As I’ve come up to speed on healthcare advertising, I’ve been surprised by how often prevalence (AKA audience quality) is used to measure whether an audience is the right one without discussing the nuance of this measurement. Of course, brands ultimately care about script lift, but there’s a lot of lag in script lift measurement, and prevalence is currently used as the gold standard. While we might talk about lookback periods for these audiences, most of the discussion of the human populations we advertise to vastly simplifies the very complex and important concepts related to time in epidemiology. We need to talk about this transparently and in detail for each audience in order to understand whether our measurement (and our audience building) is the best it can be for the indication.
In its simplest definition, prevalence measures the proportion of existing cases in the total population. However, this can be at a specific point in time, for a certain period of time, or over people’s lifetime. Which of these we choose matters, and will vary by indication. Furthermore, there are some diseases where treatment isn’t as relevant for prevalent (existing) cases, and we should instead think about incidence and look for incident (new) cases. Incidence is measured as a rate for a unit of time (e.g. 1 per 1,000 people per year), and can help us understand what to expect in terms of new diagnoses. Identifying incident cases in claims data is trickier, as it depends on the absence of diagnoses in a prior time period.
The application of these concepts in advertising audience measurement is complicated. It may make sense to measure chronic diseases that require lifelong treatment with long-term prevalence, but only if later stage treatment-switching or new treatment is high for that condition. Alternatively, a period prevalence is more appropriate for a disease may last a while until the person finds the right treatment and can be cured. Finally, incidence should be used for indications where it matters to treat them right away, and where they are later cured.
For example, for a condition like hypothyroidism, which is chronic but people tend to find treatment that works for their lifetime, advertisement should be targeted toward incident cases. On the other hand, in chronic diseases that need new treatment as they progress to more advanced stages (such as Parkinson’s or Chronic Kidney Disease) or diseases where treatment failure may be common (many autoimmune and mental health conditions), we should target advertising to prevalent cases. For most infectious diseases, advertisement should be targeted at incident cases. In contrast, for many cancers which achieve durable remission with treatment, advertisement should be targeted using period prevalence – so that we’re only targeting people who are likely still on their treatment journey, not those in durable remission.
In the advertising industry, we’re missing an opportunity to positively impact patient care because we are missing the nuance of time, prevalence, and incidence through the complex layers of the system. Communicating transparently and in a detailed and nuanced way between brands, agencies, audience targeting companies, and measurement companies about the specific audience we want to target and how we are going to define and measure it will only serve to improve patient care by increasing conversations between the right patients and their doctors.