(P-341) Methodological Considerations Related to the Use of Primary and Secondary Care Data in Identifying a Cohort of Moderate to Severe Psoriasis Patients in the UK

Background: Psoriasis as a clinical entity is well recorded in EHRs ranging from 82% to 90%. However, disease severity is not well captured but may be inferred from treatment (topical vs systemic).

Objectives: In this study, we aim to explore the ability to ascertain psoriasis severity using a primary care data source IMRD representative of general practices in the UK and a secondary care data source HTI which contains details of hospital admissions and outpatient care in NHS hospitals in England.

Methods: The study period was from January 2015 to October 2019 in HTI and to January 2020 in IMRD. Two case definitions were applied: a narrow case definition, defined as at least one READ code or ICD-10 code indicating psoriasis and at least one prescription record of 2nd line of therapy (LoT) (phototherapy, methotrexate, ciclosporin and retinoids) or 3rd LoT (apremilast, dimethyl fumarate, TNFI and Non-TNFI biologics); and a broad case definition which included the narrow case definition plus at least one auxiliary information, such as percent of body surface area affected (>10%), severity index score, physician global assessment, simple measure for assessing psoriasis activity, dermatology life quality index (DLQI) and referrals to rheumatologists and dermatologist. Descriptive analysis were used to estimate the proportions of patients fulfilling each case definition, by data source as well the proportion of patients diagnosed based on broad case definition, who were misclassified when a narrow case definition was used.

Results: Within the study, 123,551 and 139,609 patients were identified with psoriasis in IMRD and HTI respectively. Of these, 2.7% (3,318) and 8.7% (12,179) patients were identified as moderate to severe psoriasis in IMRD and HTI using the narrow case definition. The broad case definition increased the population to 64.0% (78,951) and 68.2% (95,219) in IMRD and HTI. The extent of severity status misclassification was 64% in IMRD and 73% in HTI. Of the broad case definition criteria, only referrals to rheumatology and dermatology were available in HTI, while referrals and DLQI were available in IMRD.

Conclusions: This study highlights some of the challenges in defining psoriasis disease severity using primary care and secondary care data sources in the UK. As expected more patients were observed in HTI vs. IMRD. Auxiliary information increased both moderate to severe populations significantly suggesting high potential (>60%) for disease misclassification if case definition is too narrow. These results can be used to provide recommendations for future studies on moderate to severe psoriasis populations.