Intended for healthcare professionals


Decrease in effectiveness of routine surveillance of Haemophilus influenzae disease after introduction of conjugate vaccine: comparison of routine reporting with active surveillance system

BMJ 2000; 321 doi: (Published 23 September 2000) Cite this as: BMJ 2000;321:731
  1. Babatunde Olowokure, specialist registrar in public health medicine (bolowokure{at},
  2. Jeremy Hawker, consultant epidemiologistb,
  3. Iain Blair, consultant in communicable disease controlc,
  4. Nick Spencer, professor of community child healtha
  1. a Section of Child Health, School of Postgraduate Medicine, University of Warwick, Coventry CV4 7AL
  2. b Department of Public Health and Epidemiology, University of Birmingham, Edgbaston, Birmingham B15 2TT,
  3. c Department of Public Health, Sandwell Health Authority, West Bromwich B70 9LD
  1. Correspondence to: B Olowokure, Public Health Laboratory Service Communicable Disease Surveillance Centre, West Midlands, Heartlands Hospital, Birmingham B9 5SS
  • Accepted 15 May 2000

In October 1992 routine immunisation with Haemophilus influenzae type b conjugate vaccine was introduced in the United Kingdom, and the incidence of disease was subsequently reported to have decreased 15-fold.1 The surveillance systems in place were primarily routine and were known to underestimate the burden of invasive H influenzae disease.2 This study aimed to determine whether underreporting continued after introduction of the conjugate vaccine, and how this might affect the reported success of the vaccine. Results of routine surveillance were compared with active surveillance for invasive H influenzae disease in the West Midlands health region of England.

Comparison of cases of invasive Haemophilus influenzae disease in children <5 years identified by active surveillance and reports to the Communicable Disease Surveillance Centre (routine system) before and after introduction of the vaccine: West Midlands Health Region 1990-4

View this table:

Subjects, methods, and results

Invasive H influenzae disease was defined as an illness in which the organism was isolated from a sterile site in children aged <5 years admitted to hospital in the West Midlands from October 1990 to September 1992 (pre-vaccine) and from October 1992 to September 1994 (post-vaccine).

Data from voluntary laboratory reports to the Public Health Laboratory Service Communicable Disease Surveillance Centre (routine system) were compared with a system that combined data from the Communicable Disease Surveillance Centre, statutory notifications, laboratory records, hospital paediatricians, and the British Paediatric Surveillance Unit study of conjugate vaccine failures (active system). Hospitals and laboratories were visited to validate reports and collect demographic data. Only children aged <5 years who lived in the West Midlands are included in the analysis.

Of 244 West Midlands cases identified, only 200 cases had been reported to the routine surveillance system (table). Significantly fewer cases were reported to the routine surveillance system than to the active surveillance system in the post-vaccine period than in the pre-vaccine period (P=0.0018, table). Overall, the proportion of children aged <2 years identified by the routine surveillance system was significantly lower than that of children aged >2 years (P=0.0065, table). Reporting by ethnic group, sex, or mortality did not differ significantly. Of seven cases reported to the British Paediatric Surveillance Unit, only three were reported to the routine surveillance system. Incidence rates based on reports to the routine surveillance system for the pre-vaccine and post-vaccine periods were 23.4 per 100 000 children <5 years old (95% confidence interval 19.9 to 27.2) and 5.1 (3.6 to 7.1) respectively. This compares with 27.1 (23.3 to 31.2) and 7.7 (5.8 to 10.0) using active surveillance data. The table shows that effectiveness of routine surveillance decreased by 23% after introduction of the vaccine, with consequent overestimation of the effectiveness of the immunisation programme.


The reliability of surveillance systems can affect the observed effectiveness of public health interventions. H influenzae meningitis is a life threatening disease and most cases occur before the age of 2 years.3 In this study cases of H influenzae meningitis in this age group were those most likely to be missed. The publicity surrounding the launch of the vaccine and the national study of vaccine failures4 might have been expected to improve reporting efficiency. However, we found that the introduction of the vaccine was associated with significantly increased underreporting, perhaps because of factors such as complacency following the success of the vaccine in reducing the incidence of the disease or an assumption that reporting to the British Paediatric Surveillance Unit was sufficient. These findings are relevant as continuing surveillance is needed to assess the effectiveness of the programme, identify vaccine failures, and monitor changes in predominant strains of the organism. A meningococcal vaccine is being introduced into the immunisation schedule in the United Kingdom.5 To measure the impact of this and other new vaccines accurately, the quality of the data sources for communicable disease surveillance must be ensured.


We are grateful to the Public Health Laboratory Service Communicable Disease Surveillance Centre and all the consultant microbiologists, consultants in communicable disease control, immunisation coordinators, consultant paediatricians, and medical records managers who contributed to this study by providing information on their patients. We are also grateful to the Oxford Vaccine Group for providing information on cases reported to it as part of the British Paediatric Surveillance Unit study of Hib conjugate vaccine failures.

Contributors: BO and NS conceived the study. BO collected data in the field and was responsible for data management and for preparation of the manuscript. JH and IB participated in the execution of the study. All authors contributed to data analysis and interpretation and participated in manuscript revision and approval of the final version. BO and NS are guarantors for the study.


  • Funding This study was funded by a grant from the charities of the City of Coventry and partly funded by the Warwickshire Child Health Group.

  • Competing interests None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract