Responses to Cost-Sharing by Low-Income Patients

One of the things the woefully mis-named Affordable Care Act did was create insurance exchanges that were supposed to be used by insurers to provide reasonably priced coverage for those who otherwise did not have health insurance.  Subsidies were provided to help lower-income patients buy the health insurance offered on the exchanges and in some cases, plans were required to lower cost-sharing amounts, like copays, for these low-income enrollees.  Research published by the National Bureau of Economic Research explores how patients responded to these changes in cost-sharing amounts.   (NBER Paper)   Most prior research had found, as you would expect, that higher cost-sharing tends to drive utilization lower and less or no cost-sharing tends to increase utilization.  Cost-sharing and changes in cost-sharing obviously may have different impacts on patients with different income levels, and it is likely that poorer patients are more cost-sensitive.  All-payer claims data from Utah for the years 2013-2015 was used for the study.  This allowed tracking of patients across types of coverage and payers.  The researchers were also able to look at hospital discharge and ER use databases for an extended period of time prior to the primary study period.  They were attempting to create estimates of price sensitivity, which indicates how much utilization will fall per unit of cost-sharing change.  And they looked at actual effects on health spending. Utah has low per capita health spending, for the individuals in this study, the monthly average was $384, of which the largest piece, $211, was for outpatient care, and the out-of-pocket component was $33.

In terms of raw amounts, people with high levels of cost-sharing, around 27% of total costs, had the lowest monthly average spending at $330, with an out-of-pocket of $63; while those in plans with low levels of cost-sharing, around 6%, had the highest spending at $399 per month and cost sharing of only $22.  While use of both low-value and high-value care increased as cost-sharing declined, much more low value care was used, two-thirds more, compared to only one-third more high-value care.  The increases in use occurred across all categories of care, including ER.  When adjusted for health status, those patients with the lowest cost-sharing had about 19% more spending than those patients with the highest cost-sharing.  Largely consistent with other studies, the authors found an overall price sensitivity of about -.1, which means that for a given rise in cost, utilization will decline by about 10%.  The sensitivity for inpatient care was close to zero; for outpatient care it was -.13, for ER -.2, and for drugs -.12.  Sicker patients were less cost-sharing sensitive.  The program worked as intended, at least in part.  It significantly reduced out-of-pocket spending for low-income enrollees.  But the reduced sensitivity to price meant that these patients increased spending willy-nilly; that is on both more necessary and less necessary care.  In fact, their total health spending increased by that 19% and a lot of that was for low-value care.  That is a huge, huge increase.  Now, as with most research today, keep in mind the implications if something as looney-tunes as Medicare-for-all, at least the Sanders/Warren version, were enacted, with the complete lack of any cost-sharing.   You would quite reasonably expect a massive uptick in utilization and spending.  So don’t for one second believe the nonsense estimates about the costs for these programs; in fact raise them by a factor of 33% to 50%.