Study type

Study topic

Disease /health condition

Study type

Non-interventional study

Scope of the study

Disease epidemiology
Method development or testing

Data collection methods

Secondary use of data
Non-interventional study

Non-interventional study design

Cohort
Study drug and medical condition

Medical condition to be studied

Graves' disease
Polyarteritis nodosa
Autoimmune hepatitis
Rheumatoid arthritis
Multiple sclerosis
Erythema nodosum
Systemic lupus erythematosus

Additional medical condition(s)

Hashimoto's thyroiditis, psoriatic arthritis, ulcerative colitis
Population studied

Short description of the study population

Ten disease-specific study cohorts will be created, one for each event.
Persons will be included in the incident disease-specific study cohorts if they have:
- At least 1 day of observation after 1 January 2017.
- The first recorded diagnosis of the disease of interest after 1/1/2017.
Persons will be excluded when:
- No birth year and sex information.
- Less than 365 days look-back at first diagnosis of the disease of interest, for those not born in the study period.
Follow-up will start when study persons have completed 90 days after the disease diagnosis, except for autoimmune hepatitis and erythema nodosum where follow-up will start 30, and 60 days after the diagnosis, respectively. This delay is needed for the first episode to be ended and to be at risk of a flare. Upon each flare the same eligibility delay for a next flare will be applied.
Follow-up for incidence rates will be interrupted when vaccination with any vaccine of interest occurs, from date of vaccination to 90 days later (to ensure we do not include rates in post-vaccination risk windows), and finish at death, disenrollment, or last data extraction. If follow-up is interrupted because of a vaccination, follow-up will be resumed at day 91 after vaccination until follow-up finishes. An interruption of follow-up will also happen upon a flare, with the same eligibility period as after the first diagnosis. If a vaccination happens during the eligibility interruption, the maximum duration of interruption (eligibility and post-vaccination) is used. The occurrence of a flare-up of the corresponding event will not be considered a censoring criterion since a new flare-up can occur after a gap of 90 days. For estimation of cumulative incidence, follow-up will be censored upon the earliest of a vaccination, a flare or end of follow-up.

Age groups

Paediatric Population (< 18 years)
Preterm newborn infants (0 – 27 days)
Term newborn infants (0 – 27 days)
Infants and toddlers (28 days – 23 months)
Children (2 to < 12 years)
Adolescents (12 to < 18 years)
Adult and elderly population (≥18 years)
Adults (18 to < 65 years)
Adults (18 to < 46 years)
Adults (46 to < 65 years)
Elderly (≥ 65 years)
Adults (65 to < 75 years)
Adults (75 to < 85 years)
Adults (85 years and over)
Study design details

Study design

Retrospective multi-database population-based cohort design conducted during the period from January 1st, 2017, till the last data availability, specific for each data source. Study cohorts start with a first diagnosis of
the selected auto-immune diseases during the study period.

Main study objective

- Primary objective:
To estimate the background (independent of vaccination) incidence rates of flares and 6- and 12-month cumulative incidence estimates of flares in patients with Graves’ disease, Hashimoto’s thyroiditis, polyarteritis nodosa, autoimmune hepatitis, rheumatoid arthritis, psoriatic arthritis, multiple sclerosis, erythema nodosum, systemic lupus erythematosus, and ulcerative colitis using European electronic healthcare data sources.

- Secondary objectives:
1. To assess the contribution of different provenances of data as well as the contribution of the different components of the flare definition to the incidence rates of flares.
2. To estimate the background incidence rates and 6 and 12-month cumulative risks of flares of selected auto-immune diseases in subgroups of age (< 18, 18-59, 60 years and older), sex and selected subpopulations.

Setting

The study will use data from 10 secondary electronic health record databases that are population-based in 7 countries in Europe (UK, Spain, Denmark, Finland, Norway, Italy and France). The study will be conducted during the period from January 1st, 2017, till the last data availability, specific for each data source.

Outcomes

This study will provide an estimation of the occurrence of flares of 10 selected pre-existing chronic diseases (Graves' disease, Hashimoto's thyroiditis, polyarteritis nodosa, autoimmune hepatitis, rheumatoid arthritis, psoriatic arthritis, multiple sclerosis, erythema nodosum, systemic lupus erythematosus, and ulcerative colitis) using electronic healthcare record data sources.

Data analysis plan

All analyses will be conducted using R version R-4.03 or higher or STATA:
- Descriptive analyses will be shown through histogram plots with distance from the start of follow-up to the new occurrence of potential flare for visual inspection.
- Incidence rates will be calculated based on the occurrence of a flare requiring a lag-time of 90 days (in order to be eligible) after its incident diagnosis and each flare, except for autoimmune hepatitis and erythema nodosum (30 and 60 days, respectively), and allowing re-occurrence after the last flare-up event tailored in time-period to the condition of interest and the flare-specific identification component. Cumulative incidence of a first flare will be calculated to estimate 6 - and 12-months cumulative risk of flares.
- Component analysis will describe the data diversity of each component of the flare identification algorithm in a data source-specific manner. Each identification component may differ in rates due to data source characteristics and data meanings. For each data sources, the different component algorithms and their data meanings for a same flare identification component will be identified. Flare cases, cumulative incidence of each component algorithm and their combinations in one study year will be measured. Incidence rates with the same “component”/”composite” algorithms will be compared across data sources, facilitating results benchmarking.