BACKGROUND

Rural populations experience a disproportionate burden of childhood hearing loss. The public health impact is tremendous, with lifelong effects on educational attainment, psychosocial outcomes, and future vocational opportunities. School-based health programs provide essential access to services for rural children, but limited access to specialists and loss to follow-up reduce the effectiveness of these programs. Alaska has a large rural population that relies on a well-established statewide telemedicine network to address barriers to care. However, this network has never incorporated preventive services such as school hearing screening. Furthermore, although Alaska mandates hearing screenings in schools, little evidence exists on which screening protocols are most effective.

OBJECTIVES

The purpose of our cluster randomized controlled trial was to evaluate the effectiveness of a new mobile health (mHealth) screening protocol and an expedited telemedicine specialty referral for school hearing screening to connect children to an audiologist and an otolaryngologist for evaluation and treatment of hearing loss and middle-ear disease. Our goal was to improve the timeliness of identifying and treating childhood hearing loss. We used a mixed-methods pretrial to inform trial design (exploratory sequential) and posttrial to understand the intervention experience and trial findings (explanatory sequential).

METHODS

Our cluster randomized controlled trial tested a newly designed telemedicine specialty referral among students in 15 rural communities served by a Tribal health care organization in northwest Alaska. We randomized communities to the telemedicine specialty referral (intervention) or to a standard primary care referral (comparison). A total of 1481 children (K-12) enrolled in the trial over 2 academic years (2017-2019). All participating children completed the standard school hearing screen, the mHealth screening protocol including tympanometry (mHealth plus tympanometry), and a full audiometric evaluation on the same screening day. The primary outcome was time to ear or hearing diagnosis (or both)—here, —for the 9 months following the screening date. Secondary outcomes included differences in the prevalence of hearing loss, hearing-related quality of life, and school performance; we also assessed the accuracy of the mHealth-plus-tympanometry hearing screening protocol compared with accuracy of the standard school screen. We used intention-to-treat analysis. In the second year of this project, we conducted an ancillary trial testing the same referral pathways for preschool children (N = 153). Mixed methods included focus groups with stakeholders before the start of the trial to inform study design and semistructured interviews throughout and after the trial to ensure contextual understanding of the trial findings.

RESULTS

Overall, 790 (53.3%) children required referral in at least 1 study year, with 391 and 399 in the intervention and comparison communities, respectively. Of these, 268 (68.5%) and 128 (32.1%) received an ear/hearing diagnosis within 9 months in the intervention and comparison communities, respectively. Referred children in the intervention arm were 2.3 times (95% CI, 1.4-3.8 times) more likely to receive an ear/hearing diagnosis and had a time to diagnosis that was 17.6 times (95% CI, 6.8-45.3 times) faster than those in comparison communities. The 2 treatment arms did not differ significantly in prevalence of hearing loss, child or adolescent quality of life scores, or national percentile mathematics or reading scores in year 2 of the trial. The mHealth-plus-tympanometry screen outperformed the standard school screen on metrics of sensitivity (77%; 95% CI, 73.1%-80.9%), specificity (88.8%; 95% CI, 87.3%-90.4%), positive predictive value (65.3%; 95% CI, 61.1%-69.6%), negative predictive value (93.4%; 95% CI, 92.2%-94.6%), and area under the curve (0.829; 95% CI, 0.809-0.849) compared with the gold-standard audiometric evaluation. We also found a 17.6 percentage point (pp) improvement (95% CI, 12.6 pp-22.5 pp; < .001) in sensitivity and a 0.097 (95% CI, 0.073-0.122; < .001) improvement in the area under the curve. Of the 153 preschool children enrolled in the ancillary trial (intervention = 90, comparison = 63), 71 (46.4%) were referred for diagnosis, with 39 (43.3%) and 32 (50.8%) in the intervention and comparison communities, respectively. Of those, 30 (76.9%) and 16 (50.0%) received an ear/hearing diagnosis within 9 months in the intervention and comparison communities, respectively. Preschool children in the intervention group were 1.57 times more likely (95% CI, 1.22, 2.01 times) to receive an ear/hearing diagnosis, and their time to diagnosis was 5.66 (95% CI, 2.50-12.83) times faster than that for children in the comparison group. Focus groups informed several elements of the trial design. They provided insight into community members' perceptions of hearing loss, which fell into 3 main domains: etiology, impact, and treatment. Interviews with stakeholders about the barriers and facilitators of the telemedicine intervention identified 6 key factors: clinic capacity, personnel/ownership, scheduling, telemedicine equipment/processes, communication, and awareness of the need for follow-up.

CONCLUSIONS

Telemedicine specialty referral significantly improved follow-up and reduced time to diagnosis for childhood and adolescent hearing screening in rural Alaska. Telemedicine may be applicable to other disparities in preventable health services, improving access to necessary specialty care. The mHealth-plus-tympanometry screening was the most accurate in this population. Our results suggest that tympanometry should be added to school hearing screening, especially in populations with a high prevalence of ear infections. The qualitative results played a key role in the success of trial execution in terms of design and contextualizing the trial findings.

LIMITATIONS

Effective coordination and communication between clinic and school were achieved more in some communities than in others. Challenges such as influenza outbreaks, urgent clinical care, and staffing shortages created competing priorities that sometimes delayed telemedicine specialty referrals. This study was conducted in a unique Tribal health care context; thus, the telemedicine specialty referral may not be directly generalizable to non-Tribal regions where multiple rural health entities and payers are common. Qualitative data, although extensive and inclusive of a diverse set of individuals, only represented a small portion of the communities at large.