Ten dermatologic interventions were identified as illustrative, to be presented alongside the risk of automobile injury and death. The LTR of automobile injury was 32%, defined as 1.0 RU. The LTR of automobile death was 0.89% (1/36 RU).
Two events had LTRs roughly similar to automobile injury: development of a subsequent basal cell carcinoma within 3 years (1.4 RU) and development of a squamous cell carcinoma (SCC) secondary to indoor tanning (1.6 RU). Development of SCC following organ transplantation (34 RU) was considerably more likely than automobile injury. All other identified events had lower RUs than automobile injury (Table). Three events with small RUs included tuberculosis development with a tumor necrosis factor α inhibitor (1/32 RU), Crohn disease development with isotretinoin (1/41 RU), and blindness following facial hyaluronic acid injection (1/80 RU). The LTR of death by dog attack (1/42,436 RU) and death by lightning strike (1/36,542 RU) also had small RUs.
The unit comparators from the Table were adapted into graphic form to depict risk relative to the risk of automobile injury (Figure).
Numerous interventions in dermatology offer much less risk of an adverse event than the LTR of automobile injury. However, this concept of risk includes only the likelihood of development of an event, not the severity of the measured event, as our numerical and visual tool objectively captures the related risks using an RU comparator. Such use of a standardized RU demonstrates the essence of risk; “zero risk” does not exist, and each intervention or treatment, albeit how small, must be justified in concordance with other types of risk, such as the automobile.
The development of adverse events secondary to dermatologic intervention or therapy, for which monitoring visits are utilized, were used as important comparators to the risk of automobile injury. The continuous practice of monitoring visits may increase patient’s fears regarding possible adverse events secondary to therapy. Hydroxychloroquine retinal toxicity (1/16 RU) and psoriatic arthritis development following severe psoriasis (1/3.9 RU) were less likely to occur than automobile injury. The development of abnormal blood counts or blood tests secondary to therapy or intervention could not be formatted into an RU. The use of equation (1) for the calculation of LTRadverse eventholds true only when the absolute risk of developing the adverse event in the general population—in this case, abnormal blood counts or blood tests—is low.7
Although the unit comparator allows for the comparison of different dermatologic risk, a limitation of the RU model and its visual tool are a dependence on RR, a value that changes following publication of new studies. A solution was the use of a single pooled estimate to represent the upper-limit CIs of LTR. This practice overestimates risk. As with RR, new automobile injury rates are published annually.10 In the last 5 years, the LTR of automobile injury has stayed relatively constant: between 32% and 33%.4 Although the RU calculations and Figure included a wide variety of interventions in dermatology, select clinical situations were not included. It is beyond the scope of this article to systematically review all risk in dermatology but rather introduce the concept of the RU founded on automobile-associated risks. With the introduction of a methodical framework, the reader is invited to calculate RUs pertinent to their clinical interests.
Any intervention or treatment in dermatology is accompanied by risk. The use of a unit comparator using an easily relatable event—the LTR of automobile injury—allows the patient to easily compare risk and internally justify the practice of monitoring visits. Inclusion of a visual tool, such as the Figure, might alleviate many irrational fears that accompany some of the highly effective treatments and interventions used in dermatology and thus lead to better patient outcomes.
We thank Taranjeet Singh, MS (Dunn, North Carolina), for her comments on an earlier version of the manuscript.