EMERGENCY RESPONSE SERVICES: Mortality – The Case of Cardiac Arrest 5

In order to provide further evidence about the robustness of the results, we consider the final specification, which includes only counties who changed during the year. The county which changed from Basic to Enhanced 911 saw a 14% decrease in its TIME_TO_SCENE, while the counties which changed from No 911 to Basic saw a decrease which is not statistically significant. The weaker results about changes from No 911 to Basic may reflect the fact that moving to a centralized 911 system without automated address-finding technology may have ambiguous results, especially in the short run. At a minimum, the system may require some leaming-by-doing before call-takers in a new 911 system are able to gather correct address information for a large area.

We also find that the emergency response system appears to respond to the severity of the patient’s symptoms: patients with a higher Glasgow score have somewhat higher TIME_TO_SCENE, although this result is not statistically significant. We do not, however, see differences in the TIME_TO_SCENE for different categories of insurance (Medicare is the comparison group) or for different ages, with the exceptions that Medicaid patients and younger patients tend to have faster response times.

County-level demographics are also correlated with TIME_TO_SCENE. When the largest counties are included, counties with large populations and high densities have faster response times; once the large counties are excluded, the results are reversed. In all cases, higher income is associated with faster response times.

DEPENDENT VARIABLE = L TIME_AT_SCENE
BASE TIME EXCLUDING ONLY COUNTIES
REGRESSION CONTROLS 4LARGEST WITH 911 LEVEL
(OLS) (OLS) COUNTIES CHANGES
(OLS) (FIXED EFFECTS)
TIME CONTROLS
L TIMEJTO_S CENE -0.12301 -0.1574 -0.25261
(0.00611) (0.00777) (0.02734)
L TIME_TO_HOSP -0.04943 -0.05542 -0.10568
(0.00529) (0.00669) (0.02515)
911 LEVEL
N0 911 -0.18644 -0,16716 -0.18038 -0.0343
(0.01522) (0.01501) (0.01665) (0.07951)
BASIC 911 -0.09558 -0.08014 -0.07043 0.11722
(0.01202) (0.01186) (0.01271) (0.0485)
PATIENT CHARACTERIS1 ncs
MALE -0.03912 -0.03228 -0.03712 -0.03224
(0.00744) (0.00733) (0.00932) (0.03216)
AGE 0.00536 0.00691 0.00397 0.00001
(0.00184) (0.00181) (0.00254) (0.00857)
AGE.SQUARED -0.00001 -0.00003 -0.00001 0.00002
(0.00001) (0.00001) (0.00002) (0.00007)
CARDIAC ARREST 0.02266 -0.00199 0.06571 0.13014
(0.01602) (0.01581) (0.0359) (0.13162)
DEFIB RILLATE 0.03347 0.03949 0.0247 0.06299
(0.00758) r-r-oо

о

Г0.00959) (0.03616)
GLASGOW TRAUMA SCORE (15=Least Severe; 3=Most Severe; 0=Unknown)
GLASGOW 0 2.17322 2.45456 2.54128
(0.1807) (0.1783) (0.22585)
GLASGOW 3 2.45973 2.74229 2.75759 0.17614
(0.18206) (0.17964) (0.22742) (0.19364)
GLASGOW 4-9 2.30703 2.58851 2.60049 -0.26967
(0.18212) (0.1797) (0.22805) (0.20899)
GLASGOW 10-12 2.20594 2.5016 2.55148 -0.03566
(0.18177) (0.17938) (0.22728) (0.19633)
GLASGOW 13-14 2.19037 2.48132 2.54956 -0.15936
Г0.18029) (0.17792) (0.2255) (0.1574)
GLASGOW 15 2.17405 2.47202 2.51962 -0.10864
(0.17975) (0.17741) (0.22466) (0.13614)
INSURANCE STATUS (EXCLUDED CATEGORY = MEDICARE)
MEDICAID 0.00908(0.01854) -0.00695(0.01827) 0.00054(0.02357) -0.00165(0.07775)
PRIVATE -0.02494(0.01171) -0.023(0.01153) -0.0457(0.01489) -0.10748(0.04931)
SELF PAY -0.00607(0.03419) -0.00293(0.03368) -0.01054(0.03861) -0.51433(0.22833)
OTHER 0.01131(0.02051) 0.01001 (0.0202) -0.00877(0.02758) -0.10236(0.07877)
COUNTY HOSPITAL INFRASTRUCTURE
CERT. TRAUM CENTER -0.08111(0.01321) -0.05414(0.01305) -0.11849(0.01979)
LHOSP PER SQ. MILE 0.12706(0.01066) 0.10232(0.01054) 0.08358(0.01423)
L COUNTY CARDIAC PATIENTS -0.10168(0.01055) -0.10644(0.0104) -0.104(0.01187)
COUNTY DEMOGRAPHICS
L POPULATION 0.43383(0.02746) 0.42261(0.02706) 0.37463(0.02935)
DENSITY -0.07099(0.00493) -0.06761(0.00486) -0.03212(0.04967)
L INCOME PER CAP -0.19859(0.03314) -0.20712(0.03275) -0.11367(0.04764)
VCRIMERATE 35.18246(3.80783) 36.32172(3.75116) 52.4423(4.71698)
L POLICE EXP -0.17784(0.02007) -0.16911(0.01978) -0.18477(0.02205)
L HEALTH EXP -0.01733(0.00378) -0.01811(0.00373) -0.01583(0.00406)
HOSPITAL CHARACTERISTICS
URGENT CARE CENTER -0.03078(0.0102) -0.0354(0.01005) -0.07251 (0.014 )
CATHLAB 0.0825(0.01062) 0.08175(0.01047) 0.09871(0.01386) 0.13936(0.05607)
OPENHEART FAC -0.07375(0.01114) -0.0707(0.01098) -0.11318(0.0188)
TRAUMA CENTER LEVEL 0.0035(0.01059) -0.0072(0.01044) -0.00458(0.01889)
EMERGENCY ROOM VOLUME 0.0008(0.00038) 0.00087(0.00038) 0.00174(0.00054) 0.13112(0.05946)
HOSPITAL DOCTORS -0.00019(0.00023) -0.00017(0.00022) -0.00113(0.00052) -0.12205(0.05052)
HOSPITAL RESIDENTS 0.00027(0.00006) 0.00033(0.00006) 0.00167(0.00037) -0.02503(0.00954)
CONSTANT 3.20254(0.33705)
OBSERVATIONS 24664.0000 24664.0000 16477.0000 1635.0000
R-SQUARED 0.7040 0.7170 0.7170

Table 7 analyzes the determinants of TIME_AT_SCENE, following the same set of specifications as in Table 6. TIME_AT_SCENE is negatively related to both TIME_TO_SCENE and TIME_TO_HOSPITAL. It is increasing in the level of 911, and it is longer for more severe patients. TIME_AT_SCENE is also longer for highly populated counties and especially in those with high crime rates, while it is lower in densely populated, high-income, and high-expenditure counties. A full interpretation of these results would require further investigation into the services provided by ambulances, and how they vary with TIME_AT_SCENE. For example, if longer TIME_AT_SCENE is positively correlated with more services, we can interpret the results as saying that more ambulance services are provided in counties with higher levels of 911. This interpretation seems inconsistent with the results on income and expenditures, however.