Risk-based Surveillance Strategies Research Articles

Ecological drivers for poultry farms predisposed to highly pathogenic avian influenza virus infection during the initial phase of the six outbreaks between 2010-2021: a nationwide study in South Korea.

Highly pathogenic avian influenza (HPAI) has caused substantial economic losses worldwide. An understanding of the environmental drivers that contribute to spillover transmission from wild birds to poultry farms is important for predicting areas at risk of introduction and developing risk-based surveillance strategies. We conducted an epidemiological study using data from six HPAI outbreak events in South Korea. An aggregate-level study design was implemented using third-level administrative units in South Korea. Only regions with high natural reservoir suitability were included. The incidence of HPAI at chicken and duck farms during the initial phase (30 and 45 days after the first case) of each outbreak event was used as the outcome variable, assuming that cross-species transmission from wild birds was the dominant exposure leading to infection. Candidate environmental drivers were meteorological factors, including temperature, precipitation, humidity, and altitude, as well as the proportion of protected area, farm density, deforestation level, and predator species richness. Logistic regression models were implemented; conditional autoregression models were used in cases of spatial autocorrelation of residuals. Lower temperature, higher farm density, and lower predator species richness were significantly associated with a higher risk of HPAI infection on chicken farms. Lower temperature, higher proportion of protected area, and lower predator species richness were significantly associated with a higher risk of HPAI infection on duck farms. The predicted dominant transmission routes on chicken and duck farms were horizontal and spillover, respectively. These results reveal a potential protective effect of predator species richness against HPAI outbreaks. Further studies are required to confirm a causal relationship.

Identifying target areas for risk-based surveillance and control of transboundary animal diseases: a seasonal analysis of slaughter and live-trade cattle movements in Uganda

Animal movements are a major driver for the spread of Transboundary Animal Diseases (TADs). These movements link populations that would otherwise be isolated and hence create opportunities for susceptible and infected individuals to meet. We used social network analysis to describe the seasonal network structure of cattle movements in Uganda and unravel critical network features that identify districts or sub-regions for targeted risk-based surveillance and intervention. We constructed weighted, directed networks based on 2019 between-district cattle movements using official livestock mobility data; the purpose of the movement (‘slaughter’ vs. ‘live trade’) was used to subset the network and capture the risks more reliably. Our results show that cattle trade can result in local and long-distance disease spread in Uganda. Seasonal variability appears to impact the structure of the network, with high heterogeneity of node and edge activity identified throughout the seasons. These observations mean that the structure of the live trade network can be exploited to target influential district hubs within the cattle corridor and peripheral areas in the south and west, which would result in rapid network fragmentation, reducing the contact structure-related trade risks. Similar exploitable features were observed for the slaughter network, where cattle traffic serves mainly slaughter hubs close to urban centres along the cattle corridor. Critically, analyses that target the complex livestock supply value chain offer a unique framework for understanding and quantifying risks for TADs such as Foot-and-Mouth disease in a land-locked country like Uganda. These findings can be used to inform the development of risk-based surveillance strategies and decision making on resource allocation. For instance, vaccine deployment, biosecurity enforcement and capacity building for stakeholders at the local community and across animal health services with the potential to limit the socio-economic impact of outbreaks, or indeed reduce their frequency.

Network analysis of pig movement data as an epidemiological tool: an Austrian case study

Animal movements represent a major risk for the spread of infectious diseases in the domestic swine population. In this study, we adopted methods from social network analysis to explore pig trades in Austria. We used a dataset of daily records of swine movements covering the period 2015–2021. We analyzed the topology of the network and its structural changes over time, including seasonal and long-term variations in the pig production activities. Finally, we studied the temporal dynamics of the network community structure. Our findings show that the Austrian pig production was dominated by small-sized farms while spatial farm density was heterogeneous. The network exhibited a scale-free topology but was very sparse, suggesting a moderate impact of infectious disease outbreaks. However, two regions (Upper Austria and Styria) may present a higher structural vulnerability. The network also showed very high assortativity between holdings from the same federal state. Dynamic community detection revealed a stable behavior of the clusters. Yet trade communities did not correspond to sub-national administrative divisions and may be an alternative zoning approach to managing infectious diseases. Knowledge about the topology, contact patterns, and temporal dynamics of the pig trade network can support optimized risk-based disease control and surveillance strategies.

Burden of chronic health conditions (CHCs) among people with HIV (PWH) and non-AIDS defining cancers (NADC): A report from the center for AIDS research (CFAR) network of integrated clinical systems (CNICS).

12089 Background: Advances in anti-retroviral therapy (ART) have led to improved life expectancy of PWH and a growing population of PWH who are at an increased risk of NADC. While the risk of CHCs in HIV uninfected cancer survivors is well known, the burden of morbidity (CHCs) borne by PWH-NADC remains unstudied and can adversely influence long-term survival. Methods: We utilized CNICS data for PWH ≥18y from 1995- 2018. CNICS data are harmonized across 8 academic sites, comprising patient-level clinical, morbidity, and outcomes data in the ART era. We included individuals with the five most common NADCs: anal squamous cell cancer, non-small cell lung cancer (NSCLC), prostate cancer, classic Hodgkin lymphoma (cHL), and hepatocellular cancer (HCC). For controls, we conducted a 1:2 matching (age at CNICS entry, year of CNICS entry, sex) for each NADC subgroup to select PWH and no NADC at last follow-up. Comorbidities diagnosed after cohort entry are validated in CNICS and were graded using the Common Terminology Criteria for Adverse Events into grade 1 (mild) through grade 4 (severe/life-threatening). Results: We included 693 PWH-NADC and 1,345 non-cancer PWH controls in this study. The NADC distribution was anal cancer (n = 195, 28%), NSCLC (n = 155, 22%), prostate cancer (n = 144, 21%), cHL (n = 124, 18%), and HCC (n = 75, 11%). Mean age at CNICS entry was 45y for cases and controls (SD: 10.4). At a median follow-up of 12.5y for cases and 9.5y for controls, the prevalence of all-grade and grade ≥3 CHCs was higher in PWH-NADC compared with controls (94% vs. 88%, and 25% vs. 14%, respectively [p < 0.01]). PWH-NADC had a higher prevalence of all-grade HTN (87% vs. 80%), CKD (49% vs. 46%), hyperlipidemia (56% vs. 50%), DM (26% vs. 18%), VTE (4% vs. 1%), MI (12% vs. 5%), and CHF (7% vs. 3%) compared with controls (p-value for all < 0.01). In Cox regression including type of NADC, age at CNICS cohort entry, ART use, race/ethnicity, NADCs were associated with increased risk of new-onset grade 3 or 4 CHCs (HR 2.94, p < 0.01). When examining specific CHCs, all NADCs were associated with increased risk of new-onset grade 3 DM ([anal cancer HR 2.34, HCC HR 5.02, NSCLC HR 3.89, cHL HR 2.77, and prostate cancer HR 1.88], p-value for all < 0.01). Additionally, NSCLC was associated with increased risk of G3-4 MI (HR 2.79, p = 0.04), and prostate cancer with G3-4 CHF (HR 6.19, p < 0.01). Conclusions: In this large multi-institutional study, 25% of PWH-NADC had developed severe/life-threatening CHCs at last follow up. Occurrence of NADC among PWH was associated with increased hazards of developing grade ≥3 morbidities. The specific types of CHCs observed in PWH with NADC subtypes could be used to inform risk-based anticipatory surveillance and risk reduction strategies.

Machine learning-based model for predicting the esophagogastric variceal bleeding risk in liver cirrhosis patients

BackgroundLiver cirrhosis patients are at risk for esophagogastric variceal bleeding (EGVB). Herein, we aimed to estimate the EGVB risk in patients with liver cirrhosis using an artificial neural network (ANN).MethodsWe included 999 liver cirrhosis patients hospitalized at the Beijing Ditan Hospital, Capital Medical University in the training cohort and 101 patients from Shuguang Hospital in the validation cohort. The factors independently affecting EGVB occurrence were determined via univariate analysis and used to develop an ANN model.ResultsThe 1-year cumulative EGVB incidence rates were 11.9 and 11.9% in the training and validation groups, respectively. A total of 12 independent risk factors, including gender, drinking and smoking history, decompensation, ascites, location and size of varices, alanine aminotransferase (ALT), γ-glutamyl transferase (GGT), hematocrit (HCT) and neutrophil-lymphocyte ratio (NLR) levels as well as red blood cell (RBC) count were evaluated and used to establish the ANN model, which estimated the 1-year EGVB risk.The ANN model had an area under the curve (AUC) of 0.959, which was significantly higher than the AUC for the North Italian Endoscopic Club (NIEC) (0.669) and revised North Italian Endoscopic Club (Rev-NIEC) indices (0.725) (all P < 0.001). Decision curve analyses revealed improved net benefits of the ANN compared to the NIEC and Rev-NIEC indices.ConclusionsThe ANN model accurately predicted the 1-year risk for EGVB in liver cirrhosis patients and might be used as a basis for risk-based EGVB surveillance strategies.

A risk-based post-resection follow-up strategy for hepatocellular carcinoma patients.

The optimal intervals for follow-up after hepatocellular carcinoma (HCC) patients undergo curative liver resection (LR) remain unclear. This study aimed to establish a risk-based post-resection follow-up strategy. Patients that were diagnosed with HCC and received LR from three hospitals in China were included. The risk-based strategy was established based on the random survival forest model and compared with a fixed strategy both internally and externally. In total, 3447 patients from three hospitals were included. The authors' strategy showed superiority in the early detection of tumor relapse compared with fixed surveillance. Under fewer total visits, risk-based strategy achieved analogous survival time compared to the total 20 times follow-ups based on fixed strategy. Twelve total visits (five, three, one, two, and one visits in years 1-5, respectively) for American Joint Committee on Cancer/International Union Against Cancer T1a stage patients, 13 total visits (five, four, one, two, and one visits in years 1-5, respectively) for T1b stage patients, 15 total visits (eight, three, three, zero, and one visits in years 1-5, respectively) for T2 stage patients, and 15 total visits (eight, four, one, one, and one visits in years 1-5, respectively) for T3 stage patients were advocated. The detailed follow-up arrangements were available to the public through an interactive website (https://sysuccfyz.shinyapps.io/RiskBasedFollowUp/). This risk-based surveillance strategy was demonstrated to detect relapse earlier and reduce the total number of follow-ups without compromising on survival. Based on the strategy and methodology of the authors, surgeons or patients could choose more intensive or flexible schedules depending on the requirements and economic conditions. A risk-based post-resection follow-up strategy was established by random survival forest model using a larger hepatocellular carcinoma population The strategy was demonstrated to detect tumor relapse earlier and reduce the total number of follow-ups without compromising on survival Our strategy and methodology could be widely applied by other surgeons and patients.

Brexucabtagene Autoleucel for Relapsed/Refractory Mantle Cell Lymphoma: Real-World Outcomes in the United Kingdom

Background: Brexucabtagene autoleucel (brexu-cel, Tecartus) was granted EMA approval in December 2020 for relapsed/refractory (r/r) mantle cell lymphoma (MCL) after failure of 2 lines of therapy, including a BTKi, based on the results of ZUMA-2. In the UK, patients are prospectively approved for treatment by a centralised National CAR-T Clinical Panel (NCCP) using uniform eligibility criteria. We report the safety and efficacy outcomes in UK practice with a focus on intention-to-treat (ITT). Methods: Consecutive patients with r/r MCL from 11 UK centres referred to the NCCP (or Scottish equivalent) between February 2021 and June 2022 and deemed eligible for CAR T were analysed. Eligibility required ECOG PS of 0/1, absence of active CNS disease and ≥2 lines of therapy including prior BTKi (O'Reilly et al, doi: 10.1111/bjh.18378). Organ function requirements were at the discretion of the treating centre. CRS and ICANS were graded using ASTCT criteria. Cumulative steroid dose, administered for CAR-T toxicity, was reported as dexamethasone equivalent. Results: 82 patients (M=59, F=23) of median age 67 years (range 46-80) were approved. At submission, 24.4% (n=20) had blastoid/pleomorphic morphology, 39% (n=32) had high-risk sMIPI score, 17% (n=14) had TP53 mutation/overexpression and 35.4% (n=29) had Ki-67≥30%. Median number of prior lines was 2 (range 2-7) and 26.8% (n=22) were primary refractory to ibrutinib. 86.6% (n=71) and 59.8% (n=49) were harvested and infused respectively (Figure 1). Disease features associated with failure to reach harvest or infusion include blastoid morphology (OR 6.25, 95% CI 1.00-39.09, p=0.05) and a trend towards ECOG PS>0 (OR 4.42, 95% CI 0.88-22.32, p=0.07). Manufacturing failure (MF) occurred in 24.6% (n=17) with subsequent infusion in 47% (n=8). There was no association between MF and lines of therapy, bendamustine exposure, subtype, bulky disease >5cm, peripheral blood CD3+ count or total CD3+ cells harvested. In univariate analysis (UVA), MF was associated with LDH>ULN (OR 3.62, 95% CI, 1.02-12.93, p=0.05). 41.2% of MF (n=7) had circulating disease at T-cell harvest. 61.1% of all patients with circulating disease (n=11/18) had a successful 1st manufacture with a median ALC of 4.2 x 109/L (range 0.8-32.4) 94.4% received bridging therapy (n=67) with an overall response rate (ORR) of 33.3% (n=21/63). Median time from approval and 1st harvest to cell infusion was 56 (range 41-245) and 38 days (range 28-120) respectively. CRS and ICANS were observed in 94% (12%≥grade 3) and 56% (24%≥grade 3) respectively. 77.6% (n=38) received tocilizumab, and steroids were used in 59.2% (n=29), administered at day 7 (median; range 1-17), at a median dose of 165mg (range 10-1085) over a median of 8 days (range 1-270). 16.3% (n=8) received anakinra as an adjunctive agent for ICANS. 12 (24.5%) were admitted to ICU. Non-relapse mortality was 6.1% (n=3; COVID n=1, other infection n=2). Best ORR was 89.6% (83.3% complete remission, 6.3% partial remission). At a median follow-up of 9.1 months (range 0-14.8) from approval, median PFS was 9.9 months (95% CI 5.1-N/A). Median PFS from infusion was not reached at median follow-up of 7.4 months (range 0.9-13.2) with 6 and 12-month PFS rates of 82.5% (95% CI 67.9%-90.9%) and 56.1% (95% CI 34.1%-73.3%) respectively. In UVA, inferior PFS from infusion was associated with ≥3 sites of extra-nodal disease (HR 3.92, 95% CI 1.30-11.81, p=0.02) and the presence of circulating disease at T-cell harvest (HR 7.05, 95% CI 2.17-22.90, p=0.001) ; ICANS (any grade) with a superior PFS (HR 0.28, 95% CI 0.09-0.83, p=0.02) (Table 1). Median OS from infusion was also not reached with 6-month and 12-month OS rates of 92.8% (95% CI 79.2%-97.6%) and 72.3% (95% CI 47.1%-86.9%) respectively. Conclusion: Efficacy and safety outcomes for those infused with brexu-cel in the UK are comparable with ZUMA-2 and real-world datasets. However, prospective ITT analysis highlights the challenge of disease control in r/r MCL. Earlier referral with lower disease burden and better bridging strategies may reduce ITT failures, improve manufacturing success and overall outcomes. British Society of Haematology MCL guidance proposes a risk-based surveillance strategy for potential CAR T candidates at first relapse (O'Reilly et al, doi: 10.1111/bjh.18378 ). Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Hepatocellular Carcinoma Risk Scores Predict Patients Under Surveillance at Low Risk of Benefit and High Risk of Harm.

Surveillance for hepatocellular carcinoma (HCC) is recommended for patients with cirrhosis. Multiple risk scores aim to stratify HCC risk, potentially allowing individualized surveillance strategies. We sought to validate four risk scores and quantify the consequences of surveillance via the calculation of numbers needed to benefit (NNB) and harm (NNH) according to classification by risk score strata. Data were collected on 482 patients with cirrhosis during 2013-2014, with follow-up until 31/12/2019. Risk scores (aMAP, Toronto risk index, ADRESS HCC, HCC risk score) were derived from index clinic results. The area under the receiving operating characteristic curve (AUC) was calculated for each. Additionally, per-risk strata, NNB was calculated as total surveillance ultrasounds per surveillance diagnosed early HCC (stage 0/A) and NNH as total ultrasounds performed per false positive (abnormal surveillance with normal follow-up imaging). 22 (4.6%) patients developed HCC. 77% (17/22) were diagnosed through surveillance, of which 13/17 (76%) were early stage. There were 88 false positives and no false negatives (normal surveillance result however subsequent HCC detection). Overall NNB and NNH were 241 and 36, respectively. No score was significantly superior using AUC. Patients classified as low risk demonstrated no surveillance benefit (AMAP, THRI) or had a high NNB of > 300/900 (ADRESS HCC, HCC risk score), with low NNH (24-38). Given the lack of benefit and increased harm through false positives in low-risk groups, a risk-based surveillance strategy may have the potential to reduce patient harm and increase benefit from HCC surveillance. This was not a clinical trial and the study was not pre-registered.

Validating the association of adverse pathology with distant metastasis and prostate cancer mortality 20-years after radical prostatectomy.

To assess the association of adverse pathology (AP), defined as high-grade (≥ Gleason Grade Group 3) and/or non-organ confined disease, with long-term oncologic outcomes after radical prostatectomy (RP). Using a stratified cohort sampling design, we evaluated the association of AP with the risk of distant metastasis (DM) and prostate cancer-specific mortality (PCSM) up to 20 years after RP in 428 patients treated between 1987 to 2004. Cox regression of cause-specific hazards was used to estimate the absolute risk of both endpoints, with death from other causes treated as a competing risk. Additionally, subgroup analysis in patients with low and/or intermediate-risk disease, who are potentially eligible for active surveillance (AS), was performed. Within the cohort sample, 53% of men exhibited AP at time of RP, with median follow up of 15.5 years (IQR 14.6-16.6 years) thereafter. Adverse pathology was highly associated with DM and PCSM in the overall cohort (HR 12.30, 95% confidence interval [CI] 5.30-28.55, and HR 10.03, 95% CI 3.42-29.47, respectively, both P < 0.001). Adverse pathology was also highly associated with DM and PCSM in the low/intermediate-risk subgroup (HR 10.48, 95% CI 4.18-26.28, and 8.60, 95% CI 2.40-30.48, respectively, both P < 0.001). Adverse pathology at the time of RP is highly associated with future development of DM and PCSM. Accurate prediction of AP may thus be useful for individualizing risk-based surveillance and treatment strategies.

Evaluation of risk-based surveillance strategies for Salmonella Dublin in Danish dairy herds by modelling temporal test performance and herd status classification errors

The potential risk-based improvement of the Salmonella Dublin surveillance programme in Danish dairy herds was investigated, considering herd status misclassifications due to testing errors. The programme started in October 2002. Currently (early 2021) all dairy herds are classified based on quarterly bulk tank milk (BTM) testing with an indirect antibody ELISA (iELISA). Over the last two decades, the prevalence of herds classified as “likely infected” (levels 2,3) reduced remarkably. However, since 2015, the apparent prevalence has increased again, calling for improved surveillance and control to protect animal and human health. A deterministic simulation model based on data (2018–2019) from 2283 dairy herds in level 1 (“most likely free from infection”), was developed to estimate status misclassifications as false negative (FN) and false positive (FP) herds, under two testing strategies. These were: (A) the current system based on quarterly BTM testing only, and (B) an alternative strategy based on additional blood testing of up to eight calves, within herds at high risk of infection (HR). Both strategies were evaluated using three risk classification methods (I to III) and four sensitivity analysis scenarios (SA1-4), where different temporal performances were simulated for the iELISA in BTM. To apply strategy B, the best high-risk classification method (II), which combined managerial applicability and minimized errors, would require testing approximately 1000 calves across 127 HR herds. In that case, strategy A would cause 3 FNs and 67 FPs, by assuming annual BTM sensitivity (BTMSe) 95% conditional on a 1-year disease history and specificity (BTMSp) 97%. Whereas strategy B could cause a similar number of FNs, but 7 FPs more, assuming a sensitivity (Se) of 77% and specificity (Sp) of 99% in individual blood-samples (SA1). Assuming also quarterly BTMSe 53% and BTMSp 99.9% (SA4), strategy A derived 28 FNs and 2 FPs, while strategy B resulted in 6 FNs less and 8 FPs more. Therefore, strategy B could improve early detection of infected HR herds, while strategy A would avoid more unnecessary restrictions in false-positive herds. This improves knowledge on the potential use of additional blood testing in HR herds and illustrates how deterministic modelling can be used to improve disease surveillance and control.

Temporal stability of swine movement networks in the U.S.

As a consequence of multi-site pig production practiced in North America, frequent and widespread animal movements create extensive networks of interaction between farms. Social network analysis (SNA) has been used to understand disease transmission risks within these complex and dynamic production ecosystems and is particularly relevant for designing risk-based surveillance and control strategies targeting highly connected farms. However, inferences from SNA and the effectiveness of targeted strategies may be influenced by temporal changes in network structure. Since farm movements represent a temporally dynamic network, it is also unclear how many months of data are required to gain an accurate picture of an individual farm’s connectivity pattern and the overall network structure. The extent to which shipments between two specific farms are repeated (i.e., “loyalty” of farm contacts) can influence the rate at which the structure of a network changes over time, which may influence disease dynamics. In this study, we aimed to describe temporal stability and loyalty patterns of pig movement networks in the U.S. swine industry.We analyzed a total of 282,807 animal movements among 2724 farms belonging to two production systems between 2014 and 2017. Loyalty trends were largely driven by contacts between sow farms and nurseries and between nurseries and finisher farms; mean loyalty (percent of contacts that were repeated at least once within a 52-week interval) of farm contacts was 51–60 % for farm contacts involving weaned pigs, and 12–22% for contacts involving feeder pigs. A cyclic pattern was observed for both weaned and feeder pig movements, with episodes of increased loyalty observed at intervals of 8 and 17–20 weeks, respectively. Network stability was achieved when six months of data were aggregated, and only small shifts in node-level and global network metrics were observed when adding more data. This stability is relevant for designing targeted surveillance programs for disease management, given that movements summarized over too short a period may lead to stochastic swings in network metrics. A temporal resolution of six months would be reliable for the identification of potential super-spreaders in a network for targeted intervention and disease control. The temporal stability observed in these networks suggests that identifying highly connected farms in retrospective network data (up to 24 months) is reliable for future planning, albeit with reduced effectiveness.

Modelling the Spatial Distribution of ASF-Positive Wild Boar Carcasses in South Korea Using 2019-2020 National Surveillance Data.

Simple SummarySince African swine fever (ASF) virus in wild boar populations can spill over to domestic pigs, it is crucial to understand the disease determinants in the wild compartment. However, the imperfect detection sensitivity of wild boar surveillance jeopardizes our ability to understand ASF spatial distribution. In this study, we used national surveillance data of ASF in wild boars collected in the Republic of Korea from 2019–2020 to model the spatial distribution of ASF-positive carcasses for two successive study periods associated with different surveillance intensity. The model allowed us to identify disease risk factors in the Republic of Korea, determine the spatial distribution of the risk of ASF, and estimate the sensitivity of surveillance. The outputs of this study are relevant to policy makers for developing and improving risk-based surveillance programs for ASF in wild boars.In September 2019, African swine fever (ASF) was reported in South Korea for the first time. Since then, more than 651 ASF cases in wild boars and 14 farm outbreaks have been notified in the country. Despite the efforts to eradicate ASF among wild boar populations, the number of reported ASF-positive wild boar carcasses have increased recently. The purpose of this study was to characterize the spatial distribution of ASF-positive wild boar carcasses to identify the risk factors associated with the presence and number of ASF-positive wild boar carcasses in the affected areas. Because surveillance efforts have substantially increased in early 2020, we divided the study into two periods (2 October 2019 to 19 January 2020, and 19 January to 28 April 2020) based on the number of reported cases and aggregated the number of reported ASF-positive carcasses into a regular grid of hexagons of 3-km diameter. To account for imperfect detection of positive carcasses, we adjusted spatial zero-inflated Poisson regression models to the number of ASF-positive wild boar carcasses per hexagon. During the first study period, proximity to North Korea was identified as the major risk factor for the presence of African swine fever virus. In addition, there were more positive carcasses reported in affected hexagons with high habitat suitability for wild boars, low heat load index (HLI), and high human density. During the second study period, proximity to an ASF-positive carcass reported during the first period was the only significant risk factor for the presence of ASF-positive carcasses. Additionally, low HLI and elevation were associated with an increased number of ASF-positive carcasses reported in the affected hexagons. Although the proportion of ASF-affected hexagons increased from 0.06 (95% credible interval (CrI): 0.05–0.07) to 0.09 (95% CrI: 0.08–0.10), the probability of reporting at least one positive carcass in ASF-affected hexagons increased from 0.49 (95% CrI: 0.41–0.57) to 0.73 (95% CrI: 0.66–0.81) between the two study periods. These results can be used to further advance risk-based surveillance strategies in the Republic of Korea.

Quantitative risk assessment of the introduction of low pathogenic avian influenza H5 and H7 strains into Poland via legal import of live poultry

Low pathogenic avian influenza (LPAI) caused by H5 and H7 viruses is considered a threatening disease for poultry production due to the possibility of prolonged undetected virus circulation in a poultry flock and its potential to mutate to highly pathogenic avian influenza (HPAI). The occurrence of HPAI may have devastating impact on the poultry industry and has serious economic consequences. The possibility of LPAI virus (LPAIV) being introduced into Poland via import of live poultry from EU countries was considered. The main aim of the study was to quantitatively assess the probability of LPAIV H5 and H7 introduction into Poland (PLPAI) via this pathway, to evaluate the relative contribution of exporting countries and species of poultry to this probability and to present the spatial distribution of the introduction probability in Poland. To this end, a stochastic multilevel binomial risk model, taking into account uncertainty and variability of input parameter values, was developed. The results of this model indicate that the mean annual probability of LPAIV H5 or H7 introduction into Poland is 0.088 [95 % uncertainty interval: 0.0575, 0.128], which corresponds to, on average, one outbreak every 11 years. The countries contributing most to this probability are Germany, Czech Republic and Denmark. Importations of ducks, chickens and turkeys contribute most to PLPAI, whereas importations of geese and guinea fowl represent a minor risk. The probability of LPAIV introduction is not equally distributed across Poland with the majority of counties having a high probability of LPAIV introduction being located in the Western part of the country.The results of this study can be used to support decision makers on targeted prevention or risk-based surveillance strategies for LPAI.

Provider Attitudes Toward Risk-Based Hepatocellular Carcinoma Surveillance in Patients With Cirrhosis in the United States

Hepatocellular carcinoma (HCC) surveillance rates are suboptimal in clinical practice. We aimed to elicit providers' opinions on the following aspects of HCC surveillance: preferred strategies, barriers and facilitators, and the impact of a patient's HCC risk on the choice of surveillance modality. We conducted a web-based survey among gastroenterology and hepatology providers (40% faculty physicians, 21% advanced practice providers, 39% fellow-trainees) from 26 US medical centers in 17 states. Of 654 eligible providers, 305 (47%) completed the survey. Nearly all (98.4%) of the providers endorsed semi-annual HCC surveillance in patients with cirrhosis, with 84.2% recommending ultrasound ± alpha fetoprotein (AFP) and 15.4% recommending computed tomography (CT) or magnetic resonance imaging (MRI). Barriers to surveillance included limited HCC treatment options, screening test effectiveness to reduce mortality, access to transportation, and high out-of-pocket costs. Facilitators of surveillance included professional society guidelines. Most providers (72.1%) would perform surveillance even if HCC risk was low (≤0.5% per year), while 98.7% would perform surveillance if HCC risk was ≥1% per year. As a patient's HCC risk increased from 1% to 3% to 5% per year, providers reported they would be less likely to order ultrasound ± AFP (83.6% to 68.9% to 57.4%; P < .001) and more likely to order CT or MRI ± AFP (3.9% to 26.2% to 36.1%; P < .001). Providers recommend HCC surveillance even when HCC risk is much lower than the threshold suggested by professional societies. Many appear receptive to risk-based HCC surveillance strategies that depend on patients' estimated HCC risk, instead of our current "one-size-fits all" strategy.

Assessment of a Deep Learning Model to Predict Hepatocellular Carcinoma in Patients With Hepatitis C Cirrhosis

Deep learning, a family of machine learning models that use artificial neural networks, has achieved great success at predicting outcomes in nonmedical domains. To examine whether deep learning recurrent neural network (RNN) models that use raw longitudinal data extracted directly from electronic health records outperform conventional regression models in predicting the risk of developing hepatocellular carcinoma (HCC). This prognostic study included 48 151 patients with hepatitis C virus (HCV)-related cirrhosis in the national Veterans Health Administration who had at least 3 years of follow-up after the diagnosis of cirrhosis. Patients were identified by having at least 1 positive HCV RNA test between January 1, 2000, to January 1, 2016, and were followed up from the diagnosis of cirrhosis to January 1, 2019, for the development of incident HCC. A total of 3 models predicting HCC during a 3-year period were developed and compared, as follows: (1) logistic regression (LR) with cross-sectional inputs (cross-sectional LR); (2) LR with longitudinal inputs (longitudinal LR); and (3) RNN with longitudinal inputs. Data analysis was conducted from April 2018 to August 2020. Development of HCC. Area under the receiver operating characteristic curve, area under the precision-recall curve, and Brier score. During a mean (SD) follow-up of 11.6 (5.0) years, 10 741 of 48 151 patients (22.3%) developed HCC (annual incidence, 3.1%), and a total of 52 983 samples (51 948 [98.0%] from men) were collected. Patients who developed HCC within 3 years were older than patients who did not (mean [SD] age, 58.2 [6.6] years vs 56.9 [6.9] years). RNN models had superior mean (SD) area under the receiver operating characteristic curve (0.759 [0.009]) and mean (SD) Brier score (0.136 [0.003]) than cross-sectional LR (0.689 [0.009] and 0.149 [0.003], respectively) and longitudinal LR (0.682 [0.007] and 0.150 [0.003], respectively) models. Using the RNN model, the samples with the mean (SD) highest 51% (1.5%) of HCC risk, in which 80% of all HCCs occurred, or the mean (SD) highest 66% (1.2%) of HCC risk, in which 90% of all HCCs occurred, could potentially be targeted. Among samples from patients who achieved sustained virologic response, the performance of the RNN models was even better (mean [SD] area under receiver operating characteristic curve, 0.806 [0.025]; mean [SD] Brier score, 0.117 [0.007]). In this study, deep learning RNN models outperformed conventional LR models, suggesting that RNN models could be used to identify patients with HCV-related cirrhosis with a high risk of developing HCC for risk-based HCC outreach and surveillance strategies.

Evaluation of the effectiveness of the surveillance system for tuberculosis in cattle in Spain

In Spain, the eradication of tuberculosis (TB) in cattle has had some setbacks and therefore we need to re-evaluate the current surveillance strategies. The Spanish cattle TB surveillance system consists of three components: a) routine skin testing, b) slaughterhouse surveillance, and c) pre-movement controls. Our objectives were to assess the effectiveness of the whole surveillance system and evaluate the relative contribution of each of those components, both at national and at the province level. The effectiveness was estimated based on both the sensitivity of detection per year and the time until detection. Moreover, we evaluated the impact of different factors on that effectiveness. We used a stochastic model that simulated the spread of Mycobacterium tuberculosis-complex (MTC) infection within cattle herds that was modified to incorporate the three components of TB surveillance. Input data, at the province level, on cattle movements off-farm or the frequency of the routine screening were provided by the Spanish Ministry for Agriculture, Fisheries and Food.Under the current conditions, 96.1% of Spanish herds were detected within one year after their infection (i.e. mean sensitivity of TB surveillance), although that detection took on average 178 days. The surveillance system effectiveness was highly dependent on the routine skin testing, responsible for the detection of almost 90% of the infected herds, while slaughterhouse surveillance and pre-movement controls contributed only to the identification of a small proportion of infected herds. We observed substantial differences in the effectiveness of the surveillance components among Spanish provinces, although in general, the sensitivities were high. The most influential factor on the efficiency of TB detection in Spain was, by far, the frequency of routine controls, followed by the sensitivity of the test used.In a context of reduced funding for cattle TB eradication, the frequency of testing should be adapted based on risk-based surveillance strategies, i.e. efforts should focus on herds more likely to be infected, but also in herds more likely to infect other herds. While slaughterhouse surveillance seems a cost-effective strategy, the use of pre-movement testing at least in areas of low incidence should be further evaluated.

From the Editor’s Desk: September 2019

From the Editor’s Desk: September 2019

Models estimating risk of hepatocellular carcinoma in patients with alcohol or NAFLD-related cirrhosis for risk stratification

Hepatocellular carcinoma (HCC) risk varies dramatically in patients with cirrhosis according to well-described, readily available predictors. We aimed to develop simple models estimating HCC risk in patients with alcohol-related liver disease (ALD)-cirrhosis or non-alcoholic fatty liver disease (NAFLD)-cirrhosis and calculate the net benefit that would be derived by implementing HCC surveillance strategies based on HCC risk as predicted by our models. We identified 7,068 patients with NAFLD-cirrhosis and 16,175 with ALD-cirrhosis who received care in the Veterans Affairs (VA) healthcare system in 2012. We retrospectively followed them for the development of incident HCC until January 2018. We used Cox proportional hazards regression to develop and internally validate models predicting HCC risk using baseline characteristics at entry into the cohort in 2012. We plotted decision curves of net benefit against HCC screening thresholds. We identified 1,278 incident cases of HCC during a mean follow-up period of 3.7 years. Mean annualized HCC incidence was 1.56% in NAFLD-cirrhosis and 1.44% in ALD-cirrhosis. The final models estimating HCC were developed separately for NAFLD-cirrhosis and ALD-cirrhosis and included 7 predictors: age, gender, diabetes, body mass index, platelet count, serum albumin and aspartate aminotransferase to √alanine aminotransferase ratio. The models exhibited very good measures of discrimination and calibration and an area under the receiver operating characteristic curve of 0.75 for NAFLD-cirrhosis and 0.76 for ALD-cirrhosis. Decision curves showed higher standardized net benefit of risk-based screening using our prediction models compared to the screen-all approach. We developed simple models estimating HCC risk in patients with NAFLD-cirrhosis or ALD-cirrhosis, which are available as web-based tools (www.hccrisk.com). Risk stratification can be used to inform risk-based HCC surveillance strategies in individual patients or healthcare systems or to identify high-risk patients for clinical trials. Patients with cirrhosis of the liver are at risk of getting hepatocellular carcinoma (HCC or liver cancer) and therefore it is recommended that they undergo surveillance for HCC. However, the risk of HCC varies dramatically in patients with cirrhosis, which has implications on if and how patients get surveillance, how providers counsel patients about the need for surveillance, and how healthcare systems approach and prioritize surveillance. We used readily available predictors to develop models estimating HCC risk in patients with cirrhosis, which are available as web-based tools at www.hccrisk.com.

Identifying Lookouts for Epidemio-Surveillance: Application to the Emergence of Xylella fastidiosa in France.

Recent detections of Xylella fastidiosa in Corsica Island, France, has raised concerns on its possible spread to mainland France and the rest of the Mediterranean Basin. Early detection of infected plants is paramount to prevent the spread of the bacteria, but little is known about this pathosystem in European environments, hence standard surveillance strategies may be ineffective. We present a new methodological approach for the design of risk-based surveillance strategies, adapted to the emerging risk caused by X. fastidiosa. Our proposal is based on a combination of machine learning techniques and network analysis that aims at understanding the main abiotic drivers of the infection, produce risk maps and identify lookouts for the design of future surveillance plans. The identified drivers coincide with known results in laboratory studies about the correlation between environmental variables, such as water stress and temperature, and the presence of the bacterium in plants. Furthermore, the produced risk maps overlap nicely with detected foci of infection, while they also highlight other susceptible regions where X. fastidiosa has not been found yet. We conclude the paper presenting a list of recommended regions for a risk-based surveillance campaign based on the predicted spread and probability of early detection of the disease.

Space-time cluster analysis to improve classical swine fever risk-based surveillance in Ecuador

Space-time cluster analysis to improve classical swine fever risk-based surveillance in Ecuador