In “Machines that Go ‘Ping’,” Peter Willemé and Michel Dumont explain that there are two types of new FDA drug applications and two, somewhat analogous, types of FDA submissions for medical devices: “radically” and “incrementally” innovative, basically. According to their work, these two types have different effects on health spending growth: radical stuff is cost increasing, incremental stuff is cost savings, at least on average.
I will admit that their type of work is outside my sub-sub-sub-sub-area of expertise so I’m not vouching for its methodological rigor. Someone else tell me about that, if you know.
Anyway, here’s the deal:
The FDA distinguishes new drug applications (NDAs) from NMEs [new molecular entities], the latter being new chemical structures that have never before been used in clinical practice.
NDAs are generally more “incremental” and NMEs are potentially more “radical,” the authors say. Meanwhile, after parsing the Medical Device Amendments of 1976 to the FD&C Act (yeah, my eyes glaze over too), the authors tell us,
All Class III devices require pre-market approval (PMA) by the FDA to establish safety and effectiveness. For all other medical devices a 510(k) pre-market [notifications (PMNs)] should be made in which manufacturers have to demonstrate that their device is at least as safe and effective (i.e., substantially equivalent) to a legally marketed device.
PMNs are generally more “incremental” and PMAs are potentially more “radical,” the authors say. After some estimation work with year-country (OECD) data, and controlling for GDP, proportion of health care publicly financed, BMI, and age composition, the authors conclude,
Approved new drugs (NDA) and devices (PMN) have negative effects, whereas approved pharmaceutical products based on new molecules (NME) and new ‘Class III’ devices (PMA) have positive effects [on health care spending].
The results are not small:
[T]he net positive contribution of medical innovation to the historical growth of health spending is largely due to the introduction of new Class III devices (PMA, +42%) and molecules (NME, +14%), whereas incremental innovation, both in drugs (NDA, -4%) and devices (PMN, -9%), have a negative net effect on aggregate spending.
These results suggest that ‘radically’ innovating products are cost-increasing, whereas the net effect of ‘incrementally’ innovating products appears to be cost-saving. A plausible explanation is that radical innovation leads to ‘treatment expansion’, whereas incremental innovation leads to ‘treatment substitution’. The cost-increasing effect of radical medical progress confirms the consensus view about technology as a ‘major driver’ of health expenditures. However, the results for incremental medical innovation suggest that the net effect of some new drugs and devices may be negative because they lower the use of other (costly) medical interventions.
The net cost-saving effect of some drugs and devices on health expenditures has important implications for policymakers responsible for designing and implementing cost containment programs: if the costs of reimbursing new products are evaluated separately from other medical expenses, there is a danger that the reimbursement decisions will be overly restrictive, thereby limiting the use of technology that might ultimately save costs. Our results suggest that cost containment policies should carefully take into account the potential of new drugs and devices to substitute for other medical interventions.
In conclusion, the title of their paper is great, etc.