Machine Learning Augmented Interpretation of Chest X-rays: A Systematic Review.

Limitations of the chest X-ray (CXR) have resulted in attempts to create machine learning systems to assist clinicians and improve interpretation accuracy. An understanding of the capabilities and limitations of modern machine learning systems is necessary for clinicians as these tools begin to permeate practice. This systematic review aimed to provide an overview of machine learning applications designed to facilitate CXR interpretation. A systematic search strategy was executed to identify research into machine learning algorithms capable of detecting >2 radiographic findings on CXRs published between January 2020 and September 2022. Model details and study characteristics, including risk of bias and quality, were summarized. Initially, 2248 articles were retrieved, with 46 included in the final review. Published models demonstrated strong standalone performance and were typically as accurate, or more accurate, than radiologists or non-radiologist clinicians. Multiple studies demonstrated an improvement in the clinical finding classification performance of clinicians when models acted as a diagnostic assistance device. Device performance was compared with that of clinicians in 30% of studies, while effects on clinical perception and diagnosis were evaluated in 19%. Only one study was prospectively run. On average, 128,662 images were used to train and validate models. Most classified less than eight clinical findings, while the three most comprehensive models classified 54, 72, and 124 findings. This review suggests that machine learning devices designed to facilitate CXR interpretation perform strongly, improve the detection performance of clinicians, and improve the efficiency of radiology workflow. Several limitations were identified, and clinician involvement and expertise will be key to driving the safe implementation of quality CXR machine learning systems.

Cost-Effectiveness of Pediatric Cochlear Implantation in Rural China.

OBJECTIVES: To evaluate the cost utility of cochlear implantation (CI) for severe to profound sensorineural hearing loss (SNHL) among children from rural settings in P.R. China (China). RESEARCH DESIGN: A cost-utility analysis (CUA) was undertaken using data generated from a single-center substudy of the Cochlear Pediatric Implanted Recipient Observational Study (Cochlear P-IROS). The data were projected over a 20-year time horizon using a decision tree model. SETTING: The Chinese healthcare payer and patient perspectives were adopted. INTERVENTION: Unilateral CI of children with a severe-to-profound SNHL compared with their preimplantation state of no treatment or amplification with hearing aids ("no CI" status). MAIN OUTCOME MEASURE/S: Incremental costs per quality adjusted life year (QALY) gained. RESULTS: The mean total discounted cost of unilateral CI was CNY 252,506 (37,876 USD), compared with CNY 29,005 (4,351 USD) for the no CI status from the healthcare payer plus patient perspective. A total discounted benefit of 8.9 QALYs was estimated for CI recipients compared with 6.7 QALYs for the no CI status. From the healthcare payer plus patient perspective, incremental cost-effectiveness ratio (ICER) for unilateral CI compared with no CI was CNY 100,561 (15,084 USD) per QALY. The healthcare payer perspective yielded an ICER of CNY 40,929 (6,139 USD) per QALY. Both ICERs fell within one to three times China's gross domestic product per capita (GDP, 2011-2015), considered "cost-effective" by World Health Organization (WHO) standards. CONCLUSIONS: Treatment with unilateral CI is a cost-effective hearing solution for children with severe to profound SNHL in rural China. Increased access to mainstream education and greater opportunities for employment, are potential downstream benefits of CI that may yield further societal and economic benefits. CI may be considered favorably for broader inclusion in medical insurance schemes across China.

Economic Evaluation of Treatments for Pediatric Bilateral Severe to Profound Sensorineural Hearing Loss: An Australian Perspective.

OBJECTIVES: In Australia, surgical treatment options for children with bilateral severe to profound sensorineural hearing loss exist in a continuum ranging from unilateral cochlear implantation (CI), sequential bilateral CI through to simultaneous bilateral CI, depending on the condition. When treatment options are mutually exclusive, the mean costs and benefits of each treatment group are summed together to obtain the total mean costs and benefits. This enables an incremental analysis of treatment options in the context of the treated populations.The objective was to evaluate the cost-utility of current Australian CI treatment practices in children using domestic costs and consequences when compared with bilateral hearing aids (HAs). RESEARCH DESIGN: Economic evaluation including a Markov model based on secondary sources. SETTING: The base case modeled a government health payer perspective over a child's lifetime. Primary and secondary school education costs were also assessed. INTERVENTION: Bilateral HAs compared with CI, including unilateral, sequential bilateral, or simultaneous bilateral CI weighted according to treatment. MAIN OUTCOME MEASURES: Incremental costs per quality adjusted life year. RESULTS: Approximately 42% of children in Australia with unilateral CI did not transition to sequential bilateral nor undergo simultaneous bilateral implantation. This differs from previous economic evaluations that assumed 100% of children transitioned to sequential bilateral CI treatment or were treated with simultaneous bilateral CI.The incremental cost utility of unilateral cochlear implantation compared with HAs was AUD 21,947/QALY. The weighted average incremental cost utility of the combined cochlear implantation treatment groups was AUD 31,238/QALY when compared with HAs. CONCLUSION: Previous economic evaluations of cochlear implantation assumed 100% of unilaterally treated patients would transition to sequential bilateral or be treated with simultaneous bilateral implantation. This approach does not take into account the total treated population, where a proportion of patients are treated with unilateral CI.CI was cost effective when compared with HAs, and included children treated with unilateral, sequential bilateral, and simultaneous bilateral CI.The model was sensitive to the number of assessment and habilitation visits. Alternative health service models with cost efficiencies are needed to reduce after care costs.

Cost-Utility Analysis of Cochlear Implantation in Australian Adults.

OBJECTIVES: Sequential and simultaneous bilateral cochlear implants are emerging as appropriate treatment options for Australian adults with sensory deficits in both cochleae. Current funding of Australian public hospitals does not provide for simultaneous bilateral cochlear implantation (CI) as a separate surgical procedure. Previous cost-effectiveness studies of sequential and simultaneous bilateral CI assumed 100% of unilaterally treated patients' transition to a sequential bilateral CI. This assumption does not place cochlear implantation in the context of the generally treated population. When mutually exclusive treatment options exist, such as unilateral CI, sequential bilateral CI, and simultaneous bilateral CI, the mean costs of the treated populations are weighted in the calculation of incremental cost-utility ratios. The objective was to evaluate the cost-utility of bilateral hearing aids (HAs) compared with unilateral, sequential, and simultaneous bilateral CI in Australian adults with bilateral severe to profound sensorineural hearing loss. RESEARCH DESIGN: Cost-utility analysis of secondary sources input to a Markov model. SETTING: Australian health care perspective, lifetime horizon with costs and outcomes discounted 5% annually. INTERVENTION: Bilateral HAs as treatment for bilateral severe to profound sensorineural hearing loss compared with unilateral, sequential, and simultaneous bilateral CI. MAIN OUTCOME MEASURES: Incremental costs per quality adjusted life year (AUD/QALY). RESULTS: When compared with bilateral hearing aids the incremental cost-utility ratio for the CI treatment population was AUD11,160/QALY. The incremental cost-utility ratio was weighted according to the number of patients treated unilaterally, sequentially, and simultaneously, as these were mutually exclusive treatment options. CONCLUSION: No peer-reviewed articles have reported the incremental analysis of cochlear implantation in a continuum of care for surgically treated populations with bilateral severe to profound sensorineural hearing loss. Unilateral, sequential, and simultaneous bilateral CI were cost-effective when compared with bilateral hearing aids. Technologies that reduce the total number of visits for a patient could introduce additional cost efficiencies into clinical practice.

A global patient outcomes registry: Cochlear paediatric implanted recipient observational study (Cochlear(™) P-IROS).

BACKGROUND: Currently, there is a paucity of data concerning the long-term outcomes, educational placement and quality of life of children implanted with hearing devices from large and representative samples of the population. To address this concern, a large, prospective, multicentre, multinational patient-outcomes registry for paediatric recipients of implantable hearing devices was developed. The benefits of this registry, its approach and methodology are described. METHODS/DESIGN: The Cochlear(™) Paediatric Implanted Recipient Observational Study (Cochlear P-IROS) is a prospective international patient-outcomes registry for children who are implanted in routine clinical practice with one or more hearing devices. The study aims to collect data on patient comorbidities, device use, auditory performance, quality of life and health-related utilities, across different types of implantable hearing devices from a range of manufacturers. Patients will be evaluated with a set of standardised and non-standardised questionnaires prior to initial device activation (baseline) and at six-monthly follow-up intervals up to 24 months and annually thereafter. The Cochlear P-IROS utilises a secure web interface to administer electronic case report forms to clinicians and families of implanted children. The web interface is currently available in five languages: English, Japanese, Korean, Mandarin and Russian. The interface also provides printable versions of the case report forms translated into 22 local languages for collection of data prior to entry online; additional languages may be added, as required. Participation in the Cochlear P-IROS registry is investigator-driven and voluntary. To date, the Cochlear P-IROS has recruited implant clinics across Australia, China, India, Indonesia, Turkey and Vietnam. The registry also aims to recruit multiple clinics in Cuba, Israel, Japan, Malaysia, Singapore, South Africa, South Korea and Russia. DISCUSSION: The use of a registry such as the Cochlear P-IROS will generate valuable data to support research interests of academics and clinicians around the globe. The data generated will be relevant for a wide range of stakeholders including regulators, payers, providers, policy makers, patients and their families, each with a different perspective for the acceptance and adoption of implantable hearing devices for the treatment of hearing loss.