Active surveillance of acute paediatric hospitalisations demonstrates the impact of vaccination programmes and informs vaccine policy in Canada and Australia

Sentinel surveillance of acute hospitalisations in response to infectious disease emergencies such as the 2009 influenza A(H1N1)pdm09 pandemic is well described, but recognition of its potential to supplement routine public health surveillance and provide scalability for emergency responses has been limited. We summarise the achievements of two national paediatric hospital surveillance networks relevant to vaccine programmes and emerging infectious diseases in Canada (Canadian Immunization Monitoring Program Active; IMPACT from 1991) and Australia (Paediatric Active Enhanced Disease Surveillance; PAEDS from 2007) and discuss opportunities and challenges in applying their model to other contexts. Both networks were established to enhance capacity to measure vaccine preventable disease burden, vaccine programme impact, and safety, with their scope occasionally being increased with emerging infectious diseases’ surveillance. Their active surveillance has increased data accuracy and utility for syndromic conditions (e.g. encephalitis), pathogen-specific diseases (e.g. pertussis, rotavirus, influenza), and adverse events following immunisation (e.g. febrile seizure), enabled correlation of biological specimens with clinical context and supported responses to emerging infections (e.g. pandemic influenza, parechovirus, COVID-19). The demonstrated long-term value of continuous, rather than incident-related, operation of these networks in strengthening routine surveillance, bridging research gaps, and providing scalable public health response, supports their applicability to other countries.

Sentinel surveillance of acute hospitalisations in response to infectious disease emergencies such as the 2009 influenza A(H1N1)pdm09 pandemic is well described, but recognition of its potential to supplement routine public health surveillance and provide scalability for emergency responses has been limited. We summarise the achievements of two national paediatric hospital surveillance networks relevant to vaccine programmes and emerging infectious diseases in Canada (Canadian Immunization Monitoring Program Active; IMPACT from 1991) and Australia (Paediatric Active Enhanced Disease Surveillance; PAEDS from 2007) and discuss opportunities and challenges in applying their model to other contexts. Both networks were established to enhance capacity to measure vaccine preventable disease burden, vaccine programme impact, and safety, with their scope occasionally being increased with emerging infectious diseases' surveillance. Their active surveillance has increased data accuracy and utility for syndromic conditions (e.g. encephalitis), pathogen-specific diseases (e.g. pertussis, rotavirus, influenza), and adverse events following immunisation (e.g. febrile seizure), enabled correlation of biological specimens with clinical context and supported responses to emerging infections (e.g. pandemic influenza, parechovirus, . The demonstrated long-term value of continuous, rather than incident-related, operation of these networks in strengthening routine surveillance, bridging research gaps, and providing scalable public health response, supports their applicability to other countries.

Background
The essential role of hospital-based sentinel surveillance in identifying emerging infections and measuring the incidence of severe disease was highlighted by Thomson and Nicoll in a 2010 editorial in Eurosurveillance referring to surveillance activities in Europe during the 2009 influenza A(H1N1)pdm09 pandemic [1]. They argued that given demands on clinicians in disease outbreaks, sentinel hospital-based surveillance needed to be ongoing, allowing it to be augmented, rather than established de novo, in times of crisis [1]. They identified a range of potential roles for sentinel hospital networks in Europe, primarily collection of clinical data, linked to laboratory and epidemiological data, supporting rapid, evidence-based outbreak responses. Beyond outbreak response, hospital networks have been established in low-middleand high-income countries to provide quality data for immunisation programmes, infection control, and seasonal influenza [2][3][4][5][6][7][8].
Most hospital-based surveillance networks are focused on specific diseases or syndromic targets. This article analyses the contributions of two active, paediatric hospital-based sentinel networks that, since their inception, have played wide-ranging roles in public health surveillance: the Canadian Immunization Monitoring Program, Active (IMPACT) established in 1991, and the Australian Paediatric Active Enhanced Disease Surveillance (PAEDS) network established in 2007. These networks, which contribute to surveillance

History and contributions
Both IMPACT and PAEDS arose from identified surveillance gaps in child health outcomes related to vaccine safety and VPDs. Addressing these required the establishment of national surveillance programmes due to the relatively small numbers of both paediatric patients and tertiary care centres in Canada and Australia.

Canada
The need for a hospital-based active surveillance system to reliably detect serious adverse events following immunisation (AEFIs) was recognised following detection of an increase in aseptic meningitis associated with the Urabe mumps vaccine strain in 1986-1988 by virologists at several Canadian children's hospitals [9]. Investigation of this signal, not identified by the passive system, led to replacement of the Urabe vaccine with a safer vaccine. IMPACT began in 1991 as a collaboration between Health Canada and the Canadian Paediatric Society (CPS) at five paediatric tertiary care centres in five provinces. The first surveillance targets included neurological admissions (e.g. acute flaccid paralysis (AFP), encephalopathy, seizure) and several VPDs (e.g. pertussis) ( Table 1 and Supplemental Content 1) [9]. IMPACT expanded to 12 centres in eight provinces by 1999, capturing approximately 90% of paediatric tertiary care beds in Canada [9].
For 29 years, IMPACT has collected epidemiological data for AEFIs and diseases that are current or future targets for vaccine prevention, demonstrating the effectiveness of new immunisation programmes, including, meningococcal conjugate, pneumococcal conjugate and varicella vaccines (

Australia
The PAEDS system was established in 2007 to support Australian compliance with World Health Organization (WHO) AFP surveillance standards as part of polio eradication efforts, and to conduct surveillance for varicella hospitalisations following vaccine introduction and two AEFIs potentially associated with varicella and rotavirus vaccination programmes (seizures and intussusception) ( Table 1 and Supplemental Content 1) [13]. PAEDS was funded by the Australian Government as a pilot project in four paediatric hospitals in four states. PAEDS subsequently expanded to seven hospitals in six states and territories, covering around 80% of tertiary paediatric beds. The scope of PAEDS was enlarged over time to provide key evidence regarding vaccine

Network organisation and procedures
Both IMPACT and PAEDS utilise the contribution of trained surveillance nurses at each hospital, supervised by volunteer paediatric clinicians who act as site investigators. Nurses screen hospital and emergency department admission lists for conditions under surveillance, review medical records, retrieve immunisation records, and report cases electronically on standardised case report forms to the national coordinating centre (Figure and Table 1).
The national coordinating centre submits data to national public health authorities at least quarterly for incorporation into national datasets. In Canada and Australia, centres report AEFIs directly to regional and national public health authorities.
Annual in-person meetings and standardised training have been important to maintain group cohesion and national consistency, while opportunities for data  analysis, peer-reviewed publications, and improved policy and practice sustain investigator engagement.

Ethical statement
IMPACT and PAEDS surveillance is conducted and reported in line with the Declaration of Helsinki, as revised in 2013. Ethics and/or hospital approvals are in place at participating institutions (Supplemental Content 2).

Funding and resources
IMPACT is supported primarily by federal funding and managed by a non-profit organisation (CPS). This unique arrangement has allowed alternate sources of funding from provincial governments and industry to augment federally funded activities and support additional surveillance targets (e.g. rotavirus), providing stability for the network, while ensuring investigators retain independence in data collection, analysis, and publication.
PAEDS is supported by federal, state and territory government funding, and provides a platform for researchers to use the PAEDS infrastructure on a cost-recovery basis for other serious childhood conditions, such as Kawasaki disease. PAEDS has not received pharmaceutical industry funding.
Network funding, together with funding garnered to add new conditions, provides part-time support for one nurse per site, a national nurse coordinator, and data centre staff. Site investigators provide in-kind support.

Hospital-based surveillance complements public health surveillance Canada
Public health surveillance of select VPDs is mandated by provincial and territorial governments with voluntary reporting to the Canadian Notifiable Disease Surveillance System but captures only disease onset date, sex and age. IMPACT captures additional variables for hospitalised cases including co-morbid conditions, concurrent and past infections, immunisation history, need for intensive care, and outcome at discharge. This information allows in-depth characterisation of disease burden and risk groups, as well as estimation of vaccine effectiveness, and informs cost-effectiveness analyses. Biological specimens are collected for select VPDs (e.g. Streptococcus pneumoniae, Neisseria meningitidis, rotavirus) enabling strain characterisation and monitoring for strain replacement ( Table 2) [10].
IMPACT provides the only information on paediatric hospital admissions for influenza in Canada. IMPACT data are incorporated into 'FluWatch', Canada's national influenza and influenza-like illnesses surveillance system [14]. IMPACT's weekly reporting during the influenza season allows public health to assess influenza transmission and severity by person, place and time, as well as the impact and burden of influenza epidemics in real time.
The Canadian AEFI Surveillance System (CAEFISS), the national post-market vaccine safety monitoring system, relies primarily on spontaneous reporting of AEFIs to public health [15]. IMPACT contributes > 50% of serious AEFIs and 70-90% of neurological AEFIs reported to CAEFISS [15]. In 1998, IMPACT identified an increase in disseminated Bacillus Calmette-Guérin (BCG) disease in Indigenous children with undiagnosed primary immunodeficiency, prompting changes to BCG vaccination recommendations in Canada [9].

Australia
Due to the existence of robust laboratory-based VPD surveillance through the National Notifiable Diseases Surveillance Scheme, which also captures biological specimens for select VPDs (such as those described for Canada) [16], the PAEDS network has focused on conditions where there is syndromic diagnosis (e.g. AFP, encephalitis), limited sensitivity or utilisation of laboratory tests (e.g. varicella), or where gaps in capture of immunisation status and clinical severity exist (e.g. paediatric influenza).
After 2007, PAEDS emerged as the reporting source for ca 80% of AFP cases to the Communicable Diseases Network of Australia's polio expert panel [13], resulting in Australia consistently exceeding the WHO reporting target. Enhanced studies of encephalitis via PAEDS facilitated evaluation of emerging viral infections, including enterovirus 71 and parechovirus (Table 3) [17].
Australia monitors influenza activity through a variety of complementary surveillance systems [18]. Sentinel hospital surveillance for influenza is conducted in collaboration with the InFLUenza Complications Alert Network (FluCAN) [18], which captures data from 22 sites across Australia, including seven PAEDS sites, enabling real-time tracking of a representative number of children. Paediatric influenza surveillance, including characterisation of over 1,300 paediatric hospitalisations during the 2017 influenza season, provided evidence to prompt state and territory funding of influenza vaccines for children aged 6-59 months from 2018, and inclusion on the National Immunisation Program from 2020 [2,19].

Rapid response capacity
Both networks demonstrated capacity to respond to outbreaks of emerging diseases during the 2009 influenza A(H1N1)pdm09 pandemic. IMPACT scaled up its influenza activities to continue throughout the summer (June-August) and provided one of the earliest reports on the paediatric burden of influenza A(H1N1) pdm09 disease in the Northern Hemisphere during the first pandemic wave [20]. PAEDS developed questionnaires and protocols for identifying hospitalised cases of influenza rapidly, following pilot work in individual hospitals [21]. During the coronavirus disease  pandemic, PAEDS has been capturing data on laboratory-confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) infections leading to hospitalisation or Emergency Department visit since March 2020. Surveillance for Paediatric Inflammatory Multisystem Syndrome Temporally associated with SARS-CoV-2 (PIMS-TS; and also known as Multisystem Inflammatory Syndrome in Children (MIS-C) in the USA), a newly described inflammatory syndrome occurring during or after SARS-CoV-2 infection in children, commenced in May 2020 [22,23].

Surveillance approaches of IMPACT and Australian PAEDS programmes
Eligible to report Eligible to report

Representativeness and ethics
IMPACT and PAEDS networks are based in paediatric referral centres and therefore do not cover the whole population. Calculation of disease incidence has been limited to severe diseases, such as invasive meningococcal disease (IMPACT) [11] or encephalitis (PAEDS), where either most paediatric cases are admitted or transferred to an IMPACT or PAEDS centre, or the incidence is low enough to also capture cases admitted to regional or community hospitals.
To ensure complete case capture, IMPACT operates without obtaining informed consent or enrolling individual participants. This requires the data collected to be available in a hospital chart or immunisation record. Patients or caregivers are not able to add or clarify missing data.
PAEDS originally commenced surveillance requiring informed consent to allow patients to enrol for data collection. However, this resulted in non-inclusion of patients whose parents had limited spoken English or health literacy. PAEDS now operates under a national ethical framework that allows capture of a minimal deidentified dataset for all cases. Parents or caregivers can 'opt out' of their data being used and consent is still obtained to gain additional information or to opt into additional studies via parent/caregiver interview.
Another challenge faced by both systems relates to variations in capacity in the event of severe disease epidemics. During the record breaking 2017 influenza season in Australia [2], a fivefold increase in hospitalisations (cf.d with previous years) diverted nurse time away from prompt recording of other surveillance conditions.

Opportunities for active hospital-based surveillance
Hospital-based surveillance systems in high-, middle-and low-income countries, such as the Influenza Monitoring of Vaccine Effectiveness Network (I-MOVE), Healthcare-associated Infections Surveillance Network (HAI-Net) in Europe, Global Rotavirus and Invasive Bacterial Vaccine Preventable Diseases Surveillance Networks (IB-VPD), and AEFI surveillance network in the Americas have generally focused on a specific disease or syndromic target [3][4][5][6][7]24,25]. However, collectively they represent surveillance activities similar to IMPACT or PAEDS.
Population registries and linked databases have also been used to evaluate vaccine safety and effectiveness [26,27]. However, they are limited to high-income countries, case capture may be incomplete for certain conditions (e.g. varicella), and capacity for rapid response, detailed clinical data collection and linkage to biological specimens varies [28].
IMPACT and PAEDS have demonstrated that the same platform and similar surveillance methodologies can be applied to study a broad range of diseases and syndromes of public health importance. Conditions under surveillance can be added in response to new vaccines, vaccine safety concerns and emerging diseases, while others can be discontinued or modified. This provides efficiencies with respect to staffing time, as well as flexibility and responsiveness in the event of disease outbreaks. When emerging diseases or other conditions of concern arise, established networks like I-MOVE, IB-VPD may be well placed to rapidly expand their surveillance targets. Sharing of standard surveillance protocols may also help low-and middle-income countries expand their surveillance capacity [25]. Exploration of this concept may warrant incorporation of hospital-based surveillance networks into emerging infectious disease and AEFI surveillance plans. Capacity to activate surveillance platforms to rapidly respond to communicable disease emergencies, particularly those threatening global health security, such as the COVID-19 pandemic, is recognised as essential [29].

Conclusions
IMPACT and PAEDS have been implemented successfully to address gaps in, and add value to, public health surveillance in two countries with different needs and health systems. The adaptability of both networks to changing public health priorities in their respective countries has been critical to their success. Active hospital-based sentinel surveillance systems can leverage efficiencies gained by monitoring for more than one condition to play multiple roles in informing public health policy and responding to public health emergencies. Existing surveillance systems should consider their potential to expand conditions under surveillance, particularly as the need to evaluate health interventions and monitor for emerging infectious diseases, such as COVID-19 grows.
Funding statement. The Canadian Immunization Monitoring Program Active (IMPACT) is a national surveillance initiative managed by the Canadian Paediatric Society and conducted by the IMPACT network of paediatric investigators on behalf of the Public Health Agency of Canada's Centre for Immunization and Respiratory Infectious Diseases. The Paediatric Active Enhanced Disease Surveillance (PAEDS) network is coordinated by the National Centre for Immunisation Research and Surveillance (NCIRS) and receives funding from the Australian Government Department of Health, and participating state and territory departments of health in Western Australia, South Australia, Victoria, New South Wales, Queensland and Northern Territory. Funding for work on influenza is received in part from the Australian Government Department of Health, via the Influenza Complications Alert Network (FluCAN) and from NHMRC Partnership Grant (#1113851). Funding from the NHMRC has also contributed to surveillance and studies of other PAEDS conditions.

Conflict of interest
KAT has received grants from GlaxoSmithKline and consultancy fees from Pfizer outside the submitted work. SAH has received grants from Pfizer and GlaxoSmithKline outside the submitted work. HSM is an investigator on vaccine trials sponsored by industry. Her institution receives funding for investigator led research (GSK, Pfizer, Sanofi-Pasteur). HSM acknowledges support from the National Health and Medical Research Council NHMRC (APP1155066). CCB acknowledges support from the NHMRC (APP1111596). The other authors have no disclosures.