Open Collections

Ascertainment of keratinocyte carcinoma and atopic dermatitis using health administrative data in British Columbia, Canada

University of British Columbia

Background: Disease surveillance is essential for accurate cost estimates, monitoring and evaluation of risk factors and comorbidities of the diseases. Keratinocyte carcinoma (KC) and atopic dermatitis (AD) are frequently occurring skin diseases associated with reduced quality of life and significant burdens. However, KCs are excluded from cancer registries, and AD are reliant on survey sampling. Literature publications have limited claims-based methodologies for KC and AD ascertainment. Thus, there is a lack of any population-based method of ascertainment that is complete, efficient and generalizable for these conditions. Objectives: To propose and validate the use of health insurance claims and prescription records in ascertaining two independent skin conditions. Methods: This retrospective study involved reviewing medical charts and health insurance claims data of patients receiving care by family practitioners and specialists in metropolitan Vancouver from 2010 to 2018. Information on patient demographics, histopathology, atopy, physician diagnoses, claim codes and prescriptions associated with KC and AD were collected respectively in accordance with data abstraction forms. Algorithmic models were developed with predictor variables based on combinations of demographics, claim codes (diagnosis and therapy), and prescriptions. Results: Algorithms for KC ascertainment included the diagnostic code 173 (ICD-9: “other malignant neoplasm of skin” with service codes for procedural treatment and prescriptions for imiquimod, 5-fluorouracil, or vismodegib. High sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) was achieved with values of 86.6% (confidence interval: 79.2-94.0%), 97.4% (CI: 96.2-98.6%), 78.9% (CI: 70.5-87.3%), 98.5% (CI: 97.6-99.4%) respectively. For AD ascertainment, the best-performing algorithms included predictors such as iv age, diagnostic codes 691 or 692 (ICD-9: Other atopic dermatitis and related conditions; Contact dermatitis and other eczema), service codes for ultraviolet therapies, with either prescriptions for topical steroids, topical immunomodulators, or systemic immunosuppressive therapy. High sensitivity and PPV of 94.8% (CI: 93.2-96.3%) and 87.9% (CI: 85.8-90.1%), respectively, was achieved. Conclusion: Our health insurance claims and prescription records algorithms ascertained KC and AD with high predictive accuracy. These algorithmic definitions can be utilized in population databases for evaluating epidemiologic trends and subsequently plan for effective public health initiatives and resource allocations.

Text

2019-12-09

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/72603

Experimental Medicine

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/