Table 3.

Evaluation of the Specialist Cohort Event Monitoring (SCEM) design: strengths and limitations

Limitations	PASS requirement affected	SCEM design features	Strengths	Distinguishing attributes for PASS
Selection bias from incomplete coverage, sampling process and non-response	Representativeness	Open label in secondary care setting Minimal exclusion criteria	Study off-label use Identify at-risk groups	Unique system that systematically collects good-quality information at a national scale on hospital initiation, as well as early complications and cessation, of treatment Responsive design; study fieldwork activities maintain and enhance engagement of specialist HCPs
Depletion of susceptibles—loss of sub-groups of high risk patients during follow-up	New-user/inception cohort	Targeted data collection to characterize the new-user study population Exposure data collected from dispensed prescription charts	Mitigates confounding by disease severity and immortal time bias, which can result in underestimation of risk of early-onset events	Indication for use reported from medical records and exposure from prescription chart; no assumptions made on surrogate markers Consent obtained to contact the GP for censored patients
Misclassification and underreporting	Safety outcomes	Use of well-defined case definitions based on acceptable agreed clinical standards	Expertise of the specialist provides more accurate information (through original medical records review) than that provided from hospital episode/ discharge statistics or free-text review of hospital letters sent to GPs	Can address specific regulatory questions in the context of the RMP for the product
Confounding and effect modification	Adjustment for measured confounders in statistical analysis	Relevant data on prognostic and potential confounding factors can be collected as recorded in medical charts before treatment was started Rigorous data quality review of returned questionnaires	Deeper level of granularity of relevant variables from secondary care medical records	Improved quality of information on important confounding variables relevant to the targeted outcomes and population of the study GP records may suffer from significant proportions of missing data that may be relevant for the study but not of clinical relevance

GP primary care physician, HCP health care professional, PASS post-authorization safety studies, RMP risk management plan