Thursday, March 9, 2017

Emergency Room Visits in the First Year of the ACA


Emergency Room Visits in the First Year of the ACA

Question:   What happened to emergency room visits for working-age adults between 2013 and 2014 after the enactment of the ACA?

Short Answer:  The analysis in this post relies on data from the MEPS 2013 and 2014 consolidate databases.  I look at ER visits for people aged 18 to 64, the age group most affected by the opening of state exchanges.  

I found that total ER visits in this age group fell by 2.3%.  

However, the proportion of the populations with two or fewer visits to the ER during the year did not change significantly.

The bulk of the evidence suggests that the ACA did not dramatically change ER use.

Analysis:   The estimates on ER visits were obtained from the total command in STATA on variables ERTOT14 and ERTOT!13 with population weights PERWTF14 and PERWT13F.  Results are below.



Total ER Visits 2013 and 2014
ERTOT14
3.77E+07
8898.05
3.76E+07
3.77E+07
ERTOT13
3.86E+07
8919.727
3.86E+07
3.86E+07
%  Diff.
-2.33%

The analysis on the proportion of people with two or fewer ER visits did not employ population weights.


The raw frequency data on ER visits for 2014 and 2013 is presented below. 


ER Visits 2014
Freq.
Percent
Cum.
0
18,127
86.35
86.35
1
2,041
9.72
96.07
2
533
2.54
98.61
98.61
3
158
0.75
99.37
4
65
0.31
99.68
5
24
0.11
99.79
6
17
0.08
99.87
7
5
0.02
99.9
8
7
0.03
99.93
9
4
0.02
99.95
10
3
0.01
99.96
11
3
0.01
99.98
12
1
0
99.98
13
2
0.01
99.99
14
1
0
100
18
1
0
100
1.35
Total
20,992
100


ER  Visits 2013
Freq.
Percent
Cum.
0
19,365
86.39
86.39
1
2,182
9.73
96.12
2
543
2.42
98.55
98.54
3
197
0.88
99.42
4
67
0.3
99.72
5
27
0.12
99.84
6
11
0.05
99.89
7
9
0.04
99.93
8
3
0.01
99.95
9
2
0.01
99.96
10
4
0.02
99.97
11
1
0
99.98
12
2
0.01
99.99
13
1
0
99.99
15
1
0
100
24
1
0
100
1.44
Total
22,416
100


The ER visit frequency distributions look pretty similar.

I conducted the hypothesis test that the proportion of people with 2 or fewer ER visits in 2013 and 2014 are identical.

The pooled estimator of the proportion of people with 2 or fewer ER visits is 0.9854.


The difference in proportions is 0.0007.


The standard error is SQRT(p*(1-p)*((1/n1)+(1/n2)))

Where n1 and n2 are sample sizes and p is the pooled proportion.

I got SE equal to 0.00116.

The t-statistic of 0.6 is not significant.


Concluding thoughts:  The first year of the ACA and the increase in insurance coverage did not cause much of a change in the use of emergency rooms or in the use of multiple ACA visits.    The 2015 and 2016 data is not yet available.  Perhaps when people find primary care doctors ER use patterns will change but I suspect that ER will remain a vital venue even if insurance becomes universally available.







No comments:

Post a Comment