The RAND health insurance experiment (HIE) demonstrated that increasing coinsurance rates decreases medical care utilization. The HIE also found that health outcomes did not vary between individuals with high, low and zero coinsurance rates.
A working paper by Chandra, Gruber and McKnight (“Patient Cost Sharing…“) re-examines whether or not this is the case using a more current dataset specifically focused on the elderly. The medical utilization data the authors use is for of all CalPERS retirees between January 2000 and September 2003. Almost all of the retirees are covered by Medicare, but since Medicare typically has a 20% coinsurance rate, CalPERS provides supplemental insurance to their retirees. The authors conduct a difference in difference estimation comparing copayment changes from the CalPERS decision to raise PPO copayment rates in February 2001 and then to raise HMO copayment rates beginning in January 2002.
The authors find that physician office visits and prescription drug utilization are very price sensitive. For office visits, the estimated price elasticity is between -1.38 and -1.90 and for pharmaceuticals the price elasticity is between -0.20 and -1.4. These findings are surprising since it is typically assumed that the demand for medical care is inelastic.
The authors also found that increased cost sharing led to a slight increase in hospitalizations. However, when the subpopulation of individuals with chronic health conditions is examined, large increases in hospitalization rates are found. This means that individuals with chronic health conditions forego office visits and drug purchases due to the increase in price, but this decision will worsen their health and thus increase the chance they are hospitalized.
Why would an insurance company want to increase the number of expensive hospitalizations? It turns out that the CalPERS insurance plans pay for the ‘first-dollar’ of office visit and pharmaceutical costs. Thus, by increasing copayments, office visits and drug use decrease. Since Medicare pays for the ‘last dollar’ of medical costs (i.e.: Medicare pays for expensive hospitalizations and surgical procedures), the CalPERS plans do not incur the cost of the increased hospitalizations. To summarize, CalPERS receives the majority of the cost savings from increased copayments whereas Medicare bears the cost of the increased hospitalizations when office visit and pharmaceutical demand decreases.
This papers shows that it is always important to take a more global, more systematic view whenever a researcher is investigating the medical field.
- Candra, Amitabh; Gruber, Jonathan; McKnight, Robin (2007) “Patient Cost-Sharing, Hospitalization Offsets, and the Design of Optimal Health Insurance for the Elderly,” NBER Working Paper #12972.