Abstract
The pharmaceutical industry spends billions of dollars every year for marketing its products to US health care providers. This study investigates the association between industry marketing payments and physicians’ prescription in New York and Massachusetts, and examines the effect of the Massachusetts payment restriction policy on this association in comparison with the New York State that has no payment restriction policy.
Three panel data models (fixed effects regression (FE), first difference regression (FD), and first difference with lagged independent variable (LFD)) were used to establish the association accounting for unobserved confounders and reverse causality. The main indicator is the total amount of industry payments for meals, drug samples, consulting fees, etc. (excluding research funding, and ownership). Dependent variables are a) yearly days’ supply of Medicare Part D prescriptions, b) yearly costs of prescribed prescriptions. Secular time trends, as well as differences between the two states were controlled in all models.
All three panel models showed significant positive association between industry payments and physician prescriptions. Particularly, these models show that in the base year (2014), a 1% increase in the annual industry payments was associated with (0.01%–0.04%) increase in days' supply of medication, and (0.01%–0.05%) increase in prescription costs for NYS providers. Moreover, FE and FD models show that compared to NYS physicians, the association of industry payments with days’ supply and costs of Medicare Part D prescriptions was (0.01%–0.02%) smaller for MA physicians. Because of the smaller sample size and lower power, the LFD model did not detect a significant difference between the NYS and MA providers.
The findings of this study provide additional evidence that industry payments are associated with increases in the volume and cost of Medicare prescriptions. Moreover, these findings provide preliminary evidence that payment restriction policies negatively moderate this association, which calls for further research.
Three panel data models (fixed effects regression (FE), first difference regression (FD), and first difference with lagged independent variable (LFD)) were used to establish the association accounting for unobserved confounders and reverse causality. The main indicator is the total amount of industry payments for meals, drug samples, consulting fees, etc. (excluding research funding, and ownership). Dependent variables are a) yearly days’ supply of Medicare Part D prescriptions, b) yearly costs of prescribed prescriptions. Secular time trends, as well as differences between the two states were controlled in all models.
All three panel models showed significant positive association between industry payments and physician prescriptions. Particularly, these models show that in the base year (2014), a 1% increase in the annual industry payments was associated with (0.01%–0.04%) increase in days' supply of medication, and (0.01%–0.05%) increase in prescription costs for NYS providers. Moreover, FE and FD models show that compared to NYS physicians, the association of industry payments with days’ supply and costs of Medicare Part D prescriptions was (0.01%–0.02%) smaller for MA physicians. Because of the smaller sample size and lower power, the LFD model did not detect a significant difference between the NYS and MA providers.
The findings of this study provide additional evidence that industry payments are associated with increases in the volume and cost of Medicare prescriptions. Moreover, these findings provide preliminary evidence that payment restriction policies negatively moderate this association, which calls for further research.
| Original language | English |
|---|---|
| Article number | 113942 |
| Journal | Social Science & Medicine |
| Volume | 278 |
| Early online date | 20 Apr 2021 |
| DOIs | |
| Publication status | Published - Jun 2021 |
| Externally published | Yes |