As IMD is a relatively rare disease and reporting is mandatory in Ontario, it was possible to link reportable cases to laboratory records via probabilistic linkage. The probabilistic linkage of case-level records in iPHIS and PHOL had a 77.3% (551/713) success rate and provided additional case information that enhanced our ability to describe the epidemiology of serogroup B disease in the province. While the record linkage of iPHIS to PHOL records remains incomplete (19.2% iPHIS records and 5.3% unique PHOL records were unmatched), it is not unexpected to have mismatches for a number of reasons. First, both data sources are passive systems and thus rely solely on healthcare providers, public health providers, and laboratories reporting IMD cases and forwarding N. meningitidis
isolates to the public health laboratory for serogrouping. Second, prior to the change in case definition on April 1, 2009, confirmed IMD cases were based on the presentation of relevant clinical symptoms of meningococcemia both with and without laboratory confirmation [13
]. Third, non-viable isolates from hospital laboratories could have been genogrouped using polymerase chain reaction and these would not be received at PHOL. Fourth, we did not have access to nominal data for this analysis and this limited our abilities to improve record linkage through deterministic linkage.
While it is difficult to determine whether our data is a true representation of all serogroup B cases in Ontario, we expect that most cases are captured either through the reportable disease information system or PHOL data given that this disease is associated with high public dread, has extremely high probability of hospitalization and microbiologic testing of the infected individual, and because PHOL is the only laboratory in the province that provides serogrouping of viable isolates.
In the last 11 years, as the incidence of serogroup C has decreased serogroup B has accounted for a greater proportion of IMD cases and has become the most prevalent cause of IMD in Ontario. The increase in the proportion of serogroup B IMD coincides with the introduction of MCCV into the provincially funded routine childhood immunization program in 2004 and a subsequent decrease in the incidence of serogroup C disease among all age groups; this should not be interpreted as an increasing rate in serogroup B disease. Our data were consistent with the natural fluctuation of the disease [4
] and also suggests that there were no predominant strains emerging in the province. As there was a change in case definition during the period under surveillance (i.e., April 2009), and the revised case definition was more specific, this could have resulted in a small decrease in reported cases in 2009 and 2010.
Consistent with the national data [4
], our findings reveal that people with serogroup B disease tend to be young, with infants having the highest rates. Infants represented only 1.1% of Ontario’s population in 2010 yet accounted for 21% of the overall disease burden with the majority (72.7%) of cases in younger infants less than 6 months of age. This will have important implications for vaccination programs once a vaccine that protects against this serogroup becomes available. Decision-makers will need to consider the age of vaccination and the number of cases that will be preventable through vaccination, as well as possible herd immunity effects. With an infant vaccination program, the age group with the highest incidence could be targeted, but cases occurring in very young infants will not be preventable unless they are protected by herd immunity, as has been demonstrated for serogroup C immunization. Given that novel meningococcal B vaccines are based on sub-capsular proteins rather than polysaccharides, and there are no published efficacy or effectiveness studies on these vaccines, we do not yet know whether these vaccines would result in herd immunity as observed from MCCV and other bacterial conjugate vaccines. The increase in median age throughout the surveillance period was of interest despite the lack of statistical significance, and is worth monitoring in the future, as changes in age distributions of serogroup B IMD can further impact vaccine decision-making.
Our crude NNV to prevent a single case of disease is high, in excessive of 30,000 infants, yet this is conservative as it assumes that all cases under one year would be vaccine preventable. For the calculation we used a vaccine effectiveness of between 70 and 80%, yet this is an assumption as the true value is not known. As noted in our results, approximately 70% of our infant cases occurred among infants under 6 months of age and these cases may not be vaccine preventable depending on age at vaccination and duration of time to mount an immune response. In a phase IIb clinical trial, Gossger and co-authors [15
], found that a schedule of three doses of Novartis’ novel multicomponent meningococcal B vaccine (4CMenB) given to infants at 2, 4, and 6 months of age, and in an accelerated schedule at 2, 3, and 4 months of age were necessary to achieve optimal immunogenicity. This would suggest that disease in infants less than 6 months of age, using a 2, 4, and 6 month schedule, which is typical in Canada, may not be vaccine preventable. Using this assumption the revised NNV would increase to over 120,000 infants. Although applying the number needed to treat concept to vaccines is not new, there is no agreed upon NNV threshold for vaccine decision-makers.
Given the rarity of the disease, the rates or proportions that we calculated may be unstable due to small numbers and therefore the results should be interpreted with caution. Our dataset was substantially limited in the amount of clinical data collected despite these fields being included in the reportable disease system, highlighting the importance of ensuring the collection of all relevant clinical data for reportable diseases including type of syndrome and presence of complications, to allow for a better understanding of the burden. As there are currently no licensed meningococcal B vaccines in Canada, our crude NNV calculations were based on several assumptions including vaccine efficacy, vaccine uptake and no herd immunity and is meant to give a conservative estimate until these parameters are better defined, if and when a vaccine becomes available.