Sales-based forecasting focuses on a measure of volume: units, prescriptions or treatment days. Revenue can also be used as the starting point although not recommend because price variance can distort the real underlying demand for a product.
Epi Patient Based
Epidemiological or epi-based forecasting. This is a patient-based approach where the starting point is typically a country’s population or sub-population based on age or gender. Then a disease prevalence rate, a diagnosis rate and a treatment rate are applied. This defines a target population in terms of number of treated patients.
An opportunity-based approach is where patients are split into those who are considered to be stable on a given treatment versus those who can be considered as an opportunity for a treatment start. A patient is considered as an opportunity if they are a naive patient who hasn’t previously received a treatment for the disease being modelled. Or, if they are being switched from one treatment to another. This includes adding on an additional drug to the existing treatment regime.
Patient-flow based forecasting. With this approach, cohorts of patients are followed over time. This can be important in progressive disease such as oncology, where there is a link between the treatment and outcome from one line of therapy to the next. You will need to understand how long a patient is on a treatment; and the likelihood they are to progress to another subsequent line of treatment. Other questions you will need to think about include, what is the time lag between treatments? And whether there are any restrictions applied when the patient is being retreated?