Nermina Ferizovic | Health Economics Expert

Health economics relies upon having the most appropriate data with which to make resource allocation decisions. The increasing importance of health economics in many nations and the corresponding development of national bodies who assess and scrutinise health interventions for their cost-effectiveness has led to the increased need for appropriate data.

Currently, the primary source of data in cost-effectiveness analyses are clinical trials. Clinical trials are aimed at assessing the safety and the efficacy of health interventions and it is now increasingly common that cost-effectiveness is conducted in parallel to clinical trials or at least that trial data is the basis for cost-effectiveness analyses. However, this raises the question as to whether clinical trials are the correct avenue on which to base cost-effectiveness of a health intervention. The reasons for this include the short-term trial data which cannot lend itself to long-term outcomes and the uncertainty surrounding data extrapolations, the inability of clinical trials to capture all relevant resource use and the highly controlled trial environment which does not reflect the usual standard of care that may be seen when the patient is receiving the evaluated treatment. An additional key reason is that the efficacy determined in clinical trials does not lend itself to the determination of effectiveness- a key input to cost-effectiveness analyses. 

However, many of these questions are now being re-visited in light of the increasing availability of alternative data sources. This includes the availability of insurance claims data, administrative data and medical records and, indeed, digital health applications or the more traditional registry. Historically, these have been known as observational data but are increasingly being described using the umbrella term of real-world data, a term widely used to describe data derived from non-trial settings. 

Some of these data sources might better lend themselves to cost-effectiveness analyses. For example, insurance claims data may allow analysts to determine the true macro-cost resulting from a particular health intervention or medical records may be used in micro-costing analyses. Thus, these data sources may be more reflective of true cost burdens, something which is unachievable in even the best designed clinical trials. Additionally, digital health applications or registries may provide long-term data, specifically that regarding intervention effectiveness, thus reducing data uncertainties. Finally, administrative data sources may provide data on appointments and diagnoses and may lend themselves as useful control data, particularly in orphan diseases. There are further data sources which may contribute vital information to cost-effectiveness analysis such as pharmacy data. The majority of these data sources are currently largely ignored in the cost-effectiveness of health interventions. 

The existence of alternative data sources and the many disadvantages posed from an over-reliance on clinical trial data in cost-effectiveness analysis, now raises the question as to whether the health economics profession should encourage the use of these alternative data sources? Or their development into more structured data sources, as we see in clinical trial data? Indeed, why not encourage the development of alternative data sources that better meet the requirements of health economics? This would allow resource allocation decisions in health to be based on appropriate data sources, something from which we all benefit.