For many uncommon but important paediatric conditions we lack national data on incidence, disease burden and short-term outcomes. We also lack evidence of the impacts of preventive public health measures on disease rates and outcomes. To address these important information gaps, Paediatric Surveillance Units (PSUs) have been established in 14 countries. All are based on the model developed by the British Paediatric Surveillance Unit. All PSUs facilitate national, prospective, active surveillance for a range of conditions, with monthly reporting by child health specialists. For each case reported, clinicians subsequently provide clinical and epidemiological data. This method provides considerable advantages over traditional passive reporting systems, including timeliness and increased rates of reporting and ascertainment.¹

The International Network of Paediatric Surveillance Units (INoPSU) was established in 1998. INoPSU facilitates international collaboration among member PSUs and allows for sharing of resources, simultaneous data collection and hence comparison of data from different geographical regions (table 1).² Over 150 studies have been facilitated by INoPSU members on a range of infectious and vaccine preventable diseases, genetic and metabolic conditions, childhood injuries and mental health disorders. Over 10 000 paediatricians, servicing a population of >56 million children aged <15 years, contribute data to PSUs each month.

In this paper we describe the impact of data collected by PSUs, both individually and collectively as members of INoPSU, on public health outcomes, clinical care and research.

METHODS

INoPSU has 14 member countries as described in table 2. Further details are available at http://www.inopsu.com. Members communicate through sharing annual reports, a bi-annual conference and the INoPSU website. Most communication is informal, for example when one unit seeks opinion or advice from other units when developing a new surveillance study. Often, one PSU acts as a catalyst for international collaboration by initiating a study that is later adopted by other PSUs.

All PSUs use a similar surveillance method and the quality and completeness of data collected depends on the monthly participation of clinicians. Each month a report card listing up to 16 different conditions is sent to practising paediatricians and other child health specialists. Average monthly card return rates are high (81.4%) and reflect the value with which clinicians regard this activity (table 2). This high response rate is an important achievement because many PSUs are in contact with over 1000 clinicians every month. To facilitate high participation rates some units have changed from postal to e-mail reporting or have introduced web-based reporting. Because case ascertainment is unlikely to be complete with any system, use of additional sources of cases, such as laboratory surveillance systems, is encouraged, when appropriate and available, to maximise case finding.

RESULTS

Table 3 summarises the outcomes of a selection of surveillance studies, their impact on public health policy and clinical practice, and their role in stimulating further research. Study outcomes have been categorised into seven main areas:

• monitoring public health interventions,

• informing the development of new screening policy,

• highlighting international differences that impact on local public health policy,

• describing the epidemiology and features of child mental health disorders,

• facilitating molecular epidemiological studies,

• guiding paediatric clinical practice policy, and

• identifying and quantifying product safety hazards.

Monitoring public health interventions

Surveillance studies can be used to monitor incidence rates of vaccine-preventable diseases, effectiveness of vaccination programs, and occurrence of vaccine-associated adverse effects. In some countries, PSUs are the primary source of national data supplied to the WHO, allowing governments to fulfil their obligations to the global polio eradication initiative and to obtain certification as “polio-free”.³

Sharing of study protocols among four countries enabled comparison of data on infants hospitalised with pertussis. PSU data showed that many affected infants were too young to be vaccinated according to their country’s vaccination schedules, and that many acquired pertussis from their parents, another adult or an adolescent^4,5 (personal communication from Associate Professor Cameron Grant, NZPSU). A German study confirmed that a single dose of acellular pertussis vaccine protects against hospitalisation for pertussis.⁵ This information and approval of an acellular pertussis vaccine for adults, informed changes to vaccination schedules. In the Netherlands, the recommended age for first vaccination was decreased from 3 to 2 months.⁶ In Australia, the vaccination schedule now recommends vaccination of adolescents at 15–17 years of age⁴ and a trial of vaccination at birth has been initiated (personal communication from Professor Peter McIntyre, National Centre for Immunisation Research and Surveillance, Australia).

Surveillance for Haemophilus influenzae type b infections attests to the success of a combined vaccine in dramatically reducing disease rates, mortality and morbidity caused by this bacterium.⁷ Surveillance of subacute sclerosing panencephalitis by several PSUs confirmed this condition is now very rare in countries with measles vaccination programs, and that it is not a complication of measles vaccination.^8,9

Surveillance of congenital rubella syndrome (CRS) by five PSUs showed that although the two-dose MMR vaccination regimen in place since the 1990s has been very successful in decreasing disease rates, congenital rubella still occurs.^10,11 In Australia and Canada, most children with CRS are born to immigrant mothers with incomplete or interrupted immunisation programs, or to unvaccinated mothers who travelled overseas during pregnancy to countries where rubella is still prevalent.^10,11 However, some affected children are born to mothers who missed school-based rubella vaccination or experienced primary vaccine failure.¹¹ These results have several implications. Given increasing world migration rates, ongoing surveillance programs to detect importations of rubella into developed countries should be supported. Educational campaigns are necessary to raise awareness and ensure high (>85%) uptake for primary MMR vaccination. More importance should be placed on checking rubella antibody status, and vaccinating rubella-susceptible women, particularly immigrants.^10,11

Several PSU studies have provided data on vaccine-preventable diseases prior to the availability and/or introduction of a new vaccine. Although varicella is perceived to be a benign childhood infection, studies from five PSUs documented serious complications, for example encephalitis, cerebellar ataxia and necrotising fasciitis,^12,13 and the need for hospitalisation and surgical procedures in some children.¹⁴ The effects of varicella infection in utero and in the neonatal period have also been documented.¹⁵ As varicella vaccines have become more widely used, surveillance studies have been re-activated to provide current information and allow documentation of changes in disease rates and rates of serious complications of varicella.

Surveillance of invasive pneumococcal infections provided evidence on the inpatient burden of pneumonia, septicaemia and meningitis and supported the development of new vaccination policies.¹⁶

Informing the development of new screening policies

HIV/AIDS surveillance by five PSUs confirms that most children now acquire the virus from their mother through perinatal exposure.^17–19 PSUs have documented trends in disease frequency and outcome over time. Early surveillance showed that maternal HIV infection was frequently acquired through intravenous drug use or receipt of contaminated blood products, that HIV was rarely diagnosed antenatally, that maternal infection was often identified only when children developed an AIDS-related condition and that there were substantial rates of HIV transmission from mother to infant.^17–19 Recent surveillance shows that the most common means of maternal acquisition of HIV infection is now through heterosexual contact. Earlier (antenatal) diagnosis allows uptake of interventions (anti-retroviral agents, elective caesarean section and avoidance of breastfeeding) and has led to lower transmission rates.^18–20 These findings informed the development and implementation of national policy in the UK, and more recently in New Zealand, supporting the introduction of routine HIV screening during pregnancy.¹⁸

Studies on invasive group B streptococcal disease provided national epidemiological data in five countries (UK, Canada, Germany, Netherlands, Portugal). All studies reported the higher frequency and severity of early-onset compared with late-onset disease, and supported the development of guidelines that advocate either universal screening in late pregnancy or screening based on identifiable risk factors.^21–23 One ongoing study is assessing the clinical impact of one of these guidelines.²⁴

Rates of neonatal herpes simplex infection reported through the Canadian and Australian PSUs are considerably lower than rates reported in the USA.^25,26 PSU data highlight the importance of HSV-1 in congenital and neonatal infections. HSV-1 predominates in Australia and Canada, whereas HSV-2 is responsible for most neonatal infections in the USA.²⁶ PSU studies highlight the fact that many women are unaware of their infection prior to delivery and that, despite early and intensive treatment with antiviral medications, many newborns have disseminated disease with a high fatality rate.^27,28 These results have implications for screening, vaccine program development, community education and management.

Highlighting international differences that impact on local public health policy

Using the same case definition, seven PSUs monitored the association between Shiga toxin-producing Escherichia coli (STEC) and haemolytic uraemic syndrome (HUS) (table 3). HUS is endemic in Australia, UK, Canada, Latvia, New Zealand, Portugal and Switzerland, and outbreaks were reported in the UK, Canada and Australia. Diarrhoea-associated HUS was acquired through ingestion of contaminated food (mettwurst, meat pies) or well water, swimming in contaminated sea water and environmental contamination at a kindy farm where the public can interact with the animals.^29–31 Although E coli O157 was the organism most commonly implicated, there was considerable geographic heterogeneity. In particular, E coli O157 was rare in Australia where E coli O111:H− (an organism not isolated in any other country) predominated.³¹ Data from these studies supported public health measures to raise awareness of the disease and promote means to minimise the risk of acquisition, enhance diagnostic services, and improve food handling and manufacturing practices.³¹

In 1997 the BPSU implemented surveillance of progressive intellectual and neurological deterioration (PIND) in response to the emergence of variant Creutzfeldt-Jakob disease (vCJD). Canada, which at that time was free of bovine spongiform encephalitis, served as a comparison country. Both units used the same case definition and over a 7 year period six cases of vCJD were identified through the BPSU but none in Canada.^32,33 Results highlighted the numerous causes of PIND in childhood and the large number of cases for which a specific diagnosis could not be found even after extensive investigations and review by a panel of experts.^32,33 Similarly, 21% of cases of childhood dementia identified in Australia were idiopathic, indicating the need for further research and improved diagnostic services.³⁴

Describing the epidemiology and features of child mental health disorders

Little is known about the epidemiology of many uncommon child mental health disorders and PSUs have generated data to address this knowledge gap. These studies have raised awareness of child mental health conditions, identified the need for education of paediatricians regarding their diagnosis and management, and highlighted the need for improved access to multidisciplinary diagnostic and management services. The move by PSUs to monitor mental health disorders is important because they are increasing in incidence in paediatric practice worldwide.

The early onset eating disorder studies from Australia and Canada showed that children aged 5–13 years with eating disorders present with significant weight loss (average 6–7 kg) or failure to gain weight, psychological problems and medical complications, including bradycardia, hypothermia and hypotension.³⁵ Although early onset eating disorder mainly affects girls, a significant proportion of patients (14% in Canada and 25% in Australia) are boys, particularly in the younger age groups. Many children do not meet the DSM-IV diagnostic criteria for anorexia nervosa, indicating the need for establishing pre-adolescent diagnostic criteria.³⁵

The Australian study of conversion disorder described its epidemiology and clinical features. Over half of the children identified had multiple conversion symptoms, most commonly disturbance of voluntary motor function, sensory symptoms, pseudoseizures or respiratory problems. In many children antecedent stressors were identified, including family conflict or a personal history of mental health problems, predominantly depression and anxiety.³⁶ This study highlighted the chronicity of conversion disorder, its burden on families and health professionals, and the need for a multidisciplinary approach to ensure timely diagnosis and appropriate management.

Munchausen syndrome by proxy (MSBP) involves behaviours by care givers to deliberately mislead the doctor, including fabrication, exaggeration or falsification of information regarding their child’s illness, or physical interference to create or worsen symptoms and signs. British and Australian PSU studies estimate an incidence of MSBP of 0.5/100 000 children aged <15 years. Children with MBPS place considerable demand on health resources because they often require multiple admissions and many doctors are involved in their care.³⁷ The condition is exceedingly difficult to recognise and caution is needed when clinicians are faced with seemingly unexplained signs and symptoms in children.³⁷

Facilitating molecular epidemiological studies

Studies of Rett, Prader-Willi and Smith-Lemli-Opitz syndromes and the CHARGE association have estimated national incidence rates for these rare genetic disorders.^38–40 Surveillance studies have enabled establishment of cohorts that will be followed longitudinally to assess disease prognosis, have enabled researchers to investigate genotype–phenotype correlations and have highlighted the advantage of early genetic confirmation of the diagnosis for ensuring early access to multidisciplinary care.^40,41 Demonstrating the extent of health service usage by affected children has helped to justify establishment of new models of health service delivery.⁴² Members of cohorts have also consented to participate in additional research including case–control and randomised controlled trials.^43,44 Surveillance for Smith-Lemli-Opitz syndrome and CHARGE association identified significant behavioural and co-morbid diagnoses in older children, with implications for planning of future health resources.^41,45,46 PSUs also provided data on the burden of medium chain acyl CoA dehydrogenase deficiency that informed the implementation of tandem mass spectrometry for neonatal screening.^47,48

Guiding paediatric clinical practice policy

Publication of an alleged link between intramuscular vitamin K administration in newborns and later childhood cancer led several countries to alter their prevention policy and recommend oral vitamin K as an alternative for prevention of vitamin K deficiency bleeding. Publicity surrounding this allegation led some parents to refuse consent for vitamin K. Surveillance studies in seven countries (table 3) subsequently confirmed an increased number of patients, some with intracranial bleeding and severe neurological sequelae.^49,50 International comparison showed that the lowest incidence rates of such complications occur in countries using predominantly intramuscular vitamin K, confirming this as the most effective route of administration and leading to revision of practice guidelines.^50,51

Surveillance of Reye syndrome by the BPSU commenced in 1981 following reports of an association with the consumption of aspirin.⁵² This led to a ban on aspirin use in children aged <12 years. Subsequent monitoring demonstrated an almost total elimination of Reye syndrome in this age group in the UK. However, when cases continued to be reported in children aged >12 years, the ban was extended to cover all children aged <16 years.⁵³

Studies of fetal alcohol syndrome (FAS), conducted in Australia and New Zealand, suggested under-diagnosis of this condition. The APSU study showed that Indigenous children were over-represented, about one third were in foster care and almost half had an affected sibling, indicating missed prevention opportunities.⁵⁴ The APSU study prompted two surveys of health professionals that indicated lack of knowledge about FAS diagnosis and management, fear of stigmatising the child or family, and uncertainty about the appropriate advice to give to women regarding alcohol use in pregnancy.^54,55 The study highlighted the need for review of prevention policy including the content of National Health and Medical Research Council Australian Guidelines for alcohol use in pregnancy, development of educational materials for health providers and the community, and review of diagnostic and management models used internationally. Representatives from the APSU study team are in a unique position to influence those responsible for policy on prevention, education and research through membership of the Inter Governmental Committee on Drugs Working Party on Fetal Alcohol Spectrum Disorders in Australia, convened by the Ministerial Council on Drugs Strategy. Other PSUs have shown interest in developing surveillance studies on FAS.

An NZPSU study demonstrated that bronchiectasis was most common in Pacific Islander children and also more common in Maori children than those of European ethnicity. The study highlighted the need for earlier and more effective treatment of lower respiratory disease in disadvantaged children.⁵⁶

Identifying and quantifying product safety hazards

PSUs have been used to document rare injuries in children. A British study of poisoning by toy chemistry sets identified problems with packaging and led to amendments to the European Union legislation on toy safety and instructions on the packages.⁵⁷ A Canadian study identified that injuries from baby walkers continued to occur despite a voluntary ban on their sale from 1988. This led the Canadian government to prohibit the sale, advertisement and importation of baby walkers from 2004.⁵⁸ Another Canadian study documented the severity of lumbar spine fractures and spinal cord injuries in children aged 2–15 years who were inappropriately restrained in motor vehicles by lap-only belts. In this study, 25% of children became paraplegic following injury and children aged 5–8 years were at higher risk. Advocacy for optimal restraint use in motor vehicles and use of booster seats in older children led to legislation changes in some Canadian jurisdictions.⁵⁹ A study of serious seatbelt injuries based on the Canadian study is currently under way in Australia.

DISCUSSION

Active, national surveillance facilitated by PSUs is an important, versatile and relatively cheap epidemiological tool. PSUs facilitate research in children and provide new national data on the epidemiology, diagnosis, management and short-term outcome of a wide range of uncommon, but high impact and often neglected, conditions. They complement data derived from alternative sources. Surveillance allows practising paediatricians to participate in, and contribute to, research and evaluation efforts. High participation rates by paediatricians around the world reflect the importance with which they regard this activity.

The formation of INoPSU has been important for facilitating international cooperation and collaboration between PSUs, developing new surveillance methods, and enabling international comparison of study results. It has allowed us to identify geographic differences in disease aetiology, management, prevention and outcome and these data inform local health policy. The network continues to expand: new PSUs are under development in Trinidad-Tobago and Argentina. Furthermore, the surveillance methodology developed by PSUs has been used to create specialty surveillance units including the British Neurological Surveillance Unit and the UK Obstetric Surveillance System.

As we have demonstrated, surveillance by INoPSU members allows us to monitor the impact of established public health policy and provides new and timely data to inform the development of new policy. This information also impacts on health service planning, guides national legislation pertaining to injury prevention in children and young people, informs the development of screening programs, and provides insights into genetic epidemiology. Surveillance units are able to respond quickly to recognition of emerging diseases of public health importance (for example vCJD in the UK) by rapidly initiating studies to generate national data. For most conditions studied PSUs have compiled the only national data, and in many cases PSU studies have generated hypotheses that have catalysed future research.

Despite the fact that PSUs provide important data to inform public health initiatives, clinical practice and research agendas, no PSU has ongoing core funding. PSUs rely predominantly on time-limited government and competitive research funding. INoPSU has no current funding. The success of individual PSUs and INoPSU depends on the dedication and commitment of key individuals who ensure the continuation of this important work, often under difficult circumstances. However, national surveillance is essential for low frequency conditions and INoPSU members will continue to provide high quality evidence that contributes to improved health for children and young people.