Lecture 6- Hw do hospitals respond to price changes? (Dafny, 2001) Flashcards
Who wrote how do hospitals respond to price changes and when?
Dafny in 2001
What is the Prospective Payment System? (PPS)
In 1984 the US government introduced a fixed-payment system where they were paid a set amount for each Medicaid patient in a diagnoses related group (DRG) - regardless of how much their expenses were for them
1988 is when the change in policy occurre
What is the contribution of this paper compared to previous research?
Most previous researchers study the change in hospitals behaviour after PPS was introduced.
Instead, Dafny studies behaviour as the level of fixed payment is changed. He uses a discrete (rather than continuos) change in incentives to upcode in 1988. Helps avoid OVB
He looks at the differences in the level to which reimbursements changed across differing diagnoses related groups (DRG) to see if hospitals respond the most strongly when the change in reimbursement is the highest.
Used a large data-set when compared to previous researchers
What is the main challenge when looking at the effect of price change?
It is hard to know if the price change was exogeneous because of demand and supply
What was the exogenous price change that was used by Dupas?
In 1988 there was a change in policy where within a given disease i.e heart failure patients before would be categorized as either:
- heart failure with complications or over 69 OR heart failure without complications and under 70
Reimbursements for over 69 or complications WAS always higher than below 70 and without complications
However since 1988, age is not a factor so it is:
- heart failure with complications
- heart failure without complications
This policy change lead to a rise in reimbursement level paid for top codes and a FALL in reimbursement for bottom codes
What is upcoding and does it always lead to a change in patient care?
Where a hospital will change a patients code - their Diagnoses Related Group (DRG) to a code that will result in them receiving a higher reimbursement. (Hospitals are given a different fixed payment for different DRGs)
Upcoding doesn’t necessarily result in a change in patient care.
What does Dafny mean by a nominal response to changes in fixed payment level?
When hospitals change their coding practices (i.e. upcoding)
What does Dafny mean by a real response to changes in fixed payment level?
If an increase in the fixed payment of a certain Diagnoses Related Group (DRG) leads to the hospital spending more money on this area for example on equipment for this illness so more people from this area attend the hospital.
What was the main impact on the change from the US government of removing age in the codes (i.e someone being put in the same category as with complications if they’re over 70)
There was a rise in the reimbursement rate for the top coded (those with complications)
And a fall in the reimbursement rate for the bottom coded (those without complications)
Increasing the incentive of hospitals to upcode
Describe the data set
Most of the data was from MedPAR from 1985-1991 3 years before and after the policy change.
It gave information on patients DRG code (indicate if upcoding took place, nominal change in code)
, health care intensity e.g length of stay and surgeries (see if a real change in coding could have taken place) and
patient demographics
What was the main regression
Fractionpt= a0+ +δYear T+ Z Change in Spread p88-87*Post +ept
Post is an indicator variable =1 if the year being looked at is after the policy implementation in 1988
Z shows the impact that the change in spread (bonus in reimbursent given for a top coded patient compared to bottom following the policy change had on the fraction of people admitted to hospital
Z being positive indicates upcoding
What was the impact of the policy change in terms of upcoding based on ownership (private for profit vs not)
Upcoding occurred at the greatest levels in private for profit firms and to a lesser extent in NFP and public.
What was the impact of the policy in terms of real responses (more admissions of patients that were top coded because of an increase in quality of care).
Hospitals didn’t seem to increase their quality of care to attract top coded patients
Even when looking at different types of ownership (private for profit or not) there was no change in quality of care.
Was there a larger degree of top coding for diseases where the reimbursement was larger ?
NO
What was the empirical strategy used by Dafny?
Exploited an exogenous change in prices after a change in policy in 1988
Regression :
Fractionpt = a0+δYeart + Z Change in Spread p88-87*Post + ept
Fraction shows the increase in the proportion of Medicaid patients assigned to the top coded of the DRG pair (upcoding)
Change in Spread shows the change in the amount of reimbursement that hospitals receive for topcode of a DRG pair compared to bottom code after the policy change in 1988.
Post is an indicator variable for if the year is post the reform (1988-1991)
