Ian Gemmill

The use of syndromic surveillance for decision-making during the H1N1 pandemic: A qualitative study

BackgroundAlthough an increasing number of studies are documenting uses of syndromic surveillance by front line public health, few detail the value added from linking syndromic data to public health decision-making. This study seeks to understand how syndromic data informed specific public health actions during the 2009 H1N1 pandemic.MethodsSemi-structured telephone interviews were conducted with participants from Ontario’s public health departments, the provincial ministry of health and federal public health agency to gather information about syndromic surveillance systems used and the role of syndromic data in informing specific public health actions taken during the pandemic. Responses were compared with how the same decisions were made by non-syndromic surveillance users.ResultsFindings from 56 interviews (82% response) show that syndromic data were most used for monitoring virus activity, measuring impact on the health care system and informing the opening of influenza assessment centres in several jurisdictions, and supporting communications and messaging, rather than its intended purpose of early outbreak detection. Syndromic data had limited impact on decisions that involved the operation of immunization clinics, school closures, sending information letters home with school children or providing recommendations to health care providers. Both syndromic surveillance users and non-users reported that guidance from the provincial ministry of health, communications with stakeholders and vaccine availability were driving factors in these public health decisions.ConclusionsSyndromic surveillance had limited use in decision-making during the 2009 H1N1 pandemic in Ontario. This study provides insights into the reasons why this occurred. Despite this, syndromic data were valued for providing situational awareness and confidence to support public communications and recommendations. Developing an understanding of how syndromic data are utilized during public health events provides valuable evidence to support future investments in public health surveillance.

Assessing the Relative Timeliness of Ontario's Syndromic Surveillance Systems for Early Detection of the 2009 Influenza H1N1 Pandemic Waves

OBJECTIVES: Building on previous research noting variations in the operation and perceived utility of syndromic surveillance systems in Ontario, the timeliness of these different syndromic systems for detecting the onset of both 2009 H1N1 pandemic (A(H1N1)pdm09) waves relative to laboratory testing data was assessed using a standardized analytic algorithm.METHODS: Syndromic data, specifically local emergency department (ED) visit and school absenteeism data, as well as provincial Telehealth (telephone helpline) and antiviral prescription data, were analyzed retrospectively for the period April 1, 2009 to January 31, 2010. The C2-MEDIUM aberration detection method from the US Centers for Disease Control and Prevention’s EARS software was used to detect increases above expected in syndromic data, and compared to laboratory alerts, defined as notice of confirmed A(H1N1)pdm09 cases over two consecutive days, to assess relative timeliness.RESULTS: In Wave 1, provincial-level alerts were detected for antiviral prescriptions and Telehealth respiratory calls before the laboratory alert. In Wave 2, Telehealth respiratory calls similarly alerted in advance of the laboratory, while local alerts from ED visit, antiviral prescription and school absenteeism data varied in timing relative to the laboratory alerts. Alerts from syndromic data were also observed to coincide with external factors such as media releases.CONCLUSIONS: Alerts from syndromic surveillance systems may be influenced by external factors and variation in system operations. Further understanding of both the impact of external factors on surveillance data and standardizing protocols for defining alerts is needed before the use of syndromic surveillance systems can be optimized.RésuméOBJECTIFS: À partir des résultats de recherches antérieures sur les écarts dans le fonctionnement et l’utilité perçue des systèmes de surveillance syndromique en Ontario, nous avons évalué, à l’aide d’un algorithme analytique standardisé, la rapidité avec laquelle les différents systèmes syndromiques ont détecté l’apparition des deux vagues de la pandémie de grippe H1N1 de 2009 (A(H1N1)pdm09) par rapport aux données des épreuves de laboratoire.MÉTHODE: Les données syndromiques, plus précisément les visites aux services d’urgence locaux et l’absentéisme dans les écoles, ainsi que les appels à Télésanté (une ligne d’assistance téléphonique provinciale) et les ordonnances d’antiviraux, ont été analysées rétrospectivement pour la période du 1er avril 2009 au 31 janvier 2010. Nous avons utilisé la méthode de détection des aberrations C2-MEDIUM du logiciel EARS des Centers for Disease Control and Prevention des États-Unis pour déceler les hausses supérieures aux prévisions dans les données syndromiques, et nous les avons comparées aux alertes des laboratoires, définies comme étant les avis de cas de grippe A(H1N1)pdm09 confirmés au cours de deux journées consécutives, pour évaluer la rapidité relative des systèmes de surveillance syndromique.RÉSULTATS: Durant la 1e vague, des alertes de niveau provincial ont été détectées, dans les ordonnances d’antiviraux et les appels pour problèmes respiratoires à Télésanté, avant les alertes des laboratoires. Durant la 2e vague, les appels pour problèmes respiratoires à Télésanté ont aussi précédé les alertes des laboratoires, mais les alertes locales liées aux visites aux urgences, aux ordonnances d’antiviraux et aux taux d’absentéisme dans les écoles ont varié dans le temps par rapport aux alertes des laboratoires. Il a aussi été observé que les alertes déclenchées par les données syndromiques coïncidaient avec des facteurs externes, comme les communiqués.CONCLUSIONS: Les alertes des systèmes de surveillance syndromique peuvent être influencées par des facteurs externes et des variations dans le fonctionnement des systèmes. Il faudrait pousser la recherche sur deux plans: l’impact exercé par les facteurs externes sur les données de surveillance et la normalisation des protocoles de déclenchement des alertes, avant de pouvoir optimiser l’utilisation des systèmes de surveillance syndromique.

Acute otitis media and its prevention by immunization: a survey of Canadian pediatricians' knowledge, attitudes and beliefs.

Introduction: Acute otitis media (AOM) is one of the most common bacterial infectious diseases among children and is a leading cause of child healthcare visits and antibiotic prescriptions. Few vaccines have the potential to prevent AOM. The newer pneumococcal conjugate vaccines (PCV) offer a larger spectrum of protection against AOM, as well as preventing severe diseases. The main aim of this study was to assess pediatricians’ opinions regarding AOM and its prevention by immunization. Results: Response rate was 50%. Around 60% of respondents estimated that more than 50% of their patients under the age of 3 years would suffer from at least one episode of AOM in the following year. Most respondents (79%) rated consequences of AOM as moderate. Almost all physicians (99%) considered the newer PCV as safe and effective. Most respondents considered their knowledge of the new vaccines was sufficient. More than 90% had a firm intention to recommend newer PCV to their patients. Perceived benefits of AOM prevention by immunization were: reduction of antibiotic administration and reduction of post-AOM complications. More than half of respondents (53%) considered the risk of adverse events as a barrier to AOM prevention by immunization. In multivariate analysis, the main determinant of pediatricians’ intention to recommend newer PCV was perceived safety and efficacy of the vaccines (partial R2 = 0.40, p < 0.0001). Discussion: Results of this survey show that AOM is perceived as an important health problem by paediatricians. Information about the increased protection against AOM offered by newer PCV should be disseminated to physicians. Materials and Methods: A self-administered, anonymous, mail-based questionnaire based upon the Health Belief Model and the Analytical framework for immunization programs was sent to all 1,852 Canadian pediatricians.