Advanced Course in Referral Bias, Part II

L ast month’s editorial opened the discussion of how our particular experiences can mistakenly shape our understanding of the practice of medicine. When systems are structured so that we only see one of many possible outcomes, it’s very easy to succumb to referral bias and believe that the outcome that you see is the only possible one that could occur. Let’s consider some examples. Pustular Psoriasis with Systemic Steroids I have a friend, a physician like me, who specializes in psoriasis treatment. Let’s call him Steve. Steve is a full-time dermatologist. He teaches his residents never to prescribe systemic steroids for psoriasis. And he never prescribes them himself. Steve runs a referral center for patients who have severe psoriasis, and he’s responsible for admitting and caring for all of the patients in his area who get a pustular flare after receiving systemic steroids for their psoriasis. Guess how many times Steve sees pustular flares when psoriasis patients receive a systemic steroid? If you answered 100% of the time, then you understand referral bias. Pustular flares in psoriasis may only occur in only 1 in 10,000 exposures to systemic steroids, but in this doctor’s experience, it seems like flares happen all the time. Surgeons’ Impressions of Dermatologists As dermatologists, we don’t spend much time in the operating room with other surgeons. Chances are, they only see our work in very rare instances. But would it surprise you to learn that some otolaryngologists believe that Mohs surgery causes squamous cell carcinomas (SCCs) to metastasize? If a dermatologist does micrographic surgery on 1,000 SCCs, the majority will be cured. Among the remainder, there may be a few that will have metastasized prior to Mohs surgery. Those cases, and only those cases, are the ones that will be seen by the ENT doctor. Judging Our Colleagues’ Care We have to be especially cognizant of referral bias when we consider how we view our colleagues. I think it’s safe to say (with a bit of hyperbole) that no dermatologist has ever seen a patient with skin disease that was well managed by a primary care provider. Why? The patients who were effectively treated by their primary care doctors — most of them — had no reason to visit a dermatologist. Be mindful; this bias doesn’t just apply to how we view primary care physicians. It extends to physician assistants in other doctors’ offices and even our own peers. The Really Scary Part Now that you have a better appreciation for referral bias, consider other ways it may affect your perception. Do you get a representative sample of news from around the world, or is it likely that what you get to see and hear is biased toward a particular view? Are you confident you know what’s really happening in Iraq? Steven R. Feldman, M.D., Ph.D. Chief Medical Editor

L ast month’s editorial opened the discussion of how our particular experiences can mistakenly shape our understanding of the practice of medicine. When systems are structured so that we only see one of many possible outcomes, it’s very easy to succumb to referral bias and believe that the outcome that you see is the only possible one that could occur. Let’s consider some examples. Pustular Psoriasis with Systemic Steroids I have a friend, a physician like me, who specializes in psoriasis treatment. Let’s call him Steve. Steve is a full-time dermatologist. He teaches his residents never to prescribe systemic steroids for psoriasis. And he never prescribes them himself. Steve runs a referral center for patients who have severe psoriasis, and he’s responsible for admitting and caring for all of the patients in his area who get a pustular flare after receiving systemic steroids for their psoriasis. Guess how many times Steve sees pustular flares when psoriasis patients receive a systemic steroid? If you answered 100% of the time, then you understand referral bias. Pustular flares in psoriasis may only occur in only 1 in 10,000 exposures to systemic steroids, but in this doctor’s experience, it seems like flares happen all the time. Surgeons’ Impressions of Dermatologists As dermatologists, we don’t spend much time in the operating room with other surgeons. Chances are, they only see our work in very rare instances. But would it surprise you to learn that some otolaryngologists believe that Mohs surgery causes squamous cell carcinomas (SCCs) to metastasize? If a dermatologist does micrographic surgery on 1,000 SCCs, the majority will be cured. Among the remainder, there may be a few that will have metastasized prior to Mohs surgery. Those cases, and only those cases, are the ones that will be seen by the ENT doctor. Judging Our Colleagues’ Care We have to be especially cognizant of referral bias when we consider how we view our colleagues. I think it’s safe to say (with a bit of hyperbole) that no dermatologist has ever seen a patient with skin disease that was well managed by a primary care provider. Why? The patients who were effectively treated by their primary care doctors — most of them — had no reason to visit a dermatologist. Be mindful; this bias doesn’t just apply to how we view primary care physicians. It extends to physician assistants in other doctors’ offices and even our own peers. The Really Scary Part Now that you have a better appreciation for referral bias, consider other ways it may affect your perception. Do you get a representative sample of news from around the world, or is it likely that what you get to see and hear is biased toward a particular view? Are you confident you know what’s really happening in Iraq? Steven R. Feldman, M.D., Ph.D. Chief Medical Editor