Amy Douglas
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Amy Douglas has a total of 58 co-authors. Here are the top co-authors by number of shared publications:
- Claire Jenkins
Gastrointestinal Bacteria Reference Unit, UK Health Security Agency , London NW9 5HT
4 shared publications - Orlagh Quinn
UK Health Security Agency, UK
3 shared publications - Roberto Vivancos
UK Health Security Agency
3 shared publications - Alex James Elliot
Real-time Syndromic Surveillance Team, UK Health Security Agency, Birmingham, United Kingdom
2 shared publications - Clare Sawyer
Field Service South-East and London, Health Protection Operations Division, UK Health Security Agency, London, England, United Kingdom
2 shared publications - Craig Swift
UK Health Security Agency, Colindale Avenue, London NW9 5EQ, UK
2 shared publications
Latest Publications
- Open Access
Case Fatality Risk of Norovirus in England During a Period of Strain Replacement, 2022/23 - 2024/25 Seasons
Background Norovirus causes substantial burden to healthcare systems. England experienced high activity in recent seasons alongside a shift in the dominant genotype from GII.4 to GII.17. It remains unclear whether this increased burden reflects changes in severity or other transmission mechanisms associated with the strain replacement. Methods Individual-level testing and mortality data in England from 2022/23-2024/25 seasons were linked from national surveillance systems. Piece-wise exponential additive mixed models estimated all-cause case-fatality risk (CFR) for death within 28 days of a positive test, adjusting for age, location, primary or secondary care case identification and temporal effects. Additional analysis tested the effect of the death linkage threshold. Results CFR did not differ significantly between GII.4 and GII.17, but increased markedly with age and was substantially higher for cases detected in secondary versus primary care. At the 2024/25 season peak, the CFR for the average cohort case (age 75) was 8.01% (95% CI: 6.88-9.33%) for cases identified in secondary care compared with 1.18% (95% CI: 0.61-2.25%) for primary care, likely driven by different case mixes and unmeasured confounding between patients managed in these settings. Interpretation We found no evidence that the newly dominant GII.17 genotype had higher CFR than GII.4 among test-positive cases. The increased burden is therefore more consistent with changes in transmissibility, population immunity, or testing dynamics than increased genotype-specific severity. Norovirus remains a major public health threat, and understanding genotype switching supports enhanced surveillance, expanded genotyping, routine clinical data linkage and advanced modelling to inform control strategies.
Open AccessCase Fatality Risk of Norovirus in England During a Period of Strain Replacement, 2022/23 - 2024/25 Seasons
Background Norovirus causes substantial burden to healthcare systems. England experienced high activity in recent seasons alongside a shift in the dominant genotype from GII.4 to GII.17. It remains unclear whether this increased burden reflects changes in severity or other transmission mechanisms associated with the strain replacement. Methods Individual-level testing and mortality data in England from 2022/23-2024/25 seasons were linked from national surveillance systems. Piece-wise exponential additive mixed models estimated all-cause case-fatality risk (CFR) for death within 28 days of a positive test, adjusting for age, location, primary or secondary care case identification and temporal effects. Additional analysis tested the effect of the death linkage threshold. Results CFR did not differ significantly between GII.4 and GII.17, but increased markedly with age and was substantially higher for cases detected in secondary versus primary care. At the 2024/25 season peak, the CFR for the average cohort case (age 75) was 8.01% (95% CI: 6.88-9.33%) for cases identified in secondary care compared with 1.18% (95% CI: 0.61-2.25%) for primary care, likely driven by different case mixes and unmeasured confounding between patients managed in these settings. Interpretation We found no evidence that the newly dominant GII.17 genotype had higher CFR than GII.4 among test-positive cases. The increased burden is therefore more consistent with changes in transmissibility, population immunity, or testing dynamics than increased genotype-specific severity. Norovirus remains a major public health threat, and understanding genotype switching supports enhanced surveillance, expanded genotyping, routine clinical data linkage and advanced modelling to inform control strategies.
- Open Access
- 15 July 2025
- Published by Elsevier in Journal of Infection
- Vol. 91 (2) , 106551
The epidemiology of haemolytic uraemic syndrome in England, 2009-2023: An analysis of national surveillance data
- 15 July 2025
- Published by Elsevier in Journal of Infection
- Vol. 91 (2) , 106551
Open AccessThe epidemiology of haemolytic uraemic syndrome in England, 2009-2023: An analysis of national surveillance data
- Open Access
- 12 June 2025
- Published by Microbiology Society in Journal of Medical Microbiology
- Vol. 74 (6) , 002017
Enhanced laboratory surveillance study of Campylobacter species in England
Introduction. Campylobacteriosis is the leading cause of gastroenteritis worldwide, and Campylobacter species are the most frequently reported zoonotic, bacterial foodborne pathogens in England. Gap statement. Currently, less than 2.0% of Campylobacter isolates in England undergo strain identification and typing, resulting in limited insight into their molecular epidemiology. Aim. To assess the feasibility of using high-throughput whole-genome sequencing (WGS) to generate data for microbiological and epidemiological analysis by the implementation of a 3-month enhanced laboratory surveillance for Campylobacter spp. in England, and to make recommendations for improving the current Campylobacter surveillance strategies. Methodology. All diagnostic laboratories in England were encouraged to refer isolates of Campylobacter spp. for WGS over a 3-month period (7 June–31 August 2021). Results. Over 6,000 Campylobacter species isolates were characterized, of which 87.5% were successfully identified as Campylobacter jejuni and 8.1% as Campylobacter coli. Just over half of the isolates were referred from patients who were male (53%), and C. coli isolates tended to be from older patients than C. jejuni, with median ages of 55 and 44 years, respectively. The most common multi-locus sequencing type clonal complex identified was ST-21, and within this, the sequencing type ST6175 was the most frequently identified, of which 96.8% were predicted to carry antimicrobial resistance determinants, inferring reduced susceptibility to both ciprofloxacin and tetracycline. The four largest C. jejuni 5-single nucleotide polymorphism (SNP) clusters, associated with the larger clonal complexes and sequence type groups (ST6175, ST48, ST6175 and ST5136), accounted for 23.8% (n=1,150/4,838) of SNP typable isolates. Conversely, 28.4% and 39.5% of isolates C. jejuni and C. coli, respectively, appeared to be sporadic, with each isolate assigned a unique SNP address at the 5-SNP level. Conclusion. WGS enabled identification of genetically related clusters of Campylobacter isolates in almost real time and shows potential for monitoring of inferred antimicrobial resistance. However, unlocking its full potential requires referral of sufficient and representative isolates for sequencing with parallel epidemiological data collection.
- 12 June 2025
- Published by Microbiology Society in Journal of Medical Microbiology
- Vol. 74 (6) , 002017
Open AccessEnhanced laboratory surveillance study of Campylobacter species in England
Introduction. Campylobacteriosis is the leading cause of gastroenteritis worldwide, and Campylobacter species are the most frequently reported zoonotic, bacterial foodborne pathogens in England. Gap statement. Currently, less than 2.0% of Campylobacter isolates in England undergo strain identification and typing, resulting in limited insight into their molecular epidemiology. Aim. To assess the feasibility of using high-throughput whole-genome sequencing (WGS) to generate data for microbiological and epidemiological analysis by the implementation of a 3-month enhanced laboratory surveillance for Campylobacter spp. in England, and to make recommendations for improving the current Campylobacter surveillance strategies. Methodology. All diagnostic laboratories in England were encouraged to refer isolates of Campylobacter spp. for WGS over a 3-month period (7 June–31 August 2021). Results. Over 6,000 Campylobacter species isolates were characterized, of which 87.5% were successfully identified as Campylobacter jejuni and 8.1% as Campylobacter coli. Just over half of the isolates were referred from patients who were male (53%), and C. coli isolates tended to be from older patients than C. jejuni, with median ages of 55 and 44 years, respectively. The most common multi-locus sequencing type clonal complex identified was ST-21, and within this, the sequencing type ST6175 was the most frequently identified, of which 96.8% were predicted to carry antimicrobial resistance determinants, inferring reduced susceptibility to both ciprofloxacin and tetracycline. The four largest C. jejuni 5-single nucleotide polymorphism (SNP) clusters, associated with the larger clonal complexes and sequence type groups (ST6175, ST48, ST6175 and ST5136), accounted for 23.8% (n=1,150/4,838) of SNP typable isolates. Conversely, 28.4% and 39.5% of isolates C. jejuni and C. coli, respectively, appeared to be sporadic, with each isolate assigned a unique SNP address at the 5-SNP level. Conclusion. WGS enabled identification of genetically related clusters of Campylobacter isolates in almost real time and shows potential for monitoring of inferred antimicrobial resistance. However, unlocking its full potential requires referral of sufficient and representative isolates for sequencing with parallel epidemiological data collection.
- Open Access
- 10 January 2025
- Published by Microbiology Society in Journal of Medical Microbiology
- Vol. 74 (1) , 001947
Epidemiology of Shiga toxin-producing Escherichia coli other than serotype O157:H7 in England, 2016–2023
Introduction. Shiga toxin-producing Escherichia coli (STEC) infections are of public health concern as STEC can cause large national foodborne outbreaks of severe gastrointestinal disease, particularly in the young and elderly. In recent years, the implementation of PCR by diagnostic microbiology laboratories has improved the detection of STEC, and there has been an increase in notifications of cases of non-O157 STEC. However, the extent this increase in caseload can be attributed to the improved detection by PCR, or a true increase in non-O157 STEC infections, is unknown. Gap Statement. Epidemiological and microbiological data and analyses describing the trends in non-O157 STEC in England since the implementation of PCR are limited. Aim. Demographic, microbiological and clinical characteristics of non-O157 STEC from 8 years (2016–2023) of laboratory surveillance data were analysed to understand the recent trends in non-O157 serotypes and the incidence of disease in England. Methodology. All human isolates of STEC non-O157 detected between 2016 and 2023 were extracted from the laboratory surveillance system. Microbiological data were analysed and linked to clinical outcomes. Results. There was an almost 10-fold increase in diagnoses of non-O157 STEC from 2016 (n=297) to 2023 (n=2341). A total of 9378 isolates of non-O157 STEC were detected, comprising 338 different serotypes, and were linked to 9311 individuals. A higher proportion of non-O157 STEC cases were female (56%) and aged between 20 and 39 years (27%). The most common non-O157 serotypes were O26:H11 (16%), O146:H21 (12%), O91:H14 (11%), O128:H2 (6%), O145:H28 (5%) and O103:H2 (4%). STEC O26:H11 was more frequently reported in under 5s than any other age group (38%), whereas the other common serotypes were more frequently isolated from adults. Stx2a, which has been associated with greater disease severity, was detected in 18% of cases. Where clinical details were available, 27% of non-O157 cases were admitted to the hospital and 6% developed HUS. Cases of STEC O145:H28 reported a higher rate of hospitalisation than other non-O157 STEC cases. The serotypes most likely to be associated with progression to HUS were O26:H11 (9%) and O145:H28 (7%). STEC harbouring stx2f (19%), stx2a (11%) and stx2d (11%) were most frequently isolated from cases with HUS. Conclusion. The implementation of widespread PCR testing in England has facilitated better surveillance of STEC non-O157, with respect to establishing the true incidence and burden of disease of non-O157 STEC and monitoring the emergence of highly virulent strains.
- 10 January 2025
- Published by Microbiology Society in Journal of Medical Microbiology
- Vol. 74 (1) , 001947
Open AccessEpidemiology of Shiga toxin-producing Escherichia coli other than serotype O157:H7 in England, 2016–2023
Introduction. Shiga toxin-producing Escherichia coli (STEC) infections are of public health concern as STEC can cause large national foodborne outbreaks of severe gastrointestinal disease, particularly in the young and elderly. In recent years, the implementation of PCR by diagnostic microbiology laboratories has improved the detection of STEC, and there has been an increase in notifications of cases of non-O157 STEC. However, the extent this increase in caseload can be attributed to the improved detection by PCR, or a true increase in non-O157 STEC infections, is unknown. Gap Statement. Epidemiological and microbiological data and analyses describing the trends in non-O157 STEC in England since the implementation of PCR are limited. Aim. Demographic, microbiological and clinical characteristics of non-O157 STEC from 8 years (2016–2023) of laboratory surveillance data were analysed to understand the recent trends in non-O157 serotypes and the incidence of disease in England. Methodology. All human isolates of STEC non-O157 detected between 2016 and 2023 were extracted from the laboratory surveillance system. Microbiological data were analysed and linked to clinical outcomes. Results. There was an almost 10-fold increase in diagnoses of non-O157 STEC from 2016 (n=297) to 2023 (n=2341). A total of 9378 isolates of non-O157 STEC were detected, comprising 338 different serotypes, and were linked to 9311 individuals. A higher proportion of non-O157 STEC cases were female (56%) and aged between 20 and 39 years (27%). The most common non-O157 serotypes were O26:H11 (16%), O146:H21 (12%), O91:H14 (11%), O128:H2 (6%), O145:H28 (5%) and O103:H2 (4%). STEC O26:H11 was more frequently reported in under 5s than any other age group (38%), whereas the other common serotypes were more frequently isolated from adults. Stx2a, which has been associated with greater disease severity, was detected in 18% of cases. Where clinical details were available, 27% of non-O157 cases were admitted to the hospital and 6% developed HUS. Cases of STEC O145:H28 reported a higher rate of hospitalisation than other non-O157 STEC cases. The serotypes most likely to be associated with progression to HUS were O26:H11 (9%) and O145:H28 (7%). STEC harbouring stx2f (19%), stx2a (11%) and stx2d (11%) were most frequently isolated from cases with HUS. Conclusion. The implementation of widespread PCR testing in England has facilitated better surveillance of STEC non-O157, with respect to establishing the true incidence and burden of disease of non-O157 STEC and monitoring the emergence of highly virulent strains.
- Open Access
Quantifying the impact of norovirus transmission in the community on acute kidney injury hospitalisations in England: a mathematical modelling study
Norovirus is one of the most common gastrointestinal pathogens and causes substantial morbidity and mortality worldwide. A serious consequence can be acute kidney injury (AKI), characterised by a rapid deterioration of kidney function, which is associated with prolonged hospital admissions and increased mortality. This study aimed to quantify the contribution of norovirus community transmission to AKI hospitalisations in England and its attributable cost. We developed an age-structured deterministic compartmental dynamic transmission model fitted to multiple electronic health data sources. We used a Bayesian approach to fit to hospitalisations and primary care records for AKI and gastroenteritis, as well as laboratory- confirmed norovirus surveillance data in England, from 2013-2019 and population surveys of norovirus incidence. We estimated the effect of norovirus incidence on community-acquired AKI hospitalisations and the economic burden as a product of the number of AKI hospitalisations linked to norovirus with a mean comorbidity-weighted AKI hospitalisation costs. We estimated that 25% (95% CrI: 15-36%) of norovirus infections in people aged 65 years and above were linked to community-acquired AKI hospitalisations in the winter, amounting to 313,398 (95% CrI: 187,424-449,673) hospitalisations between 2013-2019. This represents 6.5% (95% CrI: 3.9-9.4%) of all cause community-acquired AKI hospitalisations in people aged 65 years and above over this period. The healthcare cost associated with norovirus-linked AKI admissions was estimated to be £135 million (95% CrI: £81.1 million - £195 million) annually. Gastroenteritis has previously been identified as an important risk factor for AKI in vulnerable patients. Our results suggest norovirus incidence in the community plays an important role in driving AKI hospitalisations in individuals aged 65 years and above. This is the first population- level estimate of the prevalence of norovirus infections linked to hospitalisations for AKI. Norovirus presents a substantial economic burden on the healthcare system and AKI hospitalisation should be a consideration when evaluating future norovirus vaccines.
Open AccessQuantifying the impact of norovirus transmission in the community on acute kidney injury hospitalisations in England: a mathematical modelling study
Norovirus is one of the most common gastrointestinal pathogens and causes substantial morbidity and mortality worldwide. A serious consequence can be acute kidney injury (AKI), characterised by a rapid deterioration of kidney function, which is associated with prolonged hospital admissions and increased mortality. This study aimed to quantify the contribution of norovirus community transmission to AKI hospitalisations in England and its attributable cost. We developed an age-structured deterministic compartmental dynamic transmission model fitted to multiple electronic health data sources. We used a Bayesian approach to fit to hospitalisations and primary care records for AKI and gastroenteritis, as well as laboratory- confirmed norovirus surveillance data in England, from 2013-2019 and population surveys of norovirus incidence. We estimated the effect of norovirus incidence on community-acquired AKI hospitalisations and the economic burden as a product of the number of AKI hospitalisations linked to norovirus with a mean comorbidity-weighted AKI hospitalisation costs. We estimated that 25% (95% CrI: 15-36%) of norovirus infections in people aged 65 years and above were linked to community-acquired AKI hospitalisations in the winter, amounting to 313,398 (95% CrI: 187,424-449,673) hospitalisations between 2013-2019. This represents 6.5% (95% CrI: 3.9-9.4%) of all cause community-acquired AKI hospitalisations in people aged 65 years and above over this period. The healthcare cost associated with norovirus-linked AKI admissions was estimated to be £135 million (95% CrI: £81.1 million - £195 million) annually. Gastroenteritis has previously been identified as an important risk factor for AKI in vulnerable patients. Our results suggest norovirus incidence in the community plays an important role in driving AKI hospitalisations in individuals aged 65 years and above. This is the first population- level estimate of the prevalence of norovirus infections linked to hospitalisations for AKI. Norovirus presents a substantial economic burden on the healthcare system and AKI hospitalisation should be a consideration when evaluating future norovirus vaccines.
- Top Cited PapersOpen Access
- 26 September 2024
- Published by European Centre for Disease Control and Prevention (ECDC) in Eurosurveillance
- Vol. 29 (39) , 2400625
Increased circulation of GII.17 noroviruses, six European countries and the United States, 2023 to 2024
We report an increase in GII.17 norovirus outbreaks and sporadic infections of acute gastroenteritis in Austria, Germany, France, Ireland, the Netherlands, England and the United States during the 2023/24 season. A decrease in GII.4 coincided with GII.17 prevalence increasing to between 17% and 64% of all GII detections. Overall, 84% of the GII.17 strains clustered closely with strains first reported in Romania in 2021 and two new sub-lineages were identified. Norovirus surveillance and molecular characterisation should be prioritised this winter.
- 26 September 2024
- Published by European Centre for Disease Control and Prevention (ECDC) in Eurosurveillance
- Vol. 29 (39) , 2400625
Top Cited PapersOpen AccessIncreased circulation of GII.17 noroviruses, six European countries and the United States, 2023 to 2024
We report an increase in GII.17 norovirus outbreaks and sporadic infections of acute gastroenteritis in Austria, Germany, France, Ireland, the Netherlands, England and the United States during the 2023/24 season. A decrease in GII.4 coincided with GII.17 prevalence increasing to between 17% and 64% of all GII detections. Overall, 84% of the GII.17 strains clustered closely with strains first reported in Romania in 2021 and two new sub-lineages were identified. Norovirus surveillance and molecular characterisation should be prioritised this winter.
- Open Access
- 21 May 2024
- Published by Microbiology Society in Microbial Genomics
- Vol. 10 (5) , 001243
Genomic analysis of an outbreak of Shiga toxin-producing Escherichia coli O183:H18 in the United Kingdom, 2023
In June 2023, UKHSA surveillance systems detected an outbreak of severe gastrointestinal symptoms caused by a rare serotype of Shiga toxin-producing Escherichia coli, STEC O183:H18. There were 26 cases aged 6 months to 74 years (42 % cases were aged 0–9 years), distributed across the UK with onset dates range between 22 May 2023 and 4 July 2023. The epidemiological and food chain investigations were inconclusive, although meat products made from beef mince were implicated as a potential vehicle. The outbreak strain belonged to sequence type (ST) 657 and harboured a Shiga toxin (stx) subtype stx2a located on a prophage that was unique in the UKHSA stx-encoding bacteriophage database. Plasmid encoded, putative virulence genes subA, ehxA, saa, iha, lpfA and iss were detected, however, the established STEC virulence genes involved in attachment to the gut mucosa (eae and aggR) were absent. The acquisition of stx across the global population structure of ST657 appeared to correspond with the presence of subA, ehxA, saa, iha, lpfA and iss. During the outbreak investigation, we used long read sequencing to characterise the plasmid and prophage content of this atypical STEC, to look for evidence to explain its recent emergence. Although we were unable to determine source and transmission route of the outbreak strain, the genomic analysis revealed potential clues as to how novel strains for STEC evolve. With the implementation of PCR capable of detecting all STEC, and genome sequencing for typing and virulence profiling, we have the tools to enable us to monitor the changing landscape of STEC. Improvements in the standardised collection of epidemiological data and trace-back strategies within the food industry, will ensure we have a surveillance system capable of alerting us to emerging threats to public health.
- 21 May 2024
- Published by Microbiology Society in Microbial Genomics
- Vol. 10 (5) , 001243
Open AccessGenomic analysis of an outbreak of Shiga toxin-producing Escherichia coli O183:H18 in the United Kingdom, 2023
In June 2023, UKHSA surveillance systems detected an outbreak of severe gastrointestinal symptoms caused by a rare serotype of Shiga toxin-producing Escherichia coli, STEC O183:H18. There were 26 cases aged 6 months to 74 years (42 % cases were aged 0–9 years), distributed across the UK with onset dates range between 22 May 2023 and 4 July 2023. The epidemiological and food chain investigations were inconclusive, although meat products made from beef mince were implicated as a potential vehicle. The outbreak strain belonged to sequence type (ST) 657 and harboured a Shiga toxin (stx) subtype stx2a located on a prophage that was unique in the UKHSA stx-encoding bacteriophage database. Plasmid encoded, putative virulence genes subA, ehxA, saa, iha, lpfA and iss were detected, however, the established STEC virulence genes involved in attachment to the gut mucosa (eae and aggR) were absent. The acquisition of stx across the global population structure of ST657 appeared to correspond with the presence of subA, ehxA, saa, iha, lpfA and iss. During the outbreak investigation, we used long read sequencing to characterise the plasmid and prophage content of this atypical STEC, to look for evidence to explain its recent emergence. Although we were unable to determine source and transmission route of the outbreak strain, the genomic analysis revealed potential clues as to how novel strains for STEC evolve. With the implementation of PCR capable of detecting all STEC, and genome sequencing for typing and virulence profiling, we have the tools to enable us to monitor the changing landscape of STEC. Improvements in the standardised collection of epidemiological data and trace-back strategies within the food industry, will ensure we have a surveillance system capable of alerting us to emerging threats to public health.
- Open Access
- 8 May 2023
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 25 (1) , e37540
Predicting Norovirus in England Using Existing and Emerging Syndromic Data: Infodemiology Study
Background: Norovirus is associated with approximately 18% of the global burden of gastroenteritis and affects all age groups. There is currently no licensed vaccine or available antiviral treatment. However, well-designed early warning systems and forecasting can guide nonpharmaceutical approaches to norovirus infection prevention and control. Objective: This study evaluates the predictive power of existing syndromic surveillance data and emerging data sources, such as internet searches and Wikipedia page views, to predict norovirus activity across a range of age groups across England. Methods: We used existing syndromic surveillance and emerging syndromic data to predict laboratory data indicating norovirus activity. Two methods are used to evaluate the predictive potential of syndromic variables. First, the Granger causality framework was used to assess whether individual variables precede changes in norovirus laboratory reports in a given region or an age group. Then, we used random forest modeling to estimate the importance of each variable in the context of others with two methods: (1) change in the mean square error and (2) node purity. Finally, these results were combined into a visualization indicating the most influential predictors for norovirus laboratory reports in a specific age group and region. Results: Our results suggest that syndromic surveillance data include valuable predictors for norovirus laboratory reports in England. However, Wikipedia page views are less likely to provide prediction improvements on top of Google Trends and Existing Syndromic Data. Predictors displayed varying relevance across age groups and regions. For example, the random forest modeling based on selected existing and emerging syndromic variables explained 60% variance in the ≥65 years age group, 42% in the East of England, but only 13% in the South West region. Emerging data sets highlighted relative search volumes, including “flu symptoms,” “norovirus in pregnancy,” and norovirus activity in specific years, such as “norovirus 2016.” Symptoms of vomiting and gastroenteritis in multiple age groups were identified as important predictors within existing data sources. Conclusions: Existing and emerging data sources can help predict norovirus activity in England in some age groups and geographic regions, particularly, predictors concerning vomiting, gastroenteritis, and norovirus in the vulnerable populations and historical terms such as stomach flu. However, syndromic predictors were less relevant in some age groups and regions likely due to contrasting public health practices between regions and health information–seeking behavior between age groups. Additionally, predictors relevant to one norovirus season may not contribute to other seasons. Data biases, such as low spatial granularity in Google Trends and especially in Wikipedia data, also play a role in the results. Moreover, internet searches can provide insight into mental models, that is, an individual’s conceptual understanding of norovirus infection and transmission, which could be used in public health communication strategies.
- 8 May 2023
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 25 (1) , e37540
Open AccessPredicting Norovirus in England Using Existing and Emerging Syndromic Data: Infodemiology Study
Background: Norovirus is associated with approximately 18% of the global burden of gastroenteritis and affects all age groups. There is currently no licensed vaccine or available antiviral treatment. However, well-designed early warning systems and forecasting can guide nonpharmaceutical approaches to norovirus infection prevention and control. Objective: This study evaluates the predictive power of existing syndromic surveillance data and emerging data sources, such as internet searches and Wikipedia page views, to predict norovirus activity across a range of age groups across England. Methods: We used existing syndromic surveillance and emerging syndromic data to predict laboratory data indicating norovirus activity. Two methods are used to evaluate the predictive potential of syndromic variables. First, the Granger causality framework was used to assess whether individual variables precede changes in norovirus laboratory reports in a given region or an age group. Then, we used random forest modeling to estimate the importance of each variable in the context of others with two methods: (1) change in the mean square error and (2) node purity. Finally, these results were combined into a visualization indicating the most influential predictors for norovirus laboratory reports in a specific age group and region. Results: Our results suggest that syndromic surveillance data include valuable predictors for norovirus laboratory reports in England. However, Wikipedia page views are less likely to provide prediction improvements on top of Google Trends and Existing Syndromic Data. Predictors displayed varying relevance across age groups and regions. For example, the random forest modeling based on selected existing and emerging syndromic variables explained 60% variance in the ≥65 years age group, 42% in the East of England, but only 13% in the South West region. Emerging data sets highlighted relative search volumes, including “flu symptoms,” “norovirus in pregnancy,” and norovirus activity in specific years, such as “norovirus 2016.” Symptoms of vomiting and gastroenteritis in multiple age groups were identified as important predictors within existing data sources. Conclusions: Existing and emerging data sources can help predict norovirus activity in England in some age groups and geographic regions, particularly, predictors concerning vomiting, gastroenteritis, and norovirus in the vulnerable populations and historical terms such as stomach flu. However, syndromic predictors were less relevant in some age groups and regions likely due to contrasting public health practices between regions and health information–seeking behavior between age groups. Additionally, predictors relevant to one norovirus season may not contribute to other seasons. Data biases, such as low spatial granularity in Google Trends and especially in Wikipedia data, also play a role in the results. Moreover, internet searches can provide insight into mental models, that is, an individual’s conceptual understanding of norovirus infection and transmission, which could be used in public health communication strategies.
- Open Access
- 22 May 2022
- Published by JMIR Publications Inc. in JMIR Preprints
Predicting Norovirus in England Using Existing and Emerging Syndromic Data: Infodemiology Study (Preprint)
BACKGROUND Norovirus is associated with approximately 18% of the global burden of gastroenteritis and affects all age groups. There is currently no licensed vaccine or available antiviral treatment. However, well-designed early warning systems and forecasting can guide nonpharmaceutical approaches to norovirus infection prevention and control. OBJECTIVE This study evaluates the predictive power of existing syndromic surveillance data and emerging data sources, such as internet searches and Wikipedia page views, to predict norovirus activity across a range of age groups across England. METHODS We used existing syndromic surveillance and emerging syndromic data to predict laboratory data indicating norovirus activity. Two methods are used to evaluate the predictive potential of syndromic variables. First, the Granger causality framework was used to assess whether individual variables precede changes in norovirus laboratory reports in a given region or an age group. Then, we used random forest modeling to estimate the importance of each variable in the context of others with two methods: (1) change in the mean square error and (2) node purity. Finally, these results were combined into a visualization indicating the most influential predictors for norovirus laboratory reports in a specific age group and region. RESULTS Our results suggest that syndromic surveillance data include valuable predictors for norovirus laboratory reports in England. However, Wikipedia page views are less likely to provide prediction improvements on top of Google Trends and Existing Syndromic Data. Predictors displayed varying relevance across age groups and regions. For example, the random forest modeling based on selected existing and emerging syndromic variables explained 60% variance in the ≥65 years age group, 42% in the East of England, but only 13% in the South West region. Emerging data sets highlighted relative search volumes, including “flu symptoms,” “norovirus in pregnancy,” and norovirus activity in specific years, such as “norovirus 2016.” Symptoms of vomiting and gastroenteritis in multiple age groups were identified as important predictors within existing data sources. CONCLUSIONS Existing and emerging data sources can help predict norovirus activity in England in some age groups and geographic regions, particularly, predictors concerning vomiting, gastroenteritis, and norovirus in the vulnerable populations and historical terms such as stomach flu. However, syndromic predictors were less relevant in some age groups and regions likely due to contrasting public health practices between regions and health information–seeking behavior between age groups. Additionally, predictors relevant to one norovirus season may not contribute to other seasons. Data biases, such as low spatial granularity in Google Trends and especially in Wikipedia data, also play a role in the results. Moreover, internet searches can provide insight into mental models, that is, an individual’s conceptual understanding of norovirus infection and transmission, which could be used in public health communication strategies.
- 22 May 2022
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessPredicting Norovirus in England Using Existing and Emerging Syndromic Data: Infodemiology Study (Preprint)
BACKGROUND Norovirus is associated with approximately 18% of the global burden of gastroenteritis and affects all age groups. There is currently no licensed vaccine or available antiviral treatment. However, well-designed early warning systems and forecasting can guide nonpharmaceutical approaches to norovirus infection prevention and control. OBJECTIVE This study evaluates the predictive power of existing syndromic surveillance data and emerging data sources, such as internet searches and Wikipedia page views, to predict norovirus activity across a range of age groups across England. METHODS We used existing syndromic surveillance and emerging syndromic data to predict laboratory data indicating norovirus activity. Two methods are used to evaluate the predictive potential of syndromic variables. First, the Granger causality framework was used to assess whether individual variables precede changes in norovirus laboratory reports in a given region or an age group. Then, we used random forest modeling to estimate the importance of each variable in the context of others with two methods: (1) change in the mean square error and (2) node purity. Finally, these results were combined into a visualization indicating the most influential predictors for norovirus laboratory reports in a specific age group and region. RESULTS Our results suggest that syndromic surveillance data include valuable predictors for norovirus laboratory reports in England. However, Wikipedia page views are less likely to provide prediction improvements on top of Google Trends and Existing Syndromic Data. Predictors displayed varying relevance across age groups and regions. For example, the random forest modeling based on selected existing and emerging syndromic variables explained 60% variance in the ≥65 years age group, 42% in the East of England, but only 13% in the South West region. Emerging data sets highlighted relative search volumes, including “flu symptoms,” “norovirus in pregnancy,” and norovirus activity in specific years, such as “norovirus 2016.” Symptoms of vomiting and gastroenteritis in multiple age groups were identified as important predictors within existing data sources. CONCLUSIONS Existing and emerging data sources can help predict norovirus activity in England in some age groups and geographic regions, particularly, predictors concerning vomiting, gastroenteritis, and norovirus in the vulnerable populations and historical terms such as stomach flu. However, syndromic predictors were less relevant in some age groups and regions likely due to contrasting public health practices between regions and health information–seeking behavior between age groups. Additionally, predictors relevant to one norovirus season may not contribute to other seasons. Data biases, such as low spatial granularity in Google Trends and especially in Wikipedia data, also play a role in the results. Moreover, internet searches can provide insight into mental models, that is, an individual’s conceptual understanding of norovirus infection and transmission, which could be used in public health communication strategies.