Community-based ototoxicity monitoring with extended high-frequency audiometry and community health workers for drug-resistant tuberculosis

Abstract

Community-based ototoxicity monitoring with extended high-frequency audiometry and community health workers for drug-resistant tuberculosis Name: Lucia Jane Stevenson Supervisor: Professor De Wet Swanepoel Co-supervisor: Associate Professor Leigh Biagio-de Jager Department: Speech-Language Therapy and Audiology Degree: PhD Audiology

Ototoxicity occurs when certain life-saving drugs or ionising radiation are administered to patients for the treatment of illness such as cancers, cystic fibrosis and tuberculosis. Ototoxic drugs, including aminoglycoside antibiotics and platinum-based chemotherapies, cause damage to the cochlear or vestibular structures of the inner ear, or both, affecting sensory function. South Africa has a high burden of drug-resistant tuberculosis (DRTB) and until recently, aminoglycosides, were predominant in treatment regimens. Decentralised community-based ototoxicity monitoring programmes (OMPs) facilitated by community health workers (CHWs) have been implemented in response to the DRTB ototoxicity burden and to support early detection of hearing loss and increased patient access to services. This research project entailed a retrospective record review of longitudinal ototoxicity monitoring of 831 patients with DRTB, using data collected at 19 community-based clinics by six CHWs and two primary health care (PHC) audiologists, using portable audiological equipment in the City of Cape Town between 2013 and 2017. Three studies were conducted. Study I evaluated the service delivery practices of a decentralised, community-based OMP facilitated by CHWs for 831 patients (age mean = 36.1; SD = 11.0) with DRTB. The service delivery practices were evaluated against the OMP protocol and national and international recommended guidelines for ototoxicity monitoring. Approximately half (46.8%) of the patients had an initial assessment conducted in accordance with the OMP protocol recommendations. The OMP follow-up rates improved from 53.7% to 79.5% as the OMP became more established over time, higher than those of a similar DRTB treatment programme. However, the frequency and regularity of ototoxicity monitoring assessments for patients in this study did not meet the recommendations of the OMP protocol or the guidelines for ototoxicity monitoring. On average, patients were assessed 3.1 (SD = 2.31) times, with just 8% (69/831) of patients returning for the recommended six or more ototoxicity monitoring assessments. Ototoxicity monitoring was conducted on average every 58.3 (SD = 6.23) days, almost twice the 30 days recommended by the OMP protocol. Extended high-frequency (EHF) pure-tone audiometry (27.5%) was underutilised by testers and data recording was inconsistent (e.g. 37.7% of patient gender was not recorded by testers). Study II described the observed longitudinal treatment effects for DRTB and ototoxicity monitoring conducted by CHWs using conventional audiometry (0.25–8 kHz), in a decentralised community-based model of care. Of the 831 patients included in Study I, 194 (age mean = 36.2; SD = 11.3) met the selection criteria and were included in Study II. Patients’ initial assessments were conducted on average 16.8 days (SD = 86.5; range = -494 to 14 days) before treatment initiation. Follow-up rates for consecutive monitoring assessments reached as high as 80.6% for patients assessed by CHWs. However, few patients (14.2–32.6%) were assessed with the frequency (≥ 6 assessments) and regularity required for effective ototoxicity monitoring, with assessments conducted on average every 53.4 to 64.3 days. There was a significant (p = 0.019; U = 2637.0) difference between the average number of follow-up visits made by patients assessed by CHWs (average = 3.3; SD = 2.1; 148/194) and those assessed by PHC audiologists (average = 4.3; SD = 2.5; 46/194). However, the average number of days elapsing between monitoring assessments were fewer for patients assessed by CHWs (53.4 days; SD = 10.3) than for patients assessed by PHC audiologists (64.3 days; SD = 19.3). There was a high prevalence (51.5%; 100/194) of pre-existing hearing loss for patients at the time of the initial assessment. Following DRTB treatment, 51.5% (100/194) of patients presented with a significant ototoxic shift meeting one or more of the American Speech-Language-Hearing Association (ASHA) criteria with ototoxic shifts occurring most often at the high frequencies (4–8 kHz). Deterioration in hearing thresholds was bilateral and most pronounced (p < 0.05) at the high frequencies (4–8 kHz) and high frequency pure tone average (PTA) (3–8 kHz). The presence of pre-existing hearing loss, HIV co-infection, and a history of noise exposure were significant predictors (p < 0.05) of ototoxicity in DRTB patients. Study III investigated EHF audiometry for monitoring ototoxicity in a longitudinal treatment programme. Of the 831 patients included in Study I, 69 (age mean = 37.9; SD = 11.2) met the selection criteria and were included in Study III. Gender (27.5%; 19/69) and medication type (47.8%; 33/69) administered to patients was not recorded on the data collection forms by some testers and was therefore unavailable for inclusion in this retrospective study. Initial assessments were conducted on average 40.3 days (SD = 70.9; range = 0 to 301 days) after treatment initiation. At the initial assessment, 36.2% (25/69) of patients presented with a hearing loss in one or more frequency in the conventional range (0.25–8 kHz); compared to 65.2% (45/69) of patients when also considering EHF thresholds (0.25–16 kHz). Following treatment, the mean hearing threshold deterioration was significant (p < 0.05) at EHFs and the EHF PTA (9–16 kHz) of both ears. Including EHF thresholds resulted in more than half of patients (56.5%; 39/69) presenting with a significant ototoxic shift meeting one or more of the ASHA criteria, compared to 31.9% (22/69) if EHFs were not considered. Absent EHF thresholds owing to maximum equipment output limits were most pertinent at 16 kHz, with 17.4% of patient thresholds absent at the time of the initial assessment. This research project demonstrated that community-based ototoxicity monitoring is a feasible option to improve access to services and improve follow-up rates for patients. With ongoing training and supportive supervision, CHWs can facilitate community-based ototoxicity monitoring. DRTB treatment with kanamycin resulted in a significant deterioration of hearing status longitudinally, most markedly in the high frequencies and EHFs. The findings confirm that EHF audiometry is most sensitive for the early detection of ototoxicity and should be included in OMPs. However, to improve community-based OMP outcomes, OMP managers should reassess current protocols, guidelines and data recording practices and consider novel devices for ototoxicity monitoring with shortened optimised assessment approaches that target frequencies most sensitive to ototoxicity, including EHFs.