Getting a grippe on severity: a retrospective comparison of influenza-related hospitalizations and deaths captured in reportable disease and administrative data sources in Ontario, Canada

J Leigh Hobbs,Michelle Murti,Michael Whelan,Karin Hohenadel,Anne-Luise Winter

doi:10.1186/s12889-019-6924-9

J Leigh Hobbs, Michelle Murti + Show 3 more

Open Access

https://doi.org/10.1186/s12889-019-6924-9

Copy DOI

Abstract

BackgroundSince 2009, in Ontario, reportable disease surveillance data has been used for timely in-season estimates of influenza severity (i.e., hospitalizations and deaths). Due to changes in reporting requirements influenza reporting no longer captures these indicators of severity, necessitating exploration of other potential sources of data. The purpose of this study was to complete a retrospective analysis to assess the comparability of influenza-related hospitalizations and deaths captured in the Ontario reportable disease information system to those captured in Ontario’s hospital-based discharge database.MethodsHospitalizations and deaths of laboratory-confirmed influenza cases reported during the 2010–11 to 2013–14 influenza seasons were analyzed. Information on hospitalizations and deaths for laboratory-confirmed influenza cases were obtained from two databases; the integrated Public Health Information System, which is the provincial reportable disease database, and the Discharge Database, which contains information on all in-patient hospital visits using the International Classification of Diseases, 10th Revision, Canada (ICD-10-CA) coding standards. Analyses were completed using the ICD-10 J09 and J10 diagnosis codes as an indicator for laboratory-confirmed influenza, and a secondary analysis included the physician-diagnosed influenza J11 diagnosis code.ResultsFor each season, reported hospitalizations for laboratory-confirmed influenza cases in the reportable disease data were higher compared to hospitalizations with J09 and J10 diagnoses codes, but lower when J11 codes were included. The number of deaths was higher in the reportable disease data, whether or not J11 codes were included. For all four seasons, the weekly trends in the number of hospitalizations and deaths were similar for the reportable disease and hospital data (with and without J11), with seasonal peaks occurring during the same week or within 1 week of each other.ConclusionIn our retrospective analyses we found that hospital data provided a reliable estimate of the trends of influenza-related hospitalizations and deaths compared to the reportable disease data for the 2010–11 to 2013–14 influenza seasons in Ontario, but may under-estimate the total seasonal number of deaths. Hospital data could be used for retrospective end-of-season assessments of severity, but due to delays in data availability are unlikely to be timely estimates of severity during in-season surveillance.

Highlights

Since 2009, in Ontario, reportable disease surveillance data has been used for timely in-season estimates of influenza severity
Trends over time: hospitalizations For each season, the total number of hospitalizations for laboratory-confirmed influenza cases captured in the reportable disease data was higher compared to those coded as laboratory-confirmed influenza hospitalizations (J09 and J10 diagnosis codes) in the hospital data (Table 1)
In our retrospective analyses we found that hospital administrative data provided a reliable estimate of trends of influenza-related hospitalizations and deaths, as well as the total seasonal numbers of hospitalizations, compared to the reportable disease data for the 2010–11 to 2013–14 influenza seasons in Ontario

Summary

Introduction

Since 2009, in Ontario, reportable disease surveillance data has been used for timely in-season estimates of influenza severity (i.e., hospitalizations and deaths). Information on hospitalizations and deaths for laboratory-confirmed influenza cases were obtained from two databases; the integrated Public Health Information System, which is the provincial reportable disease database, and the Discharge Abstract Database, which contains information on all in-patient hospital visits using the International Classification of Diseases, 10th Revision, Canada (ICD-10-CA) coding standards. Individuals presenting for care with influenza may be diagnosed with exacerbations of a chronic disease (e.g., heart failure, chronic obstructive pulmonary disease) rather than acute influenza infection. This may result in reduced laboratory testing of patients hospitalized with influenza, leading to less detection and reporting of the virus [2]. Testing biases are observed in less severe influenza seasons, which impacts detection and reporting [3]

Objectives

Methods

Results

Discussion

Conclusion