Syndromic surveillance for influenza in two hospital emergency departments. Relationships between ICD-10 codes and notified cases, before and during a pandemic

Evaluation of an influenza-like illness sentinel surveillance system in South Korea, 2017-2023

BACKGROUND

Guided by the data from the surveillance system, public health efforts have contributed to reducing the burden of influenza in many countries. During the COVID-19 pandemic, many surveillance resources were directed at tracking the severe acute respiratory syndrome-Coronavirus 2. However, most countries have not reported surveillance evaluations during the COVID-19 pandemic.

METHODS

Using the U.S. CDC surveillance evaluation method, we evaluated the influenza-like illness (ILI) sentinel surveillance performance in South Korea between January 2017 and September 2023. For the timeliness, we measured the mean time lag between the reports from the sentinel sites to the Korea Disease Control and Prevention Agency (KDCA) and surveillance result dissemination from KDCA. For the completeness, we measured the submission rate of complete reports per overall number of reports from each sentinel site to the KDCA. For the sensitivity, we calculated the correlation coefficient between the monthly number of ILI reports and the patients with ILI from the Korea national reimbursement data by either Pearson's or Spearman's test. For the representativeness, we compared the age-specific distribution of ILI between the surveillance data and the national reimbursement data using a chi-squared test.

RESULTS

We found that the surveillance performance of timeliness (less than 2 weeks) and completeness (97 %-98 %) was stable during the study period. However, we found a reduced surveillance sensitivity (correlation coefficient: 0.73 in 2020, and 0.84 in 2021) compared to that of 2017-2019 (0.96-0.99), and it recovered in 2022-2023 (0.93-0.97). We found no statistical difference across the proportion of age groups between the surveillance and reimbursement data during the study period (all P-values > 0.05).

CONCLUSIONS

Ongoing surveillance performance monitoring is necessary to maintain efficient policy decision-making for the control of the influenza epidemic. Additional research is needed to assess the overall influenza surveillance system including laboratory and hospital-based surveillance in the country.

Near Real-Time Syndromic Surveillance of Emergency Department Triage Texts Using Natural Language Processing: Case Study in Febrile Convulsion Detection

BACKGROUND

Collecting information on adverse events following immunization from as many sources as possible is critical for promptly identifying potential safety concerns and taking appropriate actions. Febrile convulsions are recognized as an important potential reaction to vaccination in children aged <6 years.

OBJECTIVE

The primary aim of this study was to evaluate the performance of natural language processing techniques and machine learning (ML) models for the rapid detection of febrile convulsion presentations in emergency departments (EDs), especially with respect to the minimum training data requirements to obtain optimum model performance. In addition, we examined the deployment requirements for a ML model to perform real-time monitoring of ED triage notes.

METHODS

We developed a pattern matching approach as a baseline and evaluated ML models for the classification of febrile convulsions in ED triage notes to determine both their training requirements and their effectiveness in detecting febrile convulsions. We measured their performance during training and then compared the deployed models' result on new incoming ED data.

RESULTS

Although the best standard neural networks had acceptable performance and were low-resource models, transformer-based models outperformed them substantially, justifying their ongoing deployment.

CONCLUSIONS

Using natural language processing, particularly with the use of large language models, offers significant advantages in syndromic surveillance. Large language models make highly effective classifiers, and their text generation capacity can be used to enhance the quality and diversity of training data.

Real-time surveillance of severe acute respiratory infections in Scottish hospitals: an electronic register-based approach, 2017-2022

OBJECTIVES

The COVID-19 pandemic highlighted the importance of routine syndromic surveillance of respiratory infections, specifically new cases of severe acute respiratory infection (SARI). This surveillance often relies on questionnaires carried out by research nurses or transcriptions of doctor's notes, but existing, routinely collected electronic healthcare data sets are increasingly being used for such surveillance. We investigated how patient diagnosis codes, recorded within such data sets, could be used to capture SARI trends in Scotland.

STUDY DESIGN

We conducted a retrospective observational study using electronic healthcare data sets between 2017 and 2022.

METHODS

Sensitive, specific and timely case definition (CDs) based on patient diagnosis codes contained within national registers in Scotland were proposed to identify SARI cases. Representativeness and sensitivity analyses were performed to assess how well SARI cases captured by each definition matched trends in historic influenza and SARS-CoV-2 data.

RESULTS

All CDs accurately captured the peaks seen in laboratory-confirmed positive influenza and SARS-CoV-2 data, although the completeness of patient diagnosis records was discovered to vary widely. The timely CD provided the earliest detection of changes in SARI activity, whilst the sensitive CD provided insight into the burden and severity of SARI infections.

CONCLUSIONS

A universal SARI surveillance system has been developed and demonstrated to accurately capture seasonal SARI trends. It can be used as an indicator of emerging secondary care burden of emerging SARI outbreaks. The system further strengthens Scotland's existing strategies for respiratory surveillance, and the methods described here can be applied within any country with suitable electronic patient records.

The Impact of the COVID-19 Pandemic on Respiratory Illness Admissions at a Single Academic Institution in Arkansas

BACKGROUND

The first reported COVID-19 case in Arkansas was on 11 March 2020, two months after the first reported case in the United States. We sought to analyze rates of respiratory illness and influenza tests during the 2019/2020 influenza season compared to pre-pandemic years to assess whether there were higher rates of respiratory illness than expected, which may suggest undiagnosed COVID-19 cases.

METHODS

Using data collected from the data warehouse of the largest hospital in Arkansas, ICD-9 and ICD-10 codes related to respiratory illness were identified for 1 October to 1 May 2017-2020.

RESULTS

We identified 25,747 patients admitted with respiratory illness during the study. We found no significant difference in the rate of monthly admissions with respiratory illness between seasons (p = 0.14). We saw a significant increase in the number of influenza tests ordered in 2019/2020 (p < 0.01).

CONCLUSIONS

The rate of hospitalizations with respiratory illness did not significantly increase during the 2019/2020 season; however, influenza testing increased without a statistically significant difference in positivity rate. The increase in ordered influenza tests indicates an increased clinical suspicion, which may suggest a rise in pre-hospital viral illness associated with COVID-19.

Cost-Effectiveness of Extending the National Influenza Vaccination Program in South Korea: Does Vaccination of Older Adults Provide Health Benefits to the Entire Population?

The South Korean government has successfully improved influenza vaccination coverage for individuals aged 65 years or older as part of its National Immunization Program (NIP). Those aged 50–64 years without funded vaccination care have significantly lower vaccination rates and face a substantial risk of influenza-related complications. We use a dynamic epidemiological and economic model to investigate the cost-effectiveness of expanding the universal vaccine fund to include those aged 50–64. The epidemiological model is estimated using the susceptibility-infection-recovery model and influenza and influenza-like illness incidence rates, which were calculated by the National Health Insurance Service–National Sample Cohort from the 2008/09 to 2012/13 influenza seasons but excluding the 2009/10 season for pandemic influenza A (H1N1). The decision tree economic model is assessed from societal and healthcare sector perspectives. The proposed policy would eliminate 340,000 annual influenza cases and prevent 119 unnecessary deaths. From a societal perspective, the proposed policy would reduce costs by USD 68 million. From a healthcare perspective, the cost is USD 4318 per quality-adjusted life years. Within the study range, sensitivity analyses found consistent cost-effectiveness results. The influenza vaccine for adults aged 50–64 appears to be cost-saving or cost-effective and, thus, should be considered for the NIP.

Influenza Virus: Tracking, Predicting, and Forecasting

Influenza is a common respiratory infection that causes considerable morbidity and mortality worldwide each year. In recent years, along with the improvement in computational resources, there have been a number of important developments in the science of influenza surveillance and forecasting. Influenza surveillance systems have been improved by synthesizing multiple sources of information. Influenza forecasting has developed into an active field, with annual challenges in the United States that have stimulated improved methodologies. Work continues on the optimal approaches to assimilating surveillance data and information on relevant driving factors to improve estimates of the current situation (nowcasting) and to forecast future dynamics.

Barriers to hospital electronic public health reporting and implications for the COVID-19 pandemic

We sought to identify barriers to hospital reporting of electronic surveillance data to local, state, and federal public health agencies and the impact on areas projected to be overwhelmed by the COVID-19 pandemic. Using 2018 American Hospital Association data, we identified barriers to surveillance data reporting and combined this with data on the projected impact of the COVID-19 pandemic on hospital capacity at the hospital referral region level.

Our results find the most common barrier was public health agencies lacked the capacity to electronically receive data, with 41.2% of all hospitals reporting it. We also identified 31 hospital referral regions in the top quartile of projected bed capacity needed for COVID-19 patients in which over half of hospitals in the area reported that the relevant public health agency was unable to receive electronic data.

Public health agencies’ inability to receive electronic data is the most prominent hospital-reported barrier to effective syndromic surveillance. This reflects the policy commitment of investing in information technology for hospitals without a concomitant investment in IT infrastructure for state and local public health agencies.

Anatomy of a demand shock: Quantitative analysis of crowding in hospital emergency departments in Victoria, Australia during the 2009 influenza pandemic

Objective

An infectious disease outbreak such as the 2009 influenza pandemic is an unexpected demand shock to hospital emergency departments (EDs). We analysed changes in key performance metrics in (EDs) in Victoria during this pandemic to assess the impact of this demand shock.

Design and setting

Descriptive time-series analysis and longitudinal regression analysis of data from the Victorian Emergency Minimum Dataset (VEMD) using data from the 38 EDs that submit data to the state’s Department of Health and Human Services.

Main outcome measures

Daily number of presentations, influenza-like-illness (ILI) presentations, daily mean waiting time (time to first being seen by a doctor), daily number of patients who did-not-wait and daily number of access-blocked patients (admitted patients with length of stay >8 hours) at a system and hospital-level.

Results

During the influenza pandemic, mean waiting time increased by up to 25%, access block increased by 32% and did not wait presentations increased by 69% above pre-pandemic levels. The peaks of all three crowding variables corresponded approximately to the peak in admitted ILI presentations. Longitudinal fixed-effects regression analysis estimated positive and statistically significant associations between mean waiting times, did not wait presentations and access block and ILI presentations.

Conclusions

This pandemic event caused excess demand leading to increased waiting times, did-not-wait patients and access block. Increases in admitted patients were more strongly associated with crowding than non-admitted patients during the pandemic period, so policies to divert or mitigate low-complexity non-admitted patients are unlikely to be effective in reducing ED crowding.

Inpatient hospital fatality related to coding (ICD-9-CM) of the influenza diagnosis in Spain (2009-2015)

Background

To analyze hospitalization episodes with an ICD-9 diagnosis code of influenza (codes 487 and 488) in any diagnostic position from 2009 to 2015 in the Spanish hospital surveillance system.

Methods

Information about age, length of stay in hospital, mortality, comorbidity with an influenza diagnosis code between 1 October 2009 and 30 September 2015 was obtained from the National Surveillance System for Hospital Data (Conjunto Mínimo Básico de Datos, CMBD).

Results

52,884 hospital admissions were obtained. A total of 24,527 admissions corresponded to diagnoses ICD-9 code 487 (46.4%), and 28,357 (53.6%) corresponded to ICD-9 code 488. The global hospitalization rates were 8.7 and 10.6 per 100,000 people, respectively. Differences between the two diagnostic groups were found for each of the six analyzed seasons. The diagnostic ICD-9-CM 488, male gender, and high-risk patients classified by risk vaccination groups showed direct relationship with inpatient hospital death.

Conclusions

Influenza diagnosis was present in a significant number of hospital admissions. The code used for diagnosis (ICD-9-CM 488), male sex, age groups and associated risk clinical conditions showed a direct relationship with inpatient hospital fatality.

[Infectious diseases and their ICD coding : What could be improved by the introduction of ICD-11?]

Die Revision der Internationalen statistischen Klassifikation der Krankheiten und verwandter Gesundheitsprobleme (International Classification of Diseases – ICD) geht mit grundlegenden Änderungen der Morbiditäts- und Mortalitätsstatistik einher, die auch den Bereich der Infektionskrankheiten betreffen. Die Zuordnung der einzelnen Infektionskrankheiten zu den Kapiteln in der aktuellen ICD-10 erfolgt aufgrund unterschiedlicher Konzepte, teilweise nach auslösendem Agens, nach betroffenem Organsystem oder nach Lebensperiode. Besondere Herausforderungen der Klassifizierung der Infektionskrankheiten bestehen u. a. darin, dass regelmäßig ein Anpassungsbedarf durch neu auftretende Erreger entstehen kann. Außerdem reichen die Angaben hinsichtlich Umfang und Tiefe in der ICD-10 teilweise nicht aus, um epidemiologische Auswertungen der Daten durchzuführen.

Die ICD ermöglicht den weltweiten Vergleich von Statistiken zu Infektionskrankheiten. Zunehmend wird die ICD jedoch auch für die Erhebung von Surveillance- und Forschungsdaten eingesetzt, z. B. im Rahmen des Meldewesens (Identifizierung von Meldetatbeständen), aber auch in der syndromischen Surveillance akuter Atemwegsinfektionen und für den Aufbau neuer Surveillance-Systeme sowie der Evaluation der Datenqualität durch Abgleich mit Sekundärdaten.

Die Chancen der ICD-11 liegen vor allem darin, dass Infektionskrankheiten eindeutiger codiert werden können und ihre Codierung mehr relevante Informationen für die epidemiologische Bewertung enthält. Durch die hohe Komplexität können jedoch Verzerrungen in den Daten entstehen, die die Fortschreibung der Morbiditäts- und Mortalitätsstatistiken erschweren.

Right sizing for vaccine effectiveness studies: how many is enough for reliable estimation?

Background The precision of vaccine effectiveness (VE) estimates is dependent on sample size and sampling methods. In Victoria, participating general practitioners (GPs) are not limited by the number of influenza-like illness (ILI) patients they collect respiratory samples (swabs) from in sentinel surveillance. However, in the context of scarce resources it is of interest to determine the minimum sample size needed for reliable estimates. Methods Following the test-negative design, patients with ILI were recruited by GPs and tested for influenza. Descriptive analyses were conducted to assess possible selection bias introduced by GPs. VE was calculated by logistic regression as [1 – odds ratio] x 100% and adjusted for week of presentation and age. Random 20% and 50% samples were selected without replacement to estimate the effect of swab rates on VE estimates. Results GPs swabbed a smaller proportion of patients aged ≥65 years (45.9%, n=238) than those <5 (75.6%, n=288), 5–17 (67.9%, n=547) and 18–64 (75.6%, n=2662) years. Decreasing the swab rate did not alter VE point estimates significantly. However, it reduced the precision of estimates and in some instances resulted in too small a sample size to estimate VE. Conclusion Imposing a 20% or 50% swabbing rate produces less robust VE estimates. The number of swabs required per year to produce precise estimates should be dictated by seasonal severity, rather than an arbitrary rate. It would be beneficial for GPs to swab patients systematically by age group to ensure there are sufficient data to investigate VE against a particular subtype in a given age group.

Influenza-associated hospitalisation, vaccine uptake and socioeconomic deprivation in an English city region: an ecological study

OBJECTIVES

Every year, influenza poses a significant burden on the National Health Service in England. Influenza vaccination is an effective measure to prevent severe disease, hence, maximising vaccine coverage in the most vulnerable is a priority. We aimed to identify the extent to which socioeconomic status is associated with influenza-associated illness (IAI) and influenza vaccine coverage.

DESIGN

Retrospective observational study using hospital episode statistics.

SETTING

Merseyside, North-West of England, including the city of Liverpool.

PARTICIPANTS

Residents of Merseyside hospitalised with IAI between April 2004 and March 2016, and Merseyside general practice registered patients eligible for influenza vaccination in 2014/2015 and 2015/2016 influenza seasons.

EXPOSURES

Socioeconomic deprivation based on lower super output area English Indices of Deprivation scores.

PRIMARY AND SECONDARY OUTCOME MEASURES

Incidence and risk of IAI hospitalisation, and vaccine uptake.

RESULTS

There were 89 058 hospitalisations related to IAI among Merseyside residents (mean yearly rate=4.9 per 1000 population). Hospitalisations for IAI were more frequent in the most socioeconomically deprived areas compared with the least deprived in adults aged 15-39 years (incidence rate ratio (IRR) 2.08;95% CI 1.76 to 2.45; p<0.001), 60-64 years (IRR 2.65; 95% CI 2.35 to 2.99; p<0.001) and 65+ years (IRR 1.90; 95% CI 1.73 to 2.10; p<0.001), whereas rates in children were more homogeneous across deprivation strata. Vaccine uptake was lower than the nationally set targets in most neighbourhoods. The odds of vaccine uptake were 30% lower (OR 0.70; 95% CI 0.66 to 0.74; p<0.001) and 10% lower (OR 0.90; 95% CI 0.88 to 0.92; p<0.001) in the most socioeconomically deprived quintile compared with the least deprived, among children aged 24-59 months and 65+ years, respectively.

CONCLUSIONS

Higher rates of IAI hospitalisations and lower vaccine uptake in the most socioeconomically deprived populations suggest that health promotion policies and interventions that target these populations should be a priority.

Are we missing respiratory viral infections in infants and children? Comparison of a hospital-based quality management system with standard of care

OBJECTIVES

Hospital-based surveillance of influenza and acute respiratory infections relies on International Classification of Diseases (ICD) codes and hospital laboratory reports (Standard-of-Care). It is unclear how many cases are missed with either method, i.e. remain undiagnosed/coded as influenza and other respiratory virus infections. Various influenza-like illness (ILI) definitions co-exist with little guidance on how to use them. We compared the diagnostic accuracy of standard surveillance methods with a prospective quality management (QM) programme at a Berlin children's hospital with the Robert Koch Institute.

METHODS

Independent from routine care, all patients fulfilling pre-defined ILI-criteria (QM-ILI) participated in the QM programme. A separate QM team conducted standardized clinical assessments and collected nasopharyngeal specimens for blinded real-time quantitative PCR for influenza A/B viruses, respiratory syncytial virus, adenovirus, rhinovirus and human metapneumovirus.

RESULTS

Among 6073 individuals with ILI qualifying for the QM programme, only 8.7% (528/6073) would have undergone virus diagnostics during Standard-of-Care. Surveillance based on ICD codes would have missed 61% (359/587) of influenza diagnoses. Of baseline ICD codes, 53.2% (2811/5282) were non-specific, most commonly J06 ('acute upper respiratory infection'). Comparison of stakeholder case definitions revealed that QM-ILI and the WHO ILI case definition showed the highest overall sensitivities (84%-97% and 45%-68%, respectively) and the CDC ILI definition had the highest sensitivity for influenza infections (36%, 95% CI 31.4-40.8 for influenza A and 48%, 95% CI 40.5-54.7 for influenza B).

CONCLUSIONS

Disease-burden estimates and surveillance should account for the underreporting of cases in routine care. Future studies should explore the effect of ILI screening and surveillance in various age groups and settings. Diagnostic algorithms should be based on the WHO ILI case definition combined with targeted testing.

[Evaluation of an ICD-10-based electronic surveillance of acute respiratory infections (SEEDARI) in Germany]

BACKGROUND

Every year epidemic waves of influenza and other acute respiratory infections (ARIs) cause a highly variable burden of disease in the population. Thus, assessment of the situation and adaptation of prevention strategies have to rely on real time syndromic surveillance.

OBJECTIVE

We have established an ICD-10-based electronic system allowing rapid capture and transmission of information on ARI (SEED^ARI), in Germany. Here we report the evaluation of this new system based on results of the syndromic and virologic surveillance carried out by the working group on influenza in Germany (AGI).

METHODS

Consultations and ICD10-codes (J00-J22, J44.0 and B34.9) between week 16 in 2009, and week 15 in 2013, were used for comparison with AGI data. The time course and the correlation of weekly estimates of the incidence of medically attended ARI (MAARI) and ARI/100 consultations were analyzed for the different surveillance systems.

RESULTS

The number of participating medical practices in SEED^ARI almost doubled from 2009 (n = 65) to 2013 (n = 111). A total of almost 6.8 million consultations and 465,006 diagnosed ARIs were transmitted. The comparison of weekly estimated incidence of MAARI per 100,000 capita derived from SEED^ARI and the results of the AGI showed high statistical correlation (Spearman correlation coefficient r_s = 0,924; n = 209; p < 0,001). The proportion of diagnosed influenza (J09-J11) and the weekly positivity rate from virological surveillance during epidemic waves also showed high correlations.

DISCUSSION

We conclude that SEED^ARI represents a valid system for syndromic influenza surveillance. The case-based ICD-10 approach allows a detailed analysis of the actual situation and also seems suitable for population-based studies.

Detection of influenza-like illness aberrations by directly monitoring Pearson residuals of fitted negative binomial regression models

Background

Emerging novel influenza outbreaks have increasingly been a threat to the public and a major concern of public health departments. Real-time data in seamless surveillance systems such as health insurance claims data for influenza-like illnesses (ILI) are ready for analysis, making it highly desirable to develop practical techniques to analyze such readymade data for outbreak detection so that the public can receive timely influenza epidemic warnings. This study proposes a simple and effective approach to analyze area-based health insurance claims data including outpatient and emergency department (ED) visits for early detection of any aberrations of ILI.

Methods

The health insurance claims data during 2004–2009 from a national health insurance research database were used for developing early detection methods. The proposed approach fitted the daily new ILI visits and monitored the Pearson residuals directly for aberration detection. First, negative binomial regression was used for both outpatient and ED visits to adjust for potentially influential factors such as holidays, weekends, seasons, temporal dependence and temperature. Second, if the Pearson residuals exceeded 1.96, aberration signals were issued. The empirical validation of the model was done in 2008 and 2009. In addition, we designed a simulation study to compare the time of outbreak detection, non-detection probability and false alarm rate between the proposed method and modified CUSUM.

Results

The model successfully detected the aberrations of 2009 pandemic (H1N1) influenza virus in northern, central and southern Taiwan. The proposed approach was more sensitive in identifying aberrations in ED visits than those in outpatient visits. Simulation studies demonstrated that the proposed approach could detect the aberrations earlier, and with lower non-detection probability and mean false alarm rate in detecting aberrations compared to modified CUSUM methods.

Conclusions

The proposed simple approach was able to filter out temporal trends, adjust for temperature, and issue warning signals for the first wave of the influenza epidemic in a timely and accurate manner.

Electronic supplementary material

The online version of this article (doi:10.1186/s12889-015-1500-4) contains supplementary material, which is available to authorized users.

Primary care utilization among patients with influenza during the 2009 pandemic. Does risk for severe influenza disease or prior contact with the general practitioner have any influence?

BACKGROUND

Little is known about how patients belonging to risk groups for influenza used the primary care system during the influenza pandemic.

AIMS

To investigate the use of general practice and out-of-hours (OOH) services in patients with influenza-like illness (ILI) according to (i) risk for severe influenza disease and (ii) the number of regular general practitioner (GP) visits before the pandemic.

METHOD

Observational study of all ILI patients during the 2009 pandemic. Data were recorded prospectively and collected after the pandemic. Patients at risk were identified during an 18-month period by diagnoses from GPs' billing claims. Associations between risk factors for severe influenza disease and utilization of primary care were analysed using bivariate and multivariate regression analyses. Similar analyses were used for the association between number of GP visits before the pandemic and the primary care utilization during the pandemic.

RESULTS

ILI patients who were pregnant [odds ratio (OR) 1.70; 95% confidence interval (CI) 1.52, 1.89], had diabetes (OR 1.68; 95% CI 1.49, 1.89) or chronic lung disease (OR 1.44; 95 CI 1.34, 1.55) had increased risk of attending OOH services compared with patients with no risk factor. ILI patients with at least one GP visit prior to the pandemic used OOH services less during the pandemic compared with those with no GP visit.

CONCLUSION

An increased use of OOH services was found in ILI patients who were pregnant, with diabetes or with chronic lung disease. Having visited the GP before the pandemic was associated with less use of OOH services among ILI patients.

Surveillance for emerging respiratory viruses

Several new viral respiratory tract infectious diseases with epidemic potential that threaten global health security have emerged in the past 15 years. In 2003, WHO issued a worldwide alert for an unknown emerging illness, later named severe acute respiratory syndrome (SARS). The disease caused by a novel coronavirus (SARS-CoV) rapidly spread worldwide, causing more than 8000 cases and 800 deaths in more than 30 countries with a substantial economic impact. Since then, we have witnessed the emergence of several other viral respiratory pathogens including influenza viruses (avian influenza H5N1, H7N9, and H10N8; variant influenza A H3N2 virus), human adenovirus-14, and Middle East respiratory syndrome coronavirus (MERS-CoV). In response, various surveillance systems have been developed to monitor the emergence of respiratory-tract infections. These include systems based on identification of syndromes, web-based systems, systems that gather health data from health facilities (such as emergency departments and family doctors), and systems that rely on self-reporting by patients. More effective national, regional, and international surveillance systems are required to enable rapid identification of emerging respiratory epidemics, diseases with epidemic potential, their specific microbial cause, origin, mode of acquisition, and transmission dynamics.

Can epidemic detection systems at the hospital level complement regional surveillance networks: case study with the influenza epidemic?

Background

Early knowledge of influenza outbreaks in the community allows local hospital healthcare workers to recognise the clinical signs of influenza in hospitalised patients and to apply effective precautions. The objective was to assess intra-hospital surveillance systems to detect earlier than regional surveillance systems influenza outbreaks in the community.

Methods

Time series obtained from computerized medical data from patients who visited a French hospital emergency department (ED) between June 1st, 2007 and March 31st, 2011 for influenza, or were hospitalised for influenza or a respiratory syndrome after an ED visit, were compared to different regional series. Algorithms using CUSUM method were constructed to determine the epidemic detection threshold with the local data series. Sensitivity, specificity and mean timeliness were calculated to assess their performance to detect community outbreaks of influenza. A sensitivity analysis was conducted, excluding the year 2009, due to the particular epidemiological situation related to pandemic influenza this year.

Results

The local series closely followed the seasonal trends reported by regional surveillance. The algorithms achieved a sensitivity of detection equal to 100% with series of patients hospitalised with respiratory syndrome (specificity ranging from 31.9 and 92.9% and mean timeliness from −58.3 to 20.3 days) and series of patients who consulted the ED for flu (specificity ranging from 84.3 to 93.2% and mean timeliness from −32.3 to 9.8 days). The algorithm with the best balance between specificity (87.7%) and mean timeliness (0.5 day) was obtained with series built by analysis of the ICD-10 codes assigned by physicians after ED consultation. Excluding the year 2009, the same series keeps the best performance with specificity equal to 95.7% and mean timeliness equal to −1.7 day.

Conclusions

The implementation of an automatic surveillance system to detect patients with influenza or respiratory syndrome from computerized ED records could allow outbreak alerts at the intra-hospital level before the publication of regional data and could accelerate the implementation of preventive transmission-based precautions in hospital settings.

Emergency department and 'Google flu trends' data as syndromic surveillance indicators for seasonal influenza

We evaluated syndromic indicators of influenza disease activity developed using emergency department (ED) data - total ED visits attributed to influenza-like illness (ILI) ('ED ILI volume') and percentage of visits attributed to ILI ('ED ILI percent') - and Google flu trends (GFT) data (ILI cases/100 000 physician visits). Congruity and correlation among these indicators and between these indicators and weekly count of laboratory-confirmed influenza in Manitoba was assessed graphically using linear regression models. Both ED and GFT data performed well as syndromic indicators of influenza activity, and were highly correlated with each other in real time. The strongest correlations between virological data and ED ILI volume and ED ILI percent, respectively, were 0·77 and 0·71. The strongest correlation of GFT was 0·74. Seasonal influenza activity may be effectively monitored using ED and GFT data.

Syndromic surveillance for influenza in the emergency department-A systematic review

The science of surveillance is rapidly evolving due to changes in public health information and preparedness as national security issues, new information technologies and health reform. As the Emergency Department has become a much more utilized venue for acute care, it has also become a more attractive data source for disease surveillance. In recent years, influenza surveillance from the Emergency Department has increased in scope and breadth and has resulted in innovative and increasingly accepted methods of surveillance for influenza and influenza-like-illness (ILI). We undertook a systematic review of published Emergency Department-based influenza and ILI syndromic surveillance systems. A PubMed search using the keywords "syndromic", "surveillance", "influenza" and "emergency" was performed. Manuscripts were included in the analysis if they described (1) data from an Emergency Department (2) surveillance of influenza or ILI and (3) syndromic or clinical data. Meeting abstracts were excluded. The references of included manuscripts were examined for additional studies. A total of 38 manuscripts met the inclusion criteria, describing 24 discrete syndromic surveillance systems. Emergency Department-based influenza syndromic surveillance has been described worldwide. A wide variety of clinical data was used for surveillance, including chief complaint/presentation, preliminary or discharge diagnosis, free text analysis of the entire medical record, Google flu trends, calls to teletriage and help lines, ambulance dispatch calls, case reports of H1N1 in the media, markers of ED crowding, admission and Left Without Being Seen rates. Syndromes used to capture influenza rates were nearly always related to ILI (i.e. fever +/- a respiratory or constitutional complaint), however, other syndromes used for surveillance included fever alone, "respiratory complaint" and seizure. Two very large surveillance networks, the North American DiSTRIBuTE network and the European Triple S system have collected large-scale Emergency Department-based influenza and ILI syndromic surveillance data. Syndromic surveillance for influenza and ILI from the Emergency Department is becoming more prevalent as a measure of yearly influenza outbreaks.

Local implementation of a syndromic influenza surveillance system using emergency department data in Santander, Spain

BACKGROUND

We assessed the local implementation of syndromic surveillance (SyS) as part of the European project 'System for Information on, Detection and Analysis of Risks and Threats to Health' in Santander, Spain.

METHODS

We applied a cumulative sum algorithm on emergency department (ED) chief complaints for influenza-like illness in the seasons 2010-11 and 2011-12. We fine tuned the algorithm using a receiver operating characteristic analysis to identify the optimal trade-off of sensitivity and specificity and defined alert criteria. We assessed the timeliness of the SyS system to detect the onset of the influenza season.

RESULTS

The ED data correlated with the sentinel data. With the best algorithm settings we achieved 70/63% sensitivity and 89/95% specificity for 2010-11/2011-12. At least 2 consecutive days of signals defined an alert. In 2010-11 the SyS system alerted 1 week before the sentinel system and in 2011-12 in the same week. The data from the ED is available on a daily basis providing an advantage in timeliness compared with the weekly sentinel data.

CONCLUSIONS

ED-based SyS in Santander complements sentinel influenza surveillance by providing timely information. Local fine tuning and definition of alert criteria are recommended to enhance validity.

Impact of seasonal and pandemic influenza on emergency department visits, 2003-2010, Ontario, Canada

Objectives

Weekly influenza-like illness (ILI) consultation rates are an integral part of influenza surveillance. However, in most health care settings, only a small proportion of true influenza cases are clinically diagnosed as influenza or ILI. The primary objective of this study was to estimate the number and rate of visits to the emergency department (ED) that are attributable to seasonal and pandemic influenza and to describe the effect of influenza on the ED by age, diagnostic categories, and visit disposition. A secondary objective was to assess the weekly “real-time” time series of ILI ED visits as an indicator of the full burden due to influenza.

Methods

The authors performed an ecologic analysis of ED records extracted from the National Ambulatory Care Reporting System (NARCS) database for the province of Ontario, Canada, from September 2003 to March 2010 and stratified by diagnostic characteristics (International Classification of Diseases, 10th Revision [ICD-10]), age, and visit disposition. A regression model was used to estimate the seasonal baseline. The weekly number of influenza-attributable ED visits was calculated as the difference between the weekly number of visits predicted by the statistical model and the estimated baseline.

Results

The estimated rate of ED visits attributable to influenza was elevated during the H1N1/2009 pandemic period at 1,000 per 100,000 (95% confidence interval [CI] = 920 to 1,100) population compared to an average annual rate of 500 per 100,000 (95% CI = 450 to 550) for seasonal influenza. ILI or influenza was clinically diagnosed in one of 2.6 (38%) and one of 14 (7%) of these visits, respectively. While the ILI or clinical influenza diagnosis was the diagnosis most specific to influenza, only 87% and 58% of the clinically diagnosed ILI or influenza visits for pandemic and seasonal influenza, respectively, were likely directly due to an influenza infection. Rates for ILI ED visits were highest for younger age groups, while the likelihood of admission to hospital was highest in older persons. During periods of seasonal influenza activity, there was a significant increase in the number of persons who registered with nonrespiratory complaints, but left without being seen. This effect was more pronounced during the 2009 pandemic. The ratio of influenza-attributed respiratory visits to influenza-attributed ILI visits varied from 2.4:1 for the fall H1N1/2009 wave to 9:1 for the 2003/04 influenza A(H3N2) season and 28:1 for the 2007/08 H1N1 season.

Conclusions

Influenza appears to have had a much larger effect on ED visits than was captured by clinical diagnoses of influenza or ILI. Throughout the study period, ILI ED visits were strongly associated with excess respiratory complaints. However, the relationship between ILI ED visits and the estimated effect of influenza on ED visits was not consistent enough from year to year to predict the effect of influenza on the ED or downstream in-hospital resource requirements.