RKRoulla Katiri

Roulla Katiri

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Roulla Katiri has a total of 71 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Validated Methods for Synthesising Hearing Health Data for Machine Learning: A Comparative Study of KDE and VAE Approaches

    by Liam Barrett, Roulla Katiri, Yuen Bing Ooi, Isabella Moffitt, Anne G. M. Schilder, Nishchay Mehta

    Hearing loss affects approximately 1.5 billion people globally, yet access to comprehensive audiometric datasets for research remains limited due to privacy constraints. Synthetic data generation offers a promising solution, enabling broader data sharing while preserving privacy. This study developed and validated two complementary approaches for synthesising audiometric data: Kernel Density Estimation (KDE) and Variational Autoencoders (VAE). Using the National Health and Nutrition Examination Survey (NHANES) dataset comprising 36,676 participants with comprehensive hearing assessments, we trained both generative models and evaluated synthetic data quality through a rigorous Train-on-Synthetic-Test-on-Real (TSTR) machine learning validation framework and blinded expert clinical assessment by two independent audiologists. The VAE approach achieved 86.3% utility for hearing loss prediction, as compared to the benchmark real data (Train-on-Real-Test-on-Real). Both methods demonstrated strong privacy preservation, with zero exact record matches and robust membership inference attack resistance. Statistical validation confirmed equivalence within clinically negligible margins (<1 dB HL) across all audiometric frequencies. Blinded assessment of 85 patient profiles by two independent expert audiologists revealed that VAE synthetic data achieved high clinical plausibility ratings, with 96.7% of VAE profiles rated as plausible, compared to 13.3% for KDE. Inter-rater reliability was moderate (Cohen’s weighted κ=0.553, ICC =0.556), with 84.7% of ratings within one point, and both raters independently ranking VAE above real data above KDE. These findings establish validated methodologies for generating privacy-preserving synthetic audiometric data suitable for machine learning applications and clinical education, addressing a critical gap in hearing health research infrastructure.

  • Open Access

    Early Detection of Ototoxicity Using Serial Mobile Audiometry, Otoacoustic Emissions Testing, and Inner Ear Biomarker Measurement in Patients Receiving Platinum-Based Chemotherapy Treatment: It is Feasible to Implement in a National Health Service (NHS) Cancer Ambulatory Care Setting

    by Nishchay Mehta, Asma Awad, Marina Forbes, Eleftheria Iliadou, Roulla Katiri, Tanjinah Ferdous, Helen Lowe, Leah Ensell, Constantine Alifrangis, Anne G.M. Schilder, et al.

    Importance: Platinum-based chemotherapy is widely used in the management of solid tumors, and up to 60% of patients will experience ototoxic side effects, including hearing loss and tinnitus. The advent of novel otoprotective drugs generates a need for more flexible and innovative real-time approaches to drug-induced ototoxicity monitoring in cancer patients. Objective: To assess the feasibility of recruitment and compliance to monitoring of ototoxicity with mobile out-of-booth audiometry, Distortion Product Otoacoustic Emission (DPOAE) testing, and serum biomarker measurements. Design: Prospective observational study. Setting: Cancer ambulatory care. Participants: Patients undergoing platinum-based chemotherapy for bone or testicular cancer. Results: Twenty-three patients were enrolled between February 2021 and June 2023. The mean age was 17.5 years (range: 13 to 36 y). Average compliance to monitoring assessments was 86.5% for audiometry, 84.8% for DPOAEs, and 84.3% for serum biomarker measurements. End-of-treatment audiometric changes indicative of ototoxicity were found in 18 (78.2%) of the participants. Conclusions and relevance: High compliance with serial mobile audiometry and otoacoustic emissions testing, and serum biomarker measurement suggests that ototoxicity monitoring can be successfully performed in cancer ambulatory care rather than in an audiology department setting. Further investigation is required to understand the role of serum prestin and otolin-1 in monitoring ototoxicity.

  • Open Access

    Assessment of current patient reported outcome measures for three core outcome domains for single-sided deafness device intervention trials

    by Roulla Katiri, Deborah A. Hall, Derek J. Hoare, Sandra Smith, Bethany Adams, Kathryn Fackrell, Adele Horobin, Nicholas Hogan, Nóra Buggy, Pádraig T. Kitterick

    Background: Outcome reporting in clinical trials of auditory interventions for adults with Single-Sided Deafness (SSD) is inconsistent. The Core Rehabilitation Outcome Set for Single-Sided Deafness (CROSSSD) initiative has recommended three outcome domains as a minimum standard in the design of SSD intervention clinical trials. These are, Spatial orientation, Group conversations in noisy social situations, and Impact on social situations. The study objectives were to (i) understand exactly what the outcome domains mean to SSD experts, and (ii) identify and assess candidate PROMs in terms of how well they measure the experts’ conceptualisation of those SSD outcome domains. Methodology: Stakeholder representatives participated in two semi-structured online focus groups. Participants were four adults diagnosed with SSD with experience of auditory interventions, two healthcare professionals working in the field, and one clinical researcher with experience in evaluating interventions. Thematic analysis was used to determine conceptual elements of each domain. COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN) initiative recommendations were adopted to assess the relevance and comprehensiveness (content validity) of available candidate instruments. Results: Multiple key concepts were identified for each outcome domain, and presented as a taxonomy. To be acceptable, any measurement instrument would need to achieve good coverage of all concepts in this taxonomy. From the 76 candidate instruments reviewed, none met accepted standards for content validity for SSD. The best performing candidates were (i) Spatial orientation: the Spatial Hearing Questionnaire and two variants of the Speech, Spatial and Qualities scale (SSQ-12, SSQ-18-C), (ii) Group conversations in noisy situations: the Communication Profile for Hearing Impaired (CPHI) questionnaire, SSQ-12, SSQ-18-C, and a multi-item questionnaire developed by Schafer and colleagues, and (iii) Impact on social situations: the CPHI questionnaire. Conclusions: Multi-dimensional outcome domains introduce specific considerations for how they should be measured. Although some candidates instruments had reasonable comprehensiveness, modification is needed to ensure that there is overall greater relevance to the key concepts.


  • Open Access

    Application of the world guidelines for falls prevention and management’s risk stratification algorithm to patients on a frailty intervention pathway and the potential utility of sensory impairment information

    by Roulla Katiri, Jack A. Holman, Siobhán Magner, Cian O’cAheny, Colm P. Byrne

    Background: The 2022 world guidelines for falls prevention and management suggest measuring sensory function including dizziness, vision, and hearing. These variables are not included in the falls risk stratification algorithm. This study sought to investigate the utility of the guidelines and potential avenues for improvement. This study applied the falls risk stratification recommendations and reviewed the individual sensory impairment risk factor variables predictive of falls and falls risk grouping in those assessed by a frailty intervention team (FIT) based in an emergency department (ED). Methods: Patients over 65 years old who attended the ED and had a comprehensive geriatric assessment carried out by FIT over a period of four months were included in this retrospective cross-sectional study. Patient characteristics, medication, physical and sensory function status data was retrieved and analysed with respect to falls and falls risk grouping. Results: Data was gathered retrospectively from 392 patients. Excluding those with missing data, almost all attendees were in the high-risk of falls category (n = 170, 43.4%), or the low-risk category (n = 149, 38.0%). Few people were in the intermediate-risk category (n = 19, 4.8%). Hearing loss and dizziness were significantly associated with falls incidence, whereas vision and balance were not. Hearing loss, balance and dizziness were significantly associated with risk grouping, whereas vision was not. Conclusions: Most older adults included in the analysis fell into the low- or high-risk categories, with a minority in the intermediate-risk category. This suggests that the inclusion criteria for the intermediate category could be altered for greater sensitivity. While impaired balance and vision were the most common impairments, hearing status, balance and dizziness were associated with risk group. These results, through a practical application of the world guidelines for falls to an acute clinical sample, raise the possibility of refining the falls risk stratification criteria, and highlight the capacity for additional sensory intervention to mitigate falls risk.


  • Open Access

    The Core Rehabilitation Outcome Set for Single-Sided Deafness (CROSSSD) study: International consensus on outcome measures for trials of interventions for adults with single-sided deafness

    by Roulla Katiri, Deborah A. Hall, Derek J. Hoare, Kathryn Fackrell, Adele Horobin, Nicholas Hogan, Nóra Buggy, Paul H. Van de Heyning, Jill B. Firszt, Iain A. Bruce, et al.

    Background: Single-sided deafness (SSD) has functional, psychological, and social consequences. Interventions for adults with SSD include hearing aids and auditory implants. Benefits and harms (outcome domains) of these interventions are until now reported inconsistently in clinical trials. Inconsistency in reporting outcome measures prevents meaningful comparisons or syntheses of trial results. The Core Rehabilitation Outcome Set for Single-Sided Deafness (CROSSSD) international initiative used structured communication techniques to achieve consensus among healthcare users and professionals working in the field of SSD. The novel contribution is a set of core outcome domains that experts agree are critically important to assess in all clinical trials of SSD interventions. Methods: A long list of candidate outcome domains compiled from a systematic review and published qualitative data, informed the content of a two-round online Delphi survey. Overall, 308 participants from 29 countries were enrolled. Of those, 233 participants completed both rounds of the survey and scored each outcome domain on a 9-point scale. The set of core outcome domains was finalised via a web-based consensus meeting with 12 participants. Votes involved all stakeholder groups, with an approximate 2:1 ratio of professionals to healthcare users participating in the Delphi survey, and a 1:1 ratio participating in the consensus meeting. Results: The first round of the survey listed 44 potential outcome domains, organised thematically. A further five outcome domains were included in Round 2 based on participant feedback. The structured voting at round 2 identified 17 candidate outcome domains which were voted on at the consensus meeting. Consensus was reached for a core outcome domain set including three outcome domains: spatial orientation, group conversations in noisy social situations, and impact on social situations. Seventy-seven percent of the remaining Delphi participants agreed with this core outcome domain set. Conclusions: Adoption of the internationally agreed core outcome domain set would promote consistent assessment and reporting of outcomes that are meaningful and important to all relevant stakeholders. This consistency will in turn enable comparison of outcomes reported across clinical trials comparing SSD interventions in adults and reduce research waste. Further research will determine how those outcome domains should best be measured.


  • Outcomes from 7 years of a direct to audiology referral pathway

    by Roulla Katiri, Nina Sivan, Anthony Noone, Eric Farrell, Laura McLoughlin, Bronagh Lang, Bronagh O’dOnnell, Stephen M. Kieran

    Background The direct to audiology via ENT referral pathway was designed to enhance patient access to audiology services. The pathway is recommended in the Otolaryngology Head & Neck Surgery: a model of care for Ireland report, published in 2019. Aims This study aimed to review the outcomes of all patients that attended our institution over the last seven years. Methods A retrospective review of the direct to audiology referral service from March 2014 to December 2021 was conducted. Outcomes were defined by the pathway each patient took following attendance at the audiology clinic. Patients were either (i) discharged, (ii) referred for hearing aid(s) or (iii) identified as candidates for further diagnostic assessments, including a follow-up at the ENT outpatient clinic. Results During the time frame, 419 patients were triaged to the pathway. The average wait time was 13 days. The average age was 53 years (range 16–96 years, SD = 6.1). Approximately 34% (n = 143) of all patients referred were discharged back to the GP by the audiologist, but 66% (n = 276) presented with ‘red flags’ and needed further investigation in the ENT clinic, with 30% (n = 73) ultimately requiring imaging studies. Over half (n = 254, 61%) were referred for hearing aids. Conclusion The direct to audiology initiative has proven effective at reducing waiting times for ENT patients solely in need of audiological intervention. Approximately one-third of these referrals to the ENT service can be assessed comprehensively in the audiology clinic, thereby reducing the demand for ENT clinics, enhancing service provision and expediting onwards referral for amplification.


  • Open Access

    The Core Rehabilitation Outcome Set for Single-Sided Deafness (CROSSSD) study: International consensus on outcome measures for trials of interventions for adults with single-sided deafness

    by Roulla Katiri, Deborah A. Hall, Derek J. Hoare, Kathryn Fackrell, Adele Horobin, Nicholas Hogan, Nóra Buggy, Paul H. Van de Heyning, Jill B. Firszt, Iain A. Bruce, et al.

    Background: Single-sided deafness (SSD) often has functional, psychological, and social consequences. Interventions for adults with SSD include hearing aids and auditory implants. Benefits and harms of these interventions (outcome domains) are until now reported inconsistently in clinical trials. Inconsistency in reporting of outcome measures prevents meaningful comparisons or syntheses of trial results. The Core Rehabilitation Outcome Set for Single-Sided Deafness (CROSSSD) international initiative used structured communication techniques to achieve consensus among healthcare users and professionals working in the field of SSD. The product is a set of core outcome domains that experts agree are critically important to assess in all clinical trials of SSD interventions.Methods: A long list of candidate outcome domains compiled from a systematic review and published qualitative data, informed the content of a two-round online Delphi survey. Overall, 308 participants from 29 countries were enrolled. Of those, 233 participants completed both rounds of the survey and scored each outcome domain on a 9-point scale. The set of core outcome domains was finalised via a web-based consensus meeting with 12 participants. Votes involved all stakeholder groups, with an approximate 2:1 ratio of professionals to healthcare users participating in the Delphi survey, and a 1:1 ratio participating in the consensus meeting. Results: The first round of the survey listed 44 potential outcome domains, organised thematically. A further five outcome domains were included in Round 2 based on participant feedback. The structured voting at Round 2 identified 17 candidate outcome domains which were voted on at the consensus meeting. Consensus was reached for a core outcome domain set including three outcome domains: Spatial orientation, Group conversations in noisy social situations, and Impact on social situations. Seventy-seven percent of the remaining Delphi participants agreed with this core outcome domain set.Conclusions: Adoption of the internationally agreed core outcome domain set will promote consistent assessment and reporting of outcomes that are meaningful and important to all relevant stakeholders. This consistency will in turn enable comparison of outcomes reported across clinical trials comparing SSD interventions in adults and reduce research waste. Further research will determine how those outcome domains should best be measured.

  • Open Access

    Redesigning a Web-Based Stakeholder Consensus Meeting About Core Outcomes for Clinical Trials: Formative Feedback Study

    by Roulla Katiri, Deborah A Hall, Derek J Hoare, Kathryn Fackrell, Adele Horobin, Nóra Buggy, Nicholas Hogan, Pádraig T Kitterick, Core Rehabilitation Outcome Set For Single-Sided Deafness (CROSSSD) Initiative

    Background: Clinical trials that assess the benefits and harms of an intervention do so by measuring and reporting outcomes. Inconsistent selection and diversity in the choice of outcomes make it challenging to directly compare interventions. To achieve an agreed core set of outcomes, a consensus methodology is recommended, comprising a web-based Delphi survey and a face-to-face consensus meeting. However, UK government regulations to control the pandemic prohibited plans for a face-to-face consensus meeting as part of the Core Rehabilitation Outcome Set for Single-Sided Deafness (CROSSSD) study. Objective: This study aims to evaluate the modifications made by the CROSSSD study team to achieve consensus using web-based methods, but with minimal deviation from the original study protocol. Methods: The study team worked with health care users and professionals to translate the planned face-to-face consensus meeting in a web-based format, preserving the key elements of the nominal group technique. A follow-up survey gathered evaluation feedback on the experiences of the 22 participating members. Feedback covered premeeting preparation, the process of facilitated discussions and voting, ability to contribute, and perceived fairness of the outcome. Results: Overall, 98% (53/54) of feedback responses agreed or strongly agreed with the statements given, indicating that the web-based meeting achieved its original goals of open discussion, debate, and voting to agree with a core outcome set for single-sided deafness. Hearing-impaired participants were fully engaged, but there were some methodological challenges. For the participants, challenges included building rapport, understanding, and delivering the tasks in hand. For the study team, challenges included the need for thorough preparation and management of the unpredictability of tasks on the day. Conclusions: Sharing our experiences and lessons learned can benefit future core outcome set developers. Overcoming the challenges of delivering a web-based consensus exercise in the face of the pandemic can be applied more generally to maximize inclusiveness, enhance geographical access, and reduce research costs.


  • Open Access

    Systematic review of outcome domains and instruments used in designs of clinical trials for interventions that seek to restore bilateral and binaural hearing in adults with unilateral severe to profound sensorineural hearing loss (‘single-sided deafness’)

    by Roulla Katiri, Deborah A. Hall, Catherine F. Killan, Sandra Smith, Pattarawadee Prayuenyong, Pádraig T. Kitterick

    Background: This systematic review aimed to identify, compare and contrast outcome domains and outcome instruments reported in studies investigating interventions that seek to restore bilateral (two-sided) and/or binaural (both ears) hearing in adults with single-sided deafness (SSD). Findings can inform the development of evidence-based guidance to facilitate design decisions for confirmatory trials. Methods: Records were identified by searching MEDLINE, EMBASE, PubMed, CINAHL, ClinicalTrials.gov, ISRCTN, CENTRAL, WHO ICTRP and the NIHR UK clinical trials gateway. The search included records published from 1946 to March 2020. Included studies were those as follows: (a) recruiting adults aged 18 years or older diagnosed with SSD of average threshold severity worse than 70 dB HL in the worse-hearing ear and normal (or near-normal) hearing in the better-hearing ear, (b) evaluating interventions to restore bilateral and/or binaural hearing and (c) enrolling those adults in a controlled trial, before-and-after study or cross-over study. Studies that fell just short of the participant eligibility criteria were included in a separate sensitivity analysis. Results: Ninety-six studies were included (72 full inclusion, 24 sensitivity analysis). For fully included studies, 37 exclusively evaluated interventions to re-establish bilateral hearing and 29 exclusively evaluated interventions to restore binaural hearing. Overall, 520 outcome domains were identified (350 primary and 170 secondary). Speech-related outcome domains were the most common (74% of studies), followed by spatial-related domains (60% of studies). A total of 344 unique outcome instruments were reported. Speech-related outcome domains were measured by 73 different instruments and spatial-related domains by 43 different instruments. There was considerable variability in duration of follow-up, ranging from acute (baseline) testing to 10 years after the intervention. The sensitivity analysis identified no additional outcome domains. Conclusions: This review identified large variability in the reporting of outcome domains and instruments in studies evaluating the therapeutic benefits and harms of SSD interventions. Reports frequently omitted information on what domains the study intended to assess, and on what instruments were used to measure which domains. Trial registration: The systematic review protocol is registered on PROSPERO (International Prospective Register of Systematic Reviews): Registration Number CRD42018084274. Registered on 13 March 2018, last revised on 7th of May 2019.


  • Open Access

    Redesigning a Web-Based Stakeholder Consensus Meeting About Core Outcomes for Clinical Trials: Formative Feedback Study (Preprint)

    by Roulla Katiri, Deborah A Hall, Derek J Hoare, Kathryn Fackrell, Adele Horobin, Nóra Buggy, Nicholas Hogan, Pádraig T Kitterick

    BACKGROUND Clinical trials that assess the benefits and harms of an intervention do so by measuring and reporting outcomes. Inconsistent selection and diversity in the choice of outcomes make it challenging to directly compare interventions. To achieve an agreed core set of outcomes, a consensus methodology is recommended, comprising a web-based Delphi survey and a face-to-face consensus meeting. However, UK government regulations to control the pandemic prohibited plans for a face-to-face consensus meeting as part of the Core Rehabilitation Outcome Set for Single-Sided Deafness (CROSSSD) study. OBJECTIVE This study aims to evaluate the modifications made by the CROSSSD study team to achieve consensus using web-based methods, but with minimal deviation from the original study protocol. METHODS The study team worked with health care users and professionals to translate the planned face-to-face consensus meeting in a web-based format, preserving the key elements of the nominal group technique. A follow-up survey gathered evaluation feedback on the experiences of the 22 participating members. Feedback covered premeeting preparation, the process of facilitated discussions and voting, ability to contribute, and perceived fairness of the outcome. RESULTS Overall, 98% (53/54) of feedback responses agreed or strongly agreed with the statements given, indicating that the web-based meeting achieved its original goals of open discussion, debate, and voting to agree with a core outcome set for single-sided deafness. Hearing-impaired participants were fully engaged, but there were some methodological challenges. For the participants, challenges included building rapport, understanding, and delivering the tasks in hand. For the study team, challenges included the need for thorough preparation and management of the unpredictability of tasks on the day. CONCLUSIONS Sharing our experiences and lessons learned can benefit future core outcome set developers. Overcoming the challenges of delivering a web-based consensus exercise in the face of the pandemic can be applied more generally to maximize inclusiveness, enhance geographical access, and reduce research costs. CLINICALTRIAL