Vaccinations

Quality description

The quality description of the statistics on vaccinations assesses the reliability and suitability of the statistics for different purposes in accordance with the quality description recommendation of the Official Statistics of Finland (OSF).

Relevance
Method description
Accuracy and reliability
Comparability
Timeliness and punctuality
Accessibility and clarity

Relevance

A computerized, national vaccination register is an important tool for assessing the vaccine uptake in a population in real-time. The performance of vaccination programs can be evaluated, time trends monitored and sub-populations with low vaccination coverage identified. Most importantly, by linking a national vaccination register with national health registers, the vaccines' impact – both effectiveness and safety – can be evaluated more comprehensively, which in turn will aid in formulating the best possible national immunization program.

Method description

Vaccination data for the Finnish Vaccination Register is collected from patient data systems following the data model of the National Patient Data Archive (KanTA AR/Ydin¹). Most of the National Immunization Program is implemented in public health care, such as well-baby clinics. Vaccination records have been collected continuously as part of the national social and health care reporting (Avohilmo) data collection, mostly since 2012². (See the description of the primary health care statistic). Vaccinations administrated elsewhere than in public primary health care are planned to be collected in the same way from patient data systems as part of national data collections. For the moment the Finnish Vaccination Register contains only vaccinations given in public primary health care (situation of 04/2016). Additional data has been collected retrospectively from patient records of the health care centers. Retrospective data collections contain vaccination records from January 2009 onwards.

Accuracy and reliability

Vaccination data are submitted in batches containing new vaccination records as well as duplicates or corrections of previously submitted records. In case several records describe the same vaccination event (defined by the vaccinee, the administered vaccine and the date of vaccination), only the record that was received first is kept for further processing and analysis. Records from Avohilmo are preferred over retrospective data. The vaccinee is defined by a personal identification code which is assigned to all permanent Finnish residents. The vaccine is identified by batch number or trade name considering also typing errors and spelling mistakes (Table 1). Each vaccination event is further described by the health care center, where the vaccination was recorded, and the municipality, where the vaccinee was living when the vaccination was recorded.

	2012	2013	2014	2015
Vaccine identified	98.7	99.1	99.4	99.7
by batch number	93.9	95.4	96.3	97.2
by trade name	3.8	3.0	2.9	2.2
Vaccine not identified	1.3	0.9	0.5	0.3

Table 1: Percentage of rows by vaccine identification and year of vaccination (situation as of 04/2016)

Due to recording conventions and problems in data dispatch, temporal deficiencies in/of vaccination records have occurred in some health care centers. The completeness of data is assessed every month in each health care center area.

Considering the the whole population, a health care center’s records are assumed to be sufficiently complete, when the ratio of the monthly number of vaccination events other than seasonal influenza vaccination and the yearly count of residents living in the municipalities served by the health care center has reached on average in the respective and the previous five months the cut-off of 0.2%. This cut-off is based on the fact that only a few (mostly booster) vaccinations are recommended throughout the year to the general population.

Considering children younger than two years, a health care center’s records are assumed to be sufficiently complete, when the ratio of the monthly number of DTaP-IPV-Hib and MMR vaccination visits and the number of children younger than two years living in the municipalities served by the health care center has reached on average in the respective and the previous five months the cut-off of 8%. This cut-off is based on the fact that children are expected to receive three DTaP-IPV-Hib and one MMR vaccination during their first 24 months of life (monthly expectation: 16%).

Only health care centers with complete records for all the months covered by the observation period of interest are included in nationwide reports. For example, for a cohort of children up to two years of age, records need to be sufficiently complete for the whole duration of 2 years of life, whereas for influenza vaccination, records need to be sufficiently complete for the duration of the influenza season. By the end of 2015, 99% of the health care centers in Finland were submitting near online vaccination data.

Comparability

Childhood vaccination coverage from 1995 onwards

Historically, childhood vaccination coverage has been studied every second year based on a random sample of 1000 newborns, whose vaccination records scheduled to be given during the first two years of life have been collected³. The first of these studies considered children born 1995, the last study completed in March 2015 for those born 2012. The accuracy and reliability of these survey studies is high: (95% confidence interval +/- 1.5 percent). Starting from the birth cohort of children born in 2012, childhood vaccination coverage has also been reported based on the Finnish Vaccination Register and including only those children/ health care centers with presumably complete vaccination records (see Accuracy and reliability). The potential amount of missing records in the Finnish Vaccination Register can be evaluated by comparing coverage figures based on survey and register data (Table 2).

	Vaccination coverage based on survey data	Vaccination coverage based on register data
Rota (1 dose)	92.4% (90.5% - 93.9%)	91.1% (89.1% - 92.8%)
Rota (3 doses)	90.0% (87.9% - 91.7%)	83.4% (80.9% - 85.6%)
PCV (1 dose)	94.4% (92.7% - 95.7%)	92.5% (90.6% - 94.0%)
PCV (3 doses)	92.3% (90.4% - 93.8%)	85.5% (83.1% - 87.6%)
DTaP (1 dose)	98.3% (97.3% - 99.0%)	96.6% (95.3% - 97.6%)
DTaP (3 doses)	97.8% (96.6% - 98.5%)	92.1% (90.2% - 93.6%)
Polio (1 dose)	98.3% (97.3% - 99.0%)	96.7% (95.4% - 97.7%)
Polio (3 doses)	97.8% (96.6% - 98.5%)	92.1% (90.2% - 93.6%)
Hib (1 dose)	98.2% (97.1% - 98.9%)	96.6% (95.3% - 97.6%)
Hib (3 doses)	97.5% (96.3% - 98.3%)	91.9% (89.9% - 93.4%)
MMR (1 dose)	96.0% (94.5% - 97.1%)	92.7% (90.9% - 94.2%)

Table 2: Vaccination coverage (and 95% confidence intervals) in 945 sampled children born 2012 based on health care center survey and vaccination register data

Seasonal influenza vaccination coverage from 2007 onwards

From 2007 until 2009, the seasonal influenza vaccine coverage was measured based on aggregated reports provided by the Finnish health care centers. From 2009 to 2012, the coverage was assessed based on data collected retrospectively from the patient records of the health care centers. Since 2012, the coverage has been calculated using the Finnish Vaccination Register. In summary, annual influenza coverage results from 2007 onwards have been derived from three different data sources and the population base varies by season. However, the coverage estimates from the same data source (e.g. from 2012 onwards) are assumed comparable for evaluating trends over several influenza seasons.

Papilloma virus (HPV) vaccination coverage from 2013 onwards

HPV-vaccinations was introduced in the National Immunization Program in November 2013. Register-based HPV-reports have been published since April 2014. At the moment, the quality of the records can be approximated based on the validation study of childhood vaccinations (see above).

Timeliness and punctuality

The submission pace or reporting delay of records delivered using Avohilmo varies from health care center to health care center. In 2015, 87% of the health care centers submitted their records with an average reporting delay of less than seven days (situation as of 04/2016). The database is updated every night.