Friday, January 3, 2014

Emergency room and doctor utilization -- impact of the ACA.

Question:  What is the likely impact of an expansion of Medicaid on services rendered at emergency room?

Answer:   A recent Wall Street Journal article reported results of research indicating that the ACA has expanded the utilization of emergency room services. 

http://blogs.marketwatch.com/health-exchange/2014/01/02/study-finds-medicaid-use-doesnt-lower-emergency-room-visits/

A couple of years ago I conducted research on the impact of public insurance on the utilization of doctors and emergency rooms.

Bernstein, David P., Utilization of Doctors and Emergency Rooms: Implications for Health Care Reform (July 31, 2009). Available at SSRN: 

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1442031

I found that public insurance had no impact on doctor utilization and a large impact on emergency room utilization.

This research was never published in an academic journal.   Often I find referee comments to be very helpful.  In this case, I thought the referees at the journals that looked at this paper could have life support turned off.  (I am saying these refs were brain dead.)  


Key results from my paper are presented below:  


There are huge differences in both utilization of doctors and utilization of emergency rooms between individuals with public insurance and the uninsured.

Table One:  Health Care Utilization, Doctors Visits and Emergency Room Visits

Private  full year
Private Part Year
Public Insurance Part or full year
No insurance
Doctor Visits
Median (# visits)
2.0
1.0
2.0
0.0
Mean (# visits)
3.3
1.9
5.0
1.2
Proportion at least one visit (%)
70.9
54.6
71.3
34.7
Proportion at least two visits (%)
51.7
33.7
56.7
19.8
Emergency Room Visits
Median
0
0
0
0
Mean
0.14
0.18
0.38
0.17
Proportion at least one visit
10.8
13.2
25.1
12.1
Proportion at least two visits
2.3
2.7
7.7
3.1

I ran logit and tobit regressions to hold other variables like heath status constant.   The logit equations explained the probability that that a person will see a doctor or go to the emergency room during the year.   The Tobit equation explains the number of doctor and emergency room visits per person.

Doctor Utilization Results:  


Public insurance has no impact on utilization of doctors (either the number of visits per person or the likelihood of going to a doctor's office.

Table Four:  Doctor Utilization Equations

Tobit Equation for number of Doctor Visits
Logit Equation for at least one doctor visit
Logit equation for at least two doctor visits

Coeff.
P-value
Coeff.
P-value
Coeff.
P-value
Intercept
-1.9
<.0001
-0.1
0.0386
-1.1
<.0001
Family Income <100% FPL
-1.3
<.0001
-0.3
<.0001
-0.3
<.0001
100%<=Family Income < 125% FPL
-1.4
<.0001
-0.3
<.0001
-0.3
<.0001
125%<=Family Income< 200% FPL
-1.7
<.0001
-0.5
<.0001
-0.5
<.0001
200%<=Family Income<400% FPL
-0.9
<.0001
-0.2
<.0001
-0.3
<.0001
Health status very good
1.2
<.0001
0.4
<.0001
0.4
<.0001
Health status good
1.9
<.0001
0.6
<.0001
0.6
<.0001
Health status fair
3.6
<.0001
1.0
<.0001
1.0
<.0001
Health status poor
5.6
<.0001
1.2
<.0001
1.3
<.0001
Mental health status very good
-0.2
0.2611
-0.1
0.0630
0.0
0.5033
Mental health status good
-0.3
0.1302
-0.2
0.0008
-0.1
0.1736
Mental health status fair
1.0
<.0001
0.0
0.6225
0.2
0.0047
Mental health status poor
1.3
0.0005
0.0
0.8736
0.2
0.0812
Limited in climbing stairs
1.5
<.0001
0.3
<.0001
0.3
<.0001
Blind
-0.3
0.7692
-0.1
0.8969
0.6
0.1309
Coronary heart disease
1.2
0.0050
0.7
0.0015
0.8
<.0001
Stroke
-0.4
0.3178
0.5
0.0211
0.4
0.0217
Diabetes
2.1
<.0001
1.1
<.0001
0.8
<.0001
Asthma
1.6
<.0001
0.6
<.0001
0.4
<.0001
Emphysema
1.0
0.0706
0.4
0.1644
0.2
0.2943
Joint Pain
1.3
<.0001
0.4
<.0001
0.4
<.0001
Arthritis
2.0
<.0001
0.7
<.0001
0.7
<.0001
Diabetes related kidney problems
3.2
<.0001
0.0
0.9751
0.3
0.4339
Diabetes related eye problems
1.0
0.0439
0.2
0.5329
0.2
0.3174
Married
0.6
<.0001
0.2
<.0001
0.2
<.0001
Age greater than 40
0.0
<.0001
0.0
<.0001
0.0
<.0001
Number of Emergency room visits
1.2
<.0001
0.5
<.0001
0.4
<.0001
Private insurance part year
-2.4
<.0001
-0.7
<.0001
-0.8
<.0001
Public insurance part or full year
0.2
0.2350
-0.2
0.0038
-0.1
0.1648
No insurance the whole year
-4.8
<.0001
-1.5
<.0001
-1.4
<.0001
Scale
6.7

NA

NA



Emergency Room Results:


Public Insurance has a large impact on emergency room utilization both in the number of visits per person and the likelihood of going to the ER.


Table Five:  Emergency Room Use Equations

Tobit Equation for ER  Visits
Logit Equation for at Least one ER Visit
Logit Equation for at least two ER Visits

Coeff.
P-value
Coeff.
P-value
Coeff.
P-value
Intercept
-3.2
<.0001
-2.3
<.0001
-4.2
<.0001
Family Income <100% FPL
0.4
<.0001
0.3
0.0003
0.3
0.0153
100%<=Family Income < 125% FPL
0.2
0.109
0.2
0.1316
0.2
0.3842
125%<=Family Income< 200% FPL
0.1
0.5599
0.0
0.6808
0.2
0.1457
200%<=Family Income<400% FPL
-0.1
0.1446
-0.1
0.1056
0.0
0.7832
Health status very good
0.4
0.0014
0.4
0.0012
0.5
0.0410
Health status good
0.9
<.0001
0.7
<.0001
1.1
<.0001
Health status fair
1.5
<.0001
1.2
<.0001
1.7
<.0001
Health status poor
2.2
<.0001
1.7
<.0001
2.3
<.0001
Mental health status very good
-0.1
0.5443
0.0
0.5557
-0.3
0.0475
Mental health status good
-0.2
0.1128
-0.1
0.1136
-0.4
0.0195
Mental health status fair
-0.1
0.5809
-0.1
0.3991
-0.2
0.3222
Mental health status poor
0.4
0.0232
0.2
0.1495
0.1
0.5273
Limited in climbing stairs
0.2
0.0578
0.2
0.0211
0.1
0.2680
Blind
0.6
0.2551
0.4
0.2785
0.8
0.0783
Coronary heart disease
0.6
0.0035
0.5
0.0014
0.6
0.0022
Stroke
0.0
0.8588
-0.1
0.6308
0.1
0.6008
Diabetes
0.1
0.5167
0.1
0.4412
0.1
0.5371
Asthma
0.6
<.0001
0.4
<.0001
0.5
<.0001
Emphysema
-0.2
0.4648
-0.1
0.4724
-0.4
0.1887
Joint Pain
0.3
<.0001
0.2
<.0001
0.3
0.0038
Arthritis
0.3
0.0032
0.2
0.0042
0.2
0.0521
Diabetes related kidney problems
0.7
0.0222
0.6
0.0089
0.7
0.0113
Diabetes related eye problems
0.0
0.924
-0.1
0.5369
-0.2
0.4999
Married
-0.3
<.0001
-0.2
<.0001
-0.2
0.0295
Age greater than 40
0.0
<.0001
0.0
<.0001
0.0
<.0001
Private insurance part year
0.1
0.4541
0.1
0.474
0.0
0.8766
Public insurance part or full year
0.3
0.0001
0.3
0.0001
0.3
0.0329
No insurance the whole year
-0.3
0.0003
-0.2
0.0003
-0.1
0.3165
Scale
2.5

NA

NA



I am still very sad this research was not published.  

People who like this post should also look at my work on tax expenditures for employer sponsored insurance.

http://dailymathproblem.blogspot.com/2014/01/tax-expenditures-on-employer-sponsored.html

Also, people who are considering purchasing a health insurance policy on state exchanges should examine this post.





No comments:

Post a Comment