Closing the Gap in Surveillance and Audit of Invasive Mold Diseases for Antifungal Stewardship Using Machine Learning

Abstract

Clinical audit of invasive mold disease (IMD) in hematology patients is inefficient due to the difficulties of case finding. This results in antifungal stewardship (AFS) programs preferentially reporting drug cost and consumption rather than measures that actually reflect quality of care. We used machine learning-based natural language processing (NLP) to non-selectively screen chest tomography (CT) reports for pulmonary IMD, verified by clinical review against international definitions and benchmarked against key AFS measures. NLP screened 3014 reports from 1 September 2008 to 31 December 2017, generating 784 positives that after review, identified 205 IMD episodes (44% probable-proven) in 185 patients from 50,303 admissions. Breakthrough-probable/proven-IMD on antifungal prophylaxis accounted for 60% of episodes with serum monitoring of voriconazole or posaconazole in the 2 weeks prior performed in only 53% and 69% of episodes, respectively. Fiberoptic bronchoscopy within 2 days of CT scan occurred in only 54% of episodes. The average turnaround of send-away bronchoalveolar galactomannan of 12 days (range 7–22) was associated with high empiric liposomal amphotericin consumption. A random audit of 10% negative reports revealed two clinically significant misses (0.9%, 2/223). This is the first successful use of applied machine learning for institutional IMD surveillance across an entire hematology population describing process and outcome measures relevant to AFS. Compared to current methods of clinical audit, semi-automated surveillance using NLP is more efficient and inclusive by avoiding restrictions based on any underlying hematologic condition, and has the added advantage of being potentially scalable.

A population‐based analysis of attributable hospitalisation costs of invasive fungal diseases in haematological malignancy patients using data linkage of state‐wide registry and costing databases: 2009‐2015

“…Over the six‐year study period, total attributable hospitalisation costs were stable ( P = .90) with increases in IA ( P = .63) and other IFD ( P = .88) balanced against a decrease in invasive candidiasis related costs ( P = .08) with none reaching statistical significance likely due to small numbers of matched pairs. A 0.28% decrease in IFD incidence in the companion epidemiological study corroborates institutional studies from two major Victorian transplant centres suggesting that improvements in supportive care like antifungal prophylaxis have affected IFD incidence.…”

“…Despite six years of linked data with the VCDC, only a modest number of case‐control matched pairs were identified (334 matched pairs). These few matched pairs reiterate the need to improve IFD surveillance and to develop innovative machine learning solutions for prospective real‐time surveillance of IFD in cancer populations. The strengths, however, of our data linkage approach are evidenced in our ability to evaluate the longitudinal economic impact attributable to IFD at a state‐wide level, to perform cost regression modelling that adjusted for neutropenia and hospital region not included in the matching criteria and to reaffirm mucormycosis as an established IFD with a significant cost burden.…”

“…Our findings can inform future studies evaluating the cost‐effectiveness of IFD prevention initiatives while leveraging readily available datasets. Forthcoming research will focus on improving the quality of data acquisition for an uncommon but significant disease from hospitals using machine learning …”

“…13 In contrast, there are few data on the economic burden of non-Aspergillus moulds, including mucormycosis and scedosporiosis in Australia 3,14 partly due to their rarity and lack of surveillance systems in hospitals where IFD are managed. 15 To overcome the challenges of data acquisition for a rare disease, we utilised data linkage of population-based administrative coding and costing datasets with state-based cancer registry data. With linked data, we could for the first time evaluate the gross and attributable IFD hospitalisation costs adjusted for key clinical covariates and the longitudinal expenditure associated with IFD over a six-year period in Victoria, Australia.…”

“…Mucormycosis is difficult to manage and costly to treat with a median total hospitalisation cost of USD$57 324 (2014) in HM and allogeneic‐haematopoietic stem cell transplantation (allo‐HSCT) patients with similar estimates reported in Germany (€53 261, 2016) . In contrast, there are few data on the economic burden of non‐ Aspergillus moulds, including mucormycosis and scedosporiosis in Australia partly due to their rarity and lack of surveillance systems in hospitals where IFD are managed …”