Non-specific Effects of BCG in Under-five Children

Hypothesis: some of the variation observed in child and under-5 mortality, in the incidence of TB, and in severe and mild morbidity in under-5 children might be partially explained by a reduction in coverage rate of BCG since 2016. The reduction in the BCG coverage rate, that resulted from only high-risk children receiving BCG at birth, might have prevented non-vaccinated children from developing trained immunological responses as effectively as BCG vaccinated children.

Aim: to investigate the incidence of specific and non-specific effects of BCG by comparing the incidence of TB disease and infection, mild and severe morbidity and mortality in their first five years of life of children born between 2010 and 2021 according to their BCG status.

Methods Study design: Population based historical (retrospective) birth cohort study Setting : The researchers will investigate the specific and non-specific effects of BCG based on the shift in the BCG vaccination policy that occurred in Portugal in July 2016, when a consistent low incidence of TB led to the revision of the National Immunization Plan. Since then, BCG started to be administered only to new-borns considered to be at higher risk of TB infection, thus creating an opportunity to compare BCG vaccinated and non-vaccinated children.

Participants: Under-5 children born and registered in Portugal between 1 July 2010 and 30 June 2021. This time period will allow to compare the cohorts and exposed and non-exposed to BCG children before and after the 2016 change in the BCG immunization strategy and characterize children not vaccinated with BCG and their health outcomes.

Eligibility criteria: children born in Portugal, live-births and births coded as normal new-born Exclusion criteria: new borns with low-birth weight (<2kg), premature (<37 weeks) and known or suspected HIV infections will not be included in the study.

Follow-up period: until completion of 5 years of age or 30 June 2022. Data sources / measurement: birth registry, vaccination registry, the information systems of death certificates, epidemiological surveillance of notifiable diseases and TB Diagnosis Related Groups and primary health care. These systems contain information on: sex, date of birth, nationality, socioeconomic status, place of residence, health of the mother, health of the new-born (including if BCG was administered at the maternity), and primary health care unit where the child is to be followed up, cause and date of death and date of death, vaccines administrated, date of vaccination, and place of residence, hospitalizations and emergency department visits in public hospitals (including date of admission and discharge, diagnosis and type of admission, motive for admission, weight at birth, DRG code, MDC code), visits to primary health care units and motive, presentation of TB, date of diagnosis, type of treatment, other diseases before TB, risk groups and outcome.

Exposure: having received the BCG vaccine during the first year of life. Measured using the vaccination registry and variables are defined as BCG (Y/N, number of doses) and time from birth to BCG (in days).

Primary outcomes:

Death: death (Y/N); age at time of death (in days); and cause of death. Data on TB diagnosis will include confirmed case (Y/N), presentation of TB, age at diagnosis (in days), duration of treatment and outcome of treatment.

Severe morbidity: hospitalization and visits to the emergency department of the national health system and will include hospital admissions and emergency department visits during the follow-up period (Y/N and number), age (in days) at hospital admission or emergency department visits, length of stay for hospitalizations (in nights); major diagnostic categories for hospitalization and emergency department visits.

Mild morbidity: contacts with the medical doctor or nurse at a primary health centre unit due to disease or ill-health during the follow-up, besides contacts related to recommended child-surveillance scheme or other reasons besides health. Measured through visits to primary health care centre due to disease/ ill-health (Y/N and number), age at visits (in days) and diagnosis.

Mortality and morbidity pertaining to external causes and accidents will not be included in the analysis.

Secondary outcomes: mortality and morbidity profile of under-5 children between 2010 and 2022; the profile of non-vaccinated children before and after 2016, and hospital and primary health care utilization profile of under-5 children.

Statistical methods The researchers will identify, in each database, the data needed and whether a unique identifier exists and can be provided with the dataset. In case a unique identifier is provided, researchers will combine information based on that unique identifier. In case a unique identifier is not provided (most probable scenario) data linkage will be done using a range of proxy identifiers (e.g., date and place of birth, sex, place of residence) to identify probable matches. The type of data linkage method to be used will depend on the type and quality of the linkage variables available on the datasets. However, it is anticipated to have to use a combination of deterministic and probabilistic methods. Before applying linkage methods, data will be cleaned and standardized data, thus identifying and removing errors and inconsistencies (e.g. 11/04/2016 or 2016/04/11). Given the expected size of the datasets, sets of blocking attributes will be selected (e.g., sex, date of birth, initials) and record pairs with the same matched attributes within blocks compared. After, record pairs will be compared for each linkage variable and an agreement score computed. This score will be used to weight the probability of a record pair belonging to the same child.

After obtaining the final database, person-time incidence rates for primary outcomes will be computed. Kaplan Meier method will be used to compare the probability of surviving 1st and 5th year of life or of not having a hospitalization, emergency department visit, a mild/ severe morbidity episode during the follow-up period according to exposure. Log rank test will be used for differences in survival rates between exposed and non-exposed children and hazard ratios (with confidence intervals) for quantifying the differences. Proportional hazards regression analysis will be used to explore the effects of several variables on the survival outcomes. If missing data is below 5%, we will use complete-case analysis. Otherwise, the frequency and patterns of missing data will be analysed and if appropriate, multiple imputation techniques used.

Confounders to be measured (and controlled for) are sex, health unit, completeness of and timeliness of vaccination status and scheme, socioeconomic status and BCG strain.