(A). Define the reasons for establishing the register and the major scientific questions

(B). Consider sample sizes and duration needed to accrue sufficient information to address the main study questions using power calculations

(A). Define eligibility criteria for registration

(B). Consider whether capture is targeting new users or prevalent users

(C). Define ‘exposed’ population and the intended comparator. Establishment of a register with a comparator is strongly encouraged. Comparison cohort should be as similar as possible to exposed cohort, aside from exposure to drug

(D). Consider whether targeted patients will be representative of the disease in the general population (external validity) when selecting recruitment centres, and whether treated and comparator subjects should/need come from the same centres

(A). Stipulate a necessary and achievable minimum core set of variables to be collected for all individuals

(B). Clearly define each variable to minimise false interpretation. Aid data entry by examples, pop-up boxes or other means; provide predefined possible range of answers. Use coded data in favour of free-text strings where possible

(C). When capturing composite scores (eg, DAS28), collect core components if possible

(D). Collect start and stop dates of treatments of interest, including categorised reasons for discontinuation

(E). For treatments given infrequently (eg, infusions), collect the dates of each treatment cycle administration

(F). Collect dose or other means to discriminate between high and low exposure

(G). Define and collect core data on comorbidities, other medications and the treated conditions in order to allow for statistical adjustment for channelling bias and for assessments of predictors of risk

(A). Make every effort to collect outcomes in a complete, robust and transparent manner. Consider overlapping sources for information on outcome

(B). Consider methods to validate outcomes of interest

(C). Take care to accurately capture dates of events

(D). Safety outcomes should be considered beyond drug discontinuation and cover any minimum observation time

(E). Use accepted systems that allow for standardised coding of adverse events

(F). Ensure data on relevant confounders are captured so that channelling bias and confounding can be estimated (and managed in statistical analyses)

(A). Methods of follow-up should be as similar as possible between exposed and comparison cohorts, and for all subjects in each cohort

(B). Use predefined follow-up time points of data collection weighed against practicality

(C). Consider strategies to minimise losses to follow-up and maximise completeness of the data items to be entered. Consider providing clinical benefits to maximise ongoing participation: clinician-specific or centre-specific audit data, automated transfer of electronic medical data, newsletters, financial reward for contribution

(A). Define who is to provide and enter data (patient, doctor, health professional, other)

(B). Define and test how data is to be captured and entered (electronic, paper, other)

(C). Define methods to assess completeness of patient inclusion (population coverage)

(D). Define methods, processes and standards of data monitoring (eg, completeness, plausibility, quality assurance at different levels of the research process)

(E). Define a strategy to allow patients/centres to exit from further registration (NB: losses to follow-up may introduce bias if related to exposure and outcome)

(A). Ensure security of patient-identifiable information and compliance with local legislation

(B). Database architecture needs planning to ensure possibility for output on several levels (patient/site/treatment)

(C). Provide easy access to clear description of study design, variable definitions and frequently asked questions (FAQs)

(D). Define a priori the ownership and access of data

(E). Define the role of any external sponsor including data access, use and presentation

(F). Define the relationships between researchers and participating physicians

(G). Consider whether any regulatory requirements need be taken into account