidemiological, and patient-reported outcome data, we evaluated various strategies to estimate RQoL from voice, including handcrafted acoustic features, standard acoustic feature sets, and advanced deep audio embeddings derived from pretrained convolutional neural networks. We compared models using clinical features alone, voice features alone, and a combination of both. The multimodal model combining clinical and voice features demonstrated the best performance, achieving an accuracy of 70.8% and an area under the receiver operating characteristic curve (AUROC) of 0.77; an improvement of over 5% in terms of accuracy and 7% in terms of AUROC compared to model utilizing voice features alone. Incorporating vocal biomarkers significantly enhanced the predictive capacity of clinical variables across all acoustic feature types, with a net classification improvement (NRI) of up to 0.19. Our digital voice-based biomarker is capable of accurately predicting RQoL, either as an alternative to or in conjunction with clinical measures, and could be used to facilitate rapid screening and remote monitoring of respiratory health status.

Can Voice Parameters Provide Cutoff Values to Predict Dysphagia in Individuals with Multiple Sclerosis?

In dysphagia assessment, along with well-defined measurements and signs, voice parameters can potentially support clinical decision as a marker, but more evidence is needed. This study aims to determine the voice parameters that can predict the risk of dysphagia and to determine optimal cutoff values in individuals with multiple sclerosis (IwMS). Seventy-six adults participated in the study, including 39 IwMS and 37 healthy individuals (HI). The study used the Dysphagia in Multiple Sclerosis Questionnaire (DYMUS), Gugging Swallowing Screen (GUSS), and Voice Handicap Index (VHI-10) and recorded voice samples using Praat programme. Voice recordings were taken pre- and post-swallowing. The voice parameters analysed are fundamental frequency (F0), standard deviation F0 (SD F0), jitter (local), shimmer (local), and harmonic-to-noise ratio (HNR). Roc analysis was performed to examine the diagnostic accuracy performance of the risk for dysphagia/penetration. The parameters of IwMS pre-swallowing differed significantly from those of HI on the VHI-10, DYMUS, GUSS scores, and jitter (local), shimmer (local), and HNR. IwMS but not HI exhibited significant differences in shimmer (local) and HNR between the pre- and post-swallowing measurements. In IwMS, GUSS revealed significant differences in shimmer (local) pre- and post-swallowing between the groups with and without dysphagia/penetration. In the ROC analysis results, the area under the curve (AUC) for shimmer (local) pre-swallowing was 73.1% (cutoff = 1.69); post-swallowing, it was 78.6% (cutoff = 1.57). In conclusion, IwMS can be associated with differences in shimmer (local) and HNR parameters, low quality of life-related to voice, and dysphagia/penetration risk. The AUC values for shimmer (local) in IwMS pre- and post-swallowing may help to strengthen diagnostic decisions of dysphagia risk.

Women's experiences in the aftermath of the Kahramanmaraş earthquakes in Türkiye: A phenomenological study

This study aimed to understand women's experiences related to health problems and needs after the 2023 Kahramanmaraş earthquake, considered the “disaster of the century.” This study employed phenomenology, a qualitative research methodology, and used criterion sampling, a type of purposive sampling. The participants included women who had experienced both the major Kahramanmaraş earthquakes; whose houses were damaged in the earthquakes; who had to live in tents, cars, or other shelters; and who had been in the earthquake zone for about a month. Semi-structured interviews were used as a data collection tool to elicit women's thoughts and experiences after the earthquakes. The interviews were conducted face-to-face in a calm environment by taking the necessary precautions and recording the audio with the participants' written consent. The interviews lasted at least 25 min and at most 45 min. For evaluating the data, the researchers used descriptive and content analyses. After transcribing the voice recordings, the researchers created the code categories and themes. The study identified four themes: “burden of responsibility,” “security,” “care needs,” and “male dominance.” Using participants' statements directly in the findings increased the study's internal validity. The study's findings indicate that women continued to fulfill their social, childcare, and family care roles after the earthquakes. The findings suggest that during this period, women faced housing and security problems, their special needs were ignored, their needs were affected by male dominance, and they were subjected to physical and psychological violence.

Challenges experienced by nurses regarding the involvement of guardians in integrated management of childhood illnesses in Vhembe District

The involvement of families and communities in healthcare services by healthcare providers enhances accountability towards the community they serve. Guardians need to be involved in the integrated management of childhood illnesses (IMCI) to improve child health outcomes. The study aimed to explore challenges experienced by nurses regarding the involvement of guardians in the management of childhood illnesses in the Vhembe district. A qualitative, exploratory-descriptive design was followed in this study. Fifteen participants were purposefully selected for the study. One-on-one interviews following a semi-structured interview guide were conducted face-to-face with the participants. A voice recorder was used to record the interviews. Data analysis was done following Tesch's eight steps. Trustworthiness was ensured throughout the study. The study yielded two themes, namely: guardians-related challenges, which are lack of knowledge and understanding of IMCI, language barrier and lack of continuity of care and healthcare service-related challenges, which are lack of human resource, staff attitude towards the IMCI strategy, shortage of material resource and erratic water supply. For the strategy to be effective in reducing the complications of childhood illnesses and child mortality, the government needs to avail the required resources and develop measures to empower guardians in child health management.

Linguistic Signs of The Names of Leaves Used by Sikerei in Maileppet South Siberut Mentawai

This research is entitled " Linguistic Signs of the Names of Leaves Used by Sikerei in Mailepppet, South Siberut, Mentawai. This research is limited to several main ideas related to the observation of the following two questions: What are the signifier and signified of the leaves used by Sikerei in rituals in Maileppet, South Siberut, Mentawai. To explain the signifier and signified of the meaning of the leaves used by Sikerei in rituals in Maileppet, South Siberut, Mentawai. The theory used comes from Saussure's theory and Barthes' theory which is used to dissect and get answers to the objectives of this research. For the data analysis method, the author uses qualitative research by emphasizing meaning and significance in certain conditions. This research uses data collection techniques in the form of field observations, interviews, notes, voice recordings, photos. The conclusion of this research is the meaning of the leaves used by Sikerei in rituals in Maileppet, South Siberut, Mentawai. The arat sabulungan belief means belief in objects that have the spirit of the cultural heritage of the Mentawai people's ancestors. This belief is the same as the unwritten norms or rules for the Mentawai people which guide the daily life of the Mentawai people, especially in Maileppet, South Siberut.

Lexical associations in patients with neurodegenerative disorders. Two case studies: A patient with spinocerebellar ataxia type 1 (SCA-1) and a patient with Parkinson’s disease (PD

This article presents case studies of the relative ability of two subjects (OB1 and OB2) to make lexical associations; they have spinocerebellar ataxia type 1 and Parkinson’s disease, respectively. Test method: subjects gave associations (orally, with no time limit) to 63 verbal stimuli presented on a computer screen. Verbal stimuli were selected from Gatkowska’s experimental study (Gatkowska 2017). Responses were recorded on an Olympus 650 digital voice recorder, then analysed linguistically and psychologically. Results: psychological analysis – subjects within intellectual norm (Raven’s Matrices test), dysarthric disorder, difficulty in implementing instructions. Linguistic analysis: OB1 – predominance of paradigmatic relations (hyponymy, meronymy, synonymy) over syntagmatic relations (diffi culty building context), lack of causality. OB2 – richer paradigmatic relations (hyponymy, meronymy, synonymy, antonymy, complementarity), syntagmatic relations diverse. Conclusions: possible complex background of language disorders (executive, emotional), higher level of language functioning of OB2 compared to OB1. The study illustrates the need for an interdisciplinary and holistic approach in diagnostic and therapeutic management.

Voice as a Biomarker of Pediatric Health: A Scoping Review.

The human voice has the potential to serve as a valuable biomarker for the early detection, diagnosis, and monitoring of pediatric conditions. This scoping review synthesizes the current knowledge on the application of artificial intelligence (AI) in analyzing pediatric voice as a biomarker for health. The included studies featured voice recordings from pediatric populations aged 0-17 years, utilized feature extraction methods, and analyzed pathological biomarkers using AI models. Data from 62 studies were extracted, encompassing study and participant characteristics, recording sources, feature extraction methods, and AI models. Data from 39 models across 35 studies were evaluated for accuracy, sensitivity, and specificity. The review showed a global representation of pediatric voice studies, with a focus on developmental, respiratory, speech, and language conditions. The most frequently studied conditions were autism spectrum disorder, intellectual disabilities, asphyxia, and asthma. Mel-Frequency Cepstral Coefficients were the most utilized feature extraction method, while Support Vector Machines were the predominant AI model. The analysis of pediatric voice using AI demonstrates promise as a non-invasive, cost-effective biomarker for a broad spectrum of pediatric conditions. Further research is necessary to standardize the feature extraction methods and AI models utilized for the evaluation of pediatric voice as a biomarker for health. Standardization has significant potential to enhance the accuracy and applicability of these tools in clinical settings across a variety of conditions and voice recording types. Further development of this field has enormous potential for the creation of innovative diagnostic tools and interventions for pediatric populations globally.

The Use of Machine Learning in Eye Tracking Studies in Medical Imaging: A Review.

Machine learning (ML) has revolutionized medical image-based diagnostics. In this review, we cover a rapidly emerging field that can be potentially significantly impacted by ML - eye tracking in medical imaging. The review investigates the clinical, algorithmic, and hardware properties of the existing studies. In particular, it evaluates 1) the type of eye-tracking equipment used and how the equipment aligns with study aims; 2) the software required to record and process eye-tracking data, which often requires user interface development, and controller command and voice recording; 3) the ML methodology utilized depending on the anatomy of interest, gaze data representation, and target clinical application. The review concludes with a summary of recommendations for future studies, and confirms that the inclusion of gaze data broadens the ML applicability in Radiology from computer-aided diagnosis (CAD) to gaze-based image annotation, physicians' error detection, fatigue recognition, and other areas of potentially high research and clinical impact.

Parkinson Disease Detection Using CNN Algorithm

Abstract: Parkinson's disease, a progressive neurological disorder, results from the depletion of dopamine-producing neurons in the brain, leading to diminished motor function. Common symptoms include tremors, rigidity, bradykinesia, shivering, and impaired balance. This study introduces two neural network architectures: the Voice Impairment Classifier, designed for early disease detection. The research conducted a thorough assessment of convolutional neural networks (CNNs) for classifying gait signals transformed into spectrogram images, and deep dense networks for analyzing voice recordings. Results demonstrated superior performance of the proposed models, with the VGFR Spectrogram Detector achieving 88.1% accuracy and the Voice Impairment Classifier achieving 89.15% accuracy, surpassing current state-of-the-art techniques.

Women’s emotional roller coasters during pregnancy as a consequence of infertility: a qualitative phenomenological study

Women experiencing infertility tend to experience distinct emotions during pregnancy following infertility. Individuals in this population often struggle with psychological and social challenges during pregnancy. This qualitative phenomenological study determines how infertility experiences affect mothers’ mental well-being during pregnancy as well as their experiences and emotions after becoming pregnant. A face-to-face, in-depth interview with 15 women with infertility was conducted between November 2022 and May 2023. The study adhered to ethical guidelines, with informed written consent obtained prior to interviews and voice recordings. The COREQ checklist is applied to follow the EQUATOR guidelines for reporting research and the data were coded using MAXQDA.20 software. A thematic analysis revealed four main themes and 15 codes. Main themes included “Over Emotional Burden,” “Overprotection,” “Overthinking,” and “Social Activity Changes.” Pregnancy after infertility is a unique and emotionally charged experience for women, encompassing a spectrum of feelings that can be difficult to express. The importance of caring for these women and receiving support from their partners, families, and healthcare providers should not be overstated. Healthcare providers should be aware of these emotional challenges so they can provide better support and counselling to improve women’s overall pregnancy experience. It is believed that empathetic communication and tailored support can significantly improve the psychological well-being of this population. Psychologists, psychiatrists, and midwives should also be more attentive to mothers’ emotional challenges and integrate comprehensive emotional support and provide coping mechanisms in perinatal care programs.

Analysis of Students Phonological Errors in Pronouncing Vowel Sounds A Case Study of English Learners in Indonesia

The purpose of this study was to find phonological errors that students make when they pronounce vowel diphthongs. Researchers also asked what type of vocal diphthong sound students most often said in tenth grade. Grade X students of the 2021/2022 school year are the subjects of this study. This study used a quantitative descriptive approach. Researchers randomly selected 60 samples from 395 individuals. The tool used is the English Vowel Diphthong word pronunciation test, which includes eight types of English vowel diphthongs with two focus words for each type. Furthermore, a voice recorder is employed to capture the students' pronunciation. Findings showed that students made 339 word mistakes. The pronunciation of X graders shows the most frequent types of vowel diphthong errors: Diphhongs [ɔʊ] with the focus word "although" experience 72% of total errors, Diphhongs [ʊə] with the focus word "during" experience 65% of total errors, and Diphhongs [aɪ] with the focus word "sacrifice" experience 61% of total errors.

Pediatric nurses' fatigue descriptions in occupational injury reports: A descriptive qualitative study1.

Despite the high risks associated with occupational fatigue in healthcare, few organizations require nurses to screen and report fatigue symptoms. As a result, little is known about if and how nurses would report fatigue while on the job. To determine if hospital-based pediatric nurses reported fatigue as part of an active injury reporting method. This secondary analysis of qualitative data used a descriptive design with content analysis. Data from the parent study were collected at a U.S. pediatric hospital where nurses verbally reported on-shift injuries or near misses and pre- and post-shift health status via a digital voice recorder. Researchers used content analysis to independently code data for nurses' references to fatigue. Codes were then analyzed for patterns and themes. Approximately 30% (n = 104) of participants reported fatigue-related content. Emergent themes were Work Stressors, Individual Risk Factors, Fatigue Descriptors, Adverse Outcomes from Fatigue, Fatigue Buffers, Descriptors for Buffered Fatigue, and Favorable Outcomes from Buffered Fatigue. Fatigue descriptions align with prior literature, demonstrating the accuracy of the voice recorder data collection method. In addition, nurses expressed uncertainty about the appropriateness of reporting fatigue symptoms. Findings from this study support pediatric nurses will report fatigue, when provided an opportunity. Additional efforts are needed to better understand effective ways to improve fatigue reporting among nurses, including harnessing current technology for real-time reporting and how to change the culture around fatigue reporting.

A qualitative exploration of patients' perception regarding the comprehensive dental services availed at a primary health center.

Comprehensive oral care is a service centered around the patient, and individuals who need it the most often face limited access. Patient perception acts as a guide for enhancing quality, ensuring patients' future intent to utilize the services and facilitating recommendations to others. The present study aimed to assess the patients' perception of comprehensive dental services availed at a Primary Health Center (PHC). This qualitative study was based on a phenomenological interpretive approach, and judgment sampling method was employed. A validated interview guide, developed from relevant literature was employed in the local language to conduct interviews among adults visiting the PHC, gathering their views regarding the services provided. The interviews were audio recorded on a digital voice recorder, and files were password protected. Content saturation guided the determination of the final number of participants interviewed. After translating and transcribing the interviews, thematic analysis and coding were performed using ATLAS. ti 23 for Windows. A total of 12 participants were included in the study, following data saturation. Among them, there were 8(66.7%) female and 4(33.3%) male participants. Ten overarching main themes were discerned through the assigned codes, including positive views, neutral views, negative views, previous dental clinics visited, previous experience with dental treatment, treatments sought at the center, referrals, source of information about the dental center, subsequent visits and suggestions for improvement. The findings of this study revealed a positive patient perception of the comprehensive dental services offered at the PHC. Through insightful interviews, various strengths, and areas for improvement regarding the center and care provision were identified. These insights provide valuable suggestions that can be applied to elevate the utilization of dental services, ensuring continuous improvement in patient care.

Streamlining quantification of age‐related vocal atrophy

AbstractObjectivesAge‐related vocal atrophy (ARVA) has a significant impact on voice, communication, and quality of life. Vocal folds are bowed with incomplete glottic closure during phonation. Efficient quantification of vocal fold atrophy and collection of voice recordings in clinic remains challenging. The primary focus of this study is to describe a novel method for quantifying vocal atrophy and obtaining voice recordings in clinic among patients with ARVA.MethodsPatients with ARVA were included. Voice recordings were collected during the clinic visit, and acoustic analysis was subsequently performed. A novel mobile application was used to quantify the bowing index (BI).ResultsThe study included 10 patients with ARVA, with a mean age of 72.7 ± 6.8 years and body mass index (BMI) of 24.9 ± 2.4 kg/m2. Calculation of BI was feasible with a mean of 9.9 ± 1.8 units. On average, the audio recording took 2.6 ± 0.4 min, and subsequent analysis required 7.1 ± 1.8 min. Mean continuous speech f0 was 212.1 ± 10.1 and 134.2 ± 31.5 Hz for male and female patients, respectively. Smoothed cepstral peak prominence was 8.9 ± 1.5 dB (male) and 8.5 ± 0.3 dB (female), and maximum phonation time between male and female patients was 16.7 ± 9.8 and 13.8 ± 1.9 s, respectively.ConclusionWe present a feasible and streamlined method for quantification of vocal fold atrophy in the clinic among patients with ARVA. The accuracy and reliability of this new method are areas of ongoing investigation. Quantification of vocal atrophy may help with clinical decisions, including diagnosing vocal atrophy and tracking treatment progress. Moreover, this method may improve research data acquisition without burdening patients and clinicians with additional time‐consuming tasks.Level of evidence: 4.

A Scoping Review of Artificial Intelligence Detection of Voice Pathology: Challenges and Opportunities.

Survey the current literature on artificial intelligence (AI) applications for detecting and classifying vocal pathology using voice recordings, and identify challenges and opportunities for advancing the field forward. PubMed, EMBASE, CINAHL, and Scopus databases. A comprehensive literature search was performed following the Preferred Reporting Items for Systematic Reviews and Meta-analysesExtension for Scoping Reviews guidelines. Peer-reviewed journal articles in the English language were included if they used an AI approach to detect or classify pathological voices using voice recordings from patients diagnosed with vocal pathologies. Eighty-two studies were included in the review between the years 2000 and 2023, with an increase in publication rate from one study per year in 2012 to 10 per year in 2022. Seventy-two studies (88%) were aimed at detecting the presence of voice pathology, 24 (29%) at classifying the type of voice pathology present, and 4 (5%) at assessing pathological voice using the Grade, Roughness, Breathiness, Asthenia, and Strainscale. Thirty-six databases were used to collect and analyze speech samples. Fourteen articles (17%) did not provide information about their AI model validation methodology. Zero studies moved beyond the preclinical and offline AI model development stages. Zero studies specified following a reporting guideline for AI research. There is rising interest in the potential of AI technology to aid the detection and classification of voice pathology. Three challenges-and areas of opportunities-for advancing this research are heterogeneity of databases, lack of clinical validation studies, and inconsistent reporting.

Challenges to Compliance and Adherence to Personal Hygiene and Food Safety Among Foodservice Workers in Hospitals of Mogalakwena Municipality, Limpopo Province

Background: Foodborne illnesses could result from food handlers acquiring bacteria that make them dangerous for humans to consume. Non-adherence to personal hygiene and food safety can impact how well patients' conditions are managed and aggravate illnesses during hospital stay. Therefore, this study seeks to explore factors that may impact adherence to personal hygiene and food safety among hospital food service workers. Methods: Qualitative exploratory study design was used to extract data from food handlers at Mokopane and Voortrekker hospitals in Mogalakwena municipality of Limpopo province, South Africa. Twenty-one participants were purposively sampled, and face-to-face interviews were conducted using voice recorders and field notes for non-verbal cues. Tesch’s eight steps, inductive, descriptive, and open coding techniques were used to analyse data. Results: Factors reportedly affecting personal hygiene and food safety include knowledge of the importance of adherence to personal hygiene and food safety, personal preferences and attitude towards personal hygiene, and lack of uniform and inspection. Non-adherence is affected by cultural norms regarding the wearing of wedding rings and bracelets on the wrist. Conclusion: This study recommends regular internal and external food safety compliance and monitoring. The leadership of the hospital's food service department should be strengthened, and food handlers should get regular in-service training.

Violence Against Women and their Children (VAWC) Cases During Pandemic: Experiences of Women and Children Protection Desk (WCPD) Officers in Cebu City

This study aimed to explore the life experiences of the Women and Children Desk (WCPD) handling Violence against Women and their Children (VAWC) cases during the pandemic in Cebu City. Specifically, sought to address the following specific problems: the experiences of the informants in handling VAWC cases during the Covid-19 pandemic, addressing the problems encountered by the informants, and aspirations of the informants to improve the quality of service to the victims of VAWC. The phenomenological and qualitative approach in this study was utilized through interviews as the key process in the gathering of data. There were ten (10) selected police officers handling Violence against Women and their Children (VAWC) during the Covid-19 pandemic in Cebu City. Five (5) of them were individually interviewed, while the other five (5) were included in the Focused Group Discussion (FGD). They were selected randomly but met the qualifications and have broad experience in the subject matter. The research locations are the top five (5) police stations in Cebu City with a high number of VAWC cases is Cebu City namely: Police Station (PS) 4- Mabolo; PS5- Carbon; PS7- Pardo; PS9- Guadalupe and PS10- Punta. A validated interview guide was aided by the voice recorder for the basis of transcribing the responses of the informants. Utilization of Colaizzi's method, a phenomenological analysis procedure, was used to analyze the responses of the informants during the interview. For the experiences of the informants in handling VAWC cases in Cebu City during a pandemic, themes created were Satisfaction with Desired Result, Attained Productivity, and Embracing Limitations of Official Functions in addressing the problems encountered in handling (VAWC) cases during pandemic, the following themes created were: Intensified Implementation of Restrictions, Established Online Presence, and Maintaining Effective Collaboration and for the aspirations of the informants to improve the quality of service to the victims of VAWC, the following theme were created: New Normal Retooling and Equipment for Stabilization. The Philippine National Police (PNP)- Women and Children Protection Desk (WCPD) officers, as a specialized unit in enforcing laws related to the protection of women and children and regarded as the “quick-response-unit,” must maintain dynamic and standardized public service. It is suggested that the Performance Standards and Assessment Tool for police service addressing cases of violence against women be amended where the assessment tool should not be done as “Self-administered” to identify committed officers who are working diligently to end violence against women and children. The retooling of PNP-WCPD should not only cover policy-related attitudes, habits, and work ethics but should also investigate an array of services that can be extended during a crisis. The provision of additional facilities and personnel to attend VAWC cases will expedite the acquisition of relevant documents for filing legal actions and will be timely in the recent health crisis. This will enable the Department of Health to provide prompt intervention on the outbreak of mental health problems aggravated by the pandemic. PNP-WCPD and other linkages should develop process-based alternatives and act on the research of shared resources from the Covid-19 experiences because health protocols altered policing strategies and are not conducive to problem-oriented crime prevention and innovation.

Sentry of the Cyberspace during Covid-19 Pandemic: Experiences of Philippine National Police Cyber Cops

This study aimed to explore the life experiences of cyber cops as sentries in cyberspace during the COVID-19 pandemic in RACU 7. Specifically, this sought to answer the following specific problems: experiences of the informants as cyber cops during the COVID-19 pandemic, coping with the challenges they encountered, and how they tried to solve the situation and aspirations of the informants to improve quality service. The phenomenological and qualitative approach in this study was utilized through interviews as the critical process in gathering data. There were ten (10) selected police officers with a cyber cop badge. They were chosen randomly, but they met the minimum qualifications. The research location is at Cebu PPO Compound, Cebu City, where RACU 7 holds the office. A validated interview guide aided by the voice recorder to transcribe the informants' responses was used. Utilization of Colaizzi's method, a phenomenological analysis procedure, was used to analyze the responses of the informants during the interview. Results revealed that in the informants' experiences as cyber cops during the COVID-19 pandemic, the themes generated were Performing Duties Amidst the Pandemic, Maintaining a Good Image, and Encountering Insufficiency of Personnel and Supplies. In coping with the challenges encountered and how they try to solve the situation, the following themes were created: Expanding Cyber Cops Strategies and Having the Cyber Cops Initiative. For the aspirations of the informants, themes identified were the adaptation of advanced technology and the expansion of organizational support. It was recommended that PNP-ACG design a program that would address issues in times of crisis like the pandemic and extend services prospective complainants could easily access. Moreover, the PNP organization should produce more well-trained cyber cops.

Identifying bias in models that detect vocal fold paralysis from audio recordings using explainable machine learning and clinician ratings.

Detecting voice disorders from voice recordings could allow for frequent, remote, and low-cost screening before costly clinical visits and a more invasive laryngoscopy examination. Our goals were to detect unilateral vocal fold paralysis (UVFP) from voice recordings using machine learning, to identify which acoustic variables were important for prediction to increase trust, and to determine model performance relative to clinician performance. Patients with confirmed UVFP through endoscopic examination (N = 77) and controls with normal voices matched for age and sex (N = 77) were included. Voice samples were elicited by reading the Rainbow Passage and sustaining phonation of the vowel "a". Four machine learning models of differing complexity were used. SHapley Additive exPlanations (SHAP) was used to identify important features. The highest median bootstrapped ROC AUC score was 0.87 and beat clinician's performance (range: 0.74-0.81) based on the recordings. Recording durations were different between UVFP recordings and controls due to how that data was originally processed when storing, which we can show can classify both groups. And counterintuitively, many UVFP recordings had higher intensity than controls, when UVFP patients tend to have weaker voices, revealing a dataset-specific bias which we mitigate in an additional analysis. We demonstrate that recording biases in audio duration and intensity created dataset-specific differences between patients and controls, which models used to improve classification. Furthermore, clinician's ratings provide further evidence that patients were over-projecting their voices and being recorded at a higher amplitude signal than controls. Interestingly, after matching audio duration and removing variables associated with intensity in order to mitigate the biases, the models were able to achieve a similar high performance. We provide a set of recommendations to avoid bias when building and evaluating machine learning models for screening in laryngology.

Acceptance and adoption of tele-rehabilitation by physiotherapists from Nigeria, a low resource setting: a mixed-method study

BackgroundAdoption of tele-rehabilitation among physiotherapists in developed countries has risen exponentially to complement conventional physical contact physiotherapy. However, many low- and middle-income countries like Nigeria where barriers to utilization of tele-physiotherapy were earlier identified still demonstrates limited evidence on its acceptance and adoption till date.MethodThis was a mixed-method design of quantitative and qualitative study, aimed at investigating acceptance and adoption of tele-rehabilitation among physiotherapists in Nigeria. Respondents comprised 331 physiotherapists practising in Nigeria and registered with the Medical Rehabilitation Therapists (Registration) Board of Nigeria (MRTB). Quantitative approach involved completion of a 33-item structured questionnaire shared as Google Forms to respondents’ WhatsApp and emails. Qualitative aspect involved in-depth interview with 12 physiotherapists at the clinics, using a 10-item semi-structured interview guide, and recorded with a voice recorder. Quantitative data was analysed using descriptive and inferential statistics at p < 0.05 alpha level. Qualitative data was analysed using thematic content analysis.ResultsRespondents comprised 172 males and 159 females. Their mean age and years of experience were 33.42 ± 8.95 and 8.86 ± 8.037 years, respectively. Pre-lockdown, only 62 (18.7%) respondents offered tele-physiotherapy. During the 35-day total lockdown, 114 (34.4%) offered tele-physiotherapy. Majority (155 (46.8%)) used smartphone apps. Fifty-seven (17.2%) respondents were confident that tele-rehabilitation can complement conventional physiotherapy, while 72 (21.8%) kept offering post-lockdown. Use of digital technology was highest for patient education (97 (29.3%)), especially for musculoskeletal conditions (85 (25.7%)). There were significant associations between pre-lockdown duties (p = 0.036), post-lockdown duties (p = 0.005) and use of digital health technology (p = 0.001) with respondents’ age. Significant associations also existed between respondents’ years of practice and each of post-lockdown duties (p = 0.017) and use of digital health technology (p = 0.001). Qualitative approach revealed that facilitators to adoption and acceptance of tele-rehabilitation included patient education, training of physiotherapists, availability of evidence-based guidelines and tele-health software and hardware systems. Barriers were lack of or limited technological infrastructures, finances, hospital management support, and knowledge of information and communication technology.ConclusionsPhysiotherapists in Nigeria demonstrated moderate acceptance but low adoption of tele-rehabilitation, in varied comparison with clinicians and physiotherapists in other climes. Training of physiotherapists in tele-rehabilitation and provision of enabling technological environment by regulatory bodies and policy makers are recommended.