Use
of
the
Emergency
Room
in
a
Community
Hospital
H. A. WHITE, M.D., M.P.H., and P. A. O'CONNOR, M.D., M.P.H.
AN INCREASING demand for medical
carehas been noted in hospital emergency rooms across the country. Studies in metro-politan area hospitals havedocumentedthe de-mand for and the pattern ofemergency services
insuchhospitals (1-3). In these studies numer-ous problems, such as the overuse of the emer-gency roomfornonurgentmedical problems, the
substitutionof emergency room care for care by a primary physician in practice, and a
dispro-portionate use of the emergency room by lower
income level population groups, have been
identified.
To determinethe pattern ofemergency
serv-icesin a community hospital in a medium-sized
city,astudy of theservicesin theSaginaw Gen-eral Hospital emergency room wascarried out in 1966-67. The pattern of these emergency room services was compared with that of the
Yale-New Haven Hospital emergency room (1). At the time of the study, the house staff at Saginaw General
Hospital
was responsiblefor emergency services, with attending staff
seeingpatients asindicatedorbyarrangement.
Since
July
1967this
emergency roomhas beenstaffed by a corporation of
community
physicians.
Methods of StudyFour periods of 4 weeks each in July and
October
of 1966 and in January andApril
of1967werechosentodocument
possible
seasonal variation. All patientsattending theemergency roomduring these16weekswereincluded in the study. Thetotalnumberofpatients
inthestudywas3,957.
A formwas
devised
forusein datacollection.Part of the information was collected from the
patient or his family
'by
the emergency roomclerk, and Dr. White obtained the remainder
from theemergencyroom records. The
informa-tion was then transferred to standard punch-cards fortabulation.Tabulated information was compared with the 1960 census data. In 1960
the population of the city of Saginaw was 98,265,withthe
county
num-bering 190,752.Aspecialinterviewer obtained additional in-formation on socioeconomic and medical care patterns and verified information recorded by the emergency roovu clerk
during
composite
week inApril 1967.Intervals ofvarying
lengtlh
were selected at random to insure that every hour in the 7 days of the week was sampled bytheinterviewer. Thus, 214patients were in-terviewed. When similar data were obtained by clerk and
interviewer,
there was essentialagreement.
Attributes of Emergency Room Patients
Sew. Males
comprised
52.3 percent of thepatients in the study group as compared with
49.2 percent of the
population
inSaginaw
County
according
tothe 1960census.Race.
Negroes
used the emergency room morethanwhite
persons.Although
only
9.8 per-centof the total courtypopulation
wasNegro,
Dr. White isa
pediatrician
withthe Interior Mission Medical AuxiliariesTraining
School,
losBevue-Plateau,
Nigeria. Dr. O'Connor is associatepro-fessor, maternal and child
health, University
ofMichigan School of Public Hedth. This
study
wassupported
by
Public Health Service grant No. NIH 5SOI FR-5447-06.andNegroes comprised 16.8 percent of the city's
population,26.6 percent of the study group were
Negroes.
Age. When the agedistributionof thestudy group was compared withthe age distribution
in Saginaw County in the 1960 census, it was found that the two distributions were similar. The percentage distributions by age are shown
inthe following table.
Age (years) Study 1960
group1 census1 0-9 -28.4 25. 6 10-19- 19.4 17. 2 20-29 -17.8 11.7 30-39 --- 11.0 13. 4 40-49-8. 7 11.9 50-64-9. 0 12.4 65 andover -5.5 7. 9
1 May not total exactly 100 because of rounding.
Occupation.
Table 1shows thecomparison
of the occupation of thehead of the household in the interviewed group with the occupation ofmales14yearsold and over inSaginawCounty
in 1960. The unemployed made up 7.9 percent
of thestudy group as compared with 5.5 percent in the 1960 county population. Laborers
com-prised 9.3 percent of the study group as
con-trasted with6.4 percentin the county.The cate-gories of "retired," "housewife," and"student"
are not included in the 1960 census data, but wereincluded in the study.
Education.
Theeducational
level of the headsof householdintheinterviewedgroup was compared with the educational levelof persons 25yearsandoverfromthe 1960censusdataforSaginaw
County.
Themedianeducational levelwas 12 yearsin thestudygroupand 10yearsin thecounty.Thepercentage of personswith
high
school or college education was
higher
in the studygroup.CenAus
tract.Residents
ofcensustractswith low socioeconomic levels made greater use ofthe emergency room than residents of other
Saginaw
censustracts.U8ual
8oUrCe ofmwdical
care.Eighty-six
percent ofthe
patients
reportedhaving
aprivate
physician,
and 80.8 percent had had contactwith their
physician
withinthe past year.Only
10.9percenthadused
the
emergencyroomthree times or more in theprevious
year. Persons whoroutinely
received medical care atthehos-pital's
freeclinic comprised
6.6 percentof theemergency room patients. Thirty-sixpercent of
the emergency room patients attempted to con-tact theirphysician beforecomingtothe
emer-gency room, and 73.6 percent of these had talked withtheirphysician.
Eachpatientwasasked how long theproblem that had brought him to the emergency room had existed. Problemsexisting lessthan 1hour were reported by 36 percent of thepatients, and
55 percentwereadmittedto theemergency room
within4 hoursof the onset of theirillness. Ill-nesseslastingmorethan24hourswerereported by24.8percent ofthepatients.
Characteristics of Services
The patients, classified by hospital service,
were adultsurgical 37.4percent, adult medical 19.8 percent,obstetric orgynecological 7.3 per-cent, and pediatric (under 14 years old) 35.1 percent. The physician responsible for 57.9 per-cent ofthepatients was ageneral practitioner,
for17.7 percent a pediatrician, for 6.4 percent an obstetrician-gynecologist, for 6.2 percent a
generalsurgeon,for5.9percent aninternist, and
for 5.9 percentpractitioners of various
surgical
subspecialities.
Sixty-five percent of the patients were seen
by house staff,and 25 percent were seen bytheir private physician. Following examination or
Table 1. Comparison, in percentages, of occupation of head of household in inter-viewed group with that of males 14 years old and over in Saginaw County, Mich.,
1960 census
Interviewed Saginaw Occupation group County
(N= 214) (N=190,752) Unemployed 7.9 5. 5 Professional 5.6 7.8 Managerial_ 7. 0 11. 9 Clerical - -3.7 5. 1 Sales - -3.7 6.5 Craftsman 22. 4 21.3 Operatives --25.8 28. 7 Service- 3.3 4. 3 Household 1.9 . 1 Laborer -- 9.3 6. 4 Retired - -6.5 Housewife --1.9 Student -- .5 Notreported .5 2.4 Total -100. 0 100. 0
Table2. Diagnoses of patients' conditions, each monthof study, bypercentage
1967 1966
Diagnoses Total Number
January April July October
Trauma-45.5 52. 0 53.8 49. 1 50.4 1,994
Fracture ordislocation -6. 1 4. 8 4. 4 3. 3 4. 6 182
Sprain -5.3 5.4 2. 6 4.1 4.2 166
Headinjury- - 2. 6 2. 8 3.5 4.7 3.5 138
Internalinjury --- . 1 . 1 . 1 2
Laceration, abrasion, or contusion- 24. 1 32. 6 38. 0 30. 0 31. 5 1,253
Foreignbody --- 2. 0 1. 2 1. 7 3. 0 2.0 79 Burn -1.9 1. 4 1.5 1. 6 1. 6 62 Poisoning -- 2.5 1. 9 1. 4 1. 6 1. 8 72 Otherornonspecific- . 9 1.8 .7 . 8 1.1 40 Nontrauma- -54.5 48. 0 46. 2 50.9 49. 6 1,963 Infectiousdisease--. 6 . 7 1. 1 .6 . 8 30 Psychiatricdisorder_ -2. 0 1. 6 2. 1 2.9 2. 2 86
Diseaseof thecentral nervoussystem-._ . 6 1. 5 . 7 . 6 . 8 33
Eyedisease ---- . 7 . 8 1.1 .6 . 8 32 Ear disease -1. 4 1. 1 1.6 1.3 1.4 54 Circulatorydisease -3. 4 3. 2 2. 9 2. 7 3. 0 119 Respiratorydisease -8. 9 7. 0 6. 1 9. 3 7. 7 306 Gastrointestinaldisease -7. 2 6. 4 5.4 5. 0 5.9 234 Genitourinarydisease -4. 8 4. 0 3. 2 3. 7 3. 8 152 Pregnancy-2. 9 1. 9 2. 6 2. 9 2. 6 103
Allergy (includingasthma) -1.9 3. 4 2. 6 1. 7 2. 4 95
Endocrinedisease -. 5 . 2 .2 .4 .3 13
Blooddisease - -. 1 .3 . 1 .2 6
Skindisease -2. 6 .9 1.7 1.9 1. 8 71
Boneand jointdisease -1. 2 1. 4 1.3 1. 3 1.3 51
Congenital malformations ordiseases of
infancy- -_---__--- .5 .1 .1 .2 6
Symptomsofsenility, ill-defined -9. 8 10. 1 6. 5 9. 6 & 8 349
Neoplasms --- 1. 4 . 3 .6 . 4 . 7 26
Nodiagnosedillness ---- 2. 0 1. 5 3. 4 2. 6 2. 4 97
Followupand other visits -2. 1 1. 8 2.7 3. 3 2. 5 100
Total -100. 0 100. 0 100.0 100. 0 100. 0 3,957
treatment or both in the
emergency
room, 14.8 percent of thepatientswereadmittedtothehos-pitalfor further care, 1.4 percent were referred to anotherfacility,forexample,thephysician's
office,
and 83.1percent were dischargedtotheir homes.Urgency.
White
determined the degree ofurgency from the emergency room records or
fromthephysician incaseswheredoubtexisted.
Aclassification ofconditions similarto that of
Weinerman
andco-workers(1)
wasused.Emergent: Threatening to life or limb without im-mediate medical attention. Acute and usually severe.
Urgent: Requiring medical attention within a few
hours. Acute butnotnecessarilysevere.
Nonurgent: Notrequiring the resources ofan emer-gencyroom.Nonacute, minor in severity.
Emergent cases resulted in 3.3 percent of
emergency room
visits,
urgent cases57.1percent,and nonurgent cases 35.9 percent.
Scheduled
procedures, such as minor surgery or cast
changes, prompted 3.3
percent
of the visits.Diagnoses.
Thediagnoses
wereindividually
coded and then grouped into 29
categories
foranalysis.
The 4 months studied showed verylittlevariation.Lacerations,
abrasions,
andcon-tusions, the largest single group in the
series,
accounted formorethan30percent ofthecases.
Ihe incidenceof these
diagnoses
increasedsig-nificantly in
July.
Theincidence
ofrespiratory
diseases increased
during
the monthsofOctober
andJanuary.
Overall,
50.4percentofcases weredue to trauma; 49.6 percent were nontrauma cases
(table 2).
Source
of
payment.
More thantwo-thirds of thepatients
had health insurance, 50.5 percentbeing
covered by BlueCross
and 16.2 percenthaving
othertypesof health insurance.Welfare
agencies, Medicare,
and Medicaidprovided
health carefor9.2 percent of the
patients,
and7.6 percent were coveredbycompensation. Only
9.3 percent of the patients had no assistance frominsuranceoragenciestomeettheirmedical
expense&
Distribution of Services by Time
The week wasdivided into six periods. Each daywasdivided into three 8-hour periods:
mid-nightto 8 a.m., 8 a.m. to 4p.m., and 4 p.m. to
midnight.Because of similarities indistribution
of patient
visits,
Saturdaysund
Sundays were consideredtogether, giving three weekend timeperiods.Thedays Monday through Friday were
also considered together to give three weekday
time periods.
Urgency ratings varied for thedifferent time intervals (table 3). The highest percent of emergent cases, 6.6 percent of the total cases,
occurred in the midnight to 8 a.m. period on weekends. For the corresponding time period
onweekdays, 5.4 percent of the cases were
emer-gent.Thehighest percentofurgentcases was in
the4p.m. tomidnight periods.
Table 4shows thedistributionofpatients by
service for the
six time
periods.
Thepediatric
service
wasresponsible
for thehighest
percent of patients in the 4 p.m. tomidnight
periods.
More patients
required
surgical
services frommidnight
to 8 am.onweekends and from8a.m.to 4 p.m. on
weekdays.
Thehighest
percentage ofmedical
cases inan8-hourperiod
wasin themidnightto8a.m.
periods.
Ahigher
proportion
ofthese cases werefoundtobeemergent.
Source of Referral
Distribution of service. Of the 3,957
pa-tients, 57.4 percentwere
self-referred
(table
5). Abouttwo-thirds of thepediatric patients-912
of 1,366-were self or parent referred, and 437
were referredby
physicians,
their office staff,oranswering
service.
About one-half of
surgical
and medicalpa-tients wereself-referred. As
might
be expected,alarge
majority
ofpatients
referredby
thepo-liceand
employers
weresurgical
patients.
Self-referral was more common for
pediatric
pa-tients than for
patients
in otherservices.
Urgency rating. The distribution of
pa-Table 3. Urgency ratings, in percentages, for patients seen during 8-hour intervals
Weekend Weekday
Total
Urgencyrating 12p.m.- 8a.m.- 4-12p.m. 12p.m.- 8 a.m.- 4-12p.m. (N=3,957)
8a.m. 4p.m. (N=555) 8a.m. 4p.m. (N=1,244) (N= 197) (N=563) (N=279) (N=1,119) Emergent 6.6 2.1 3. 8 5.4 3. 2 2. 9 3.3 Urgent -54.9 48. 0 62. 8 49.4 53.9 63. 2 57. 1 Nonurgent 36. 5 43.5 33. 0 44. 1 35. 4 32. 5 35.9 Scheduled 1.0 6. 2 .2 .4 6.9 1.1 3.3 Undetermined--- 1.0 .2 . 2 .7 .6 .3 . 4 Total 100.0 100. 0 100. 0 100. 0 100. 0 100. 0 100.0
Table 4. Percentage of patients seen during 8-hour intervals according to service responsible
for their care
Weekend Weekday
Total Service 12p.m.- 8a.m.- 4-12 p.m. 12 p.m.- 8a.m.- 4-12p.m. (N=3,957)
8a.m. 4p.m. (N=555) 8 a.m. 4p.m. (N=1,244) (N=197) (N=563) (N=279) (N=1,119) Pediatrics 18. 8 38. 2 43. 2 19. 4 29. 5 41.0 35. 1 Surgery -45.6 35.8 31.4 33. 0 45.6 33. 4 37. 4 Medicine- 28.0 18.3 18.9 35.4 17.4 18. 2 19.8 Obstetricsand gynecology 6.1 7. 3 6.3 11.1 7.1 7. 2 7. 3 Undetermined-- 1.5 .4 . 2 1.1 .4 .2 .4 Total 100.0 100.0 100.0 100.0 100.0 100.0 100. 0
Table5. Classification of patients according to service and source of referral
Physician's
Service Self Physician staffor Employer Police Total answering
service
Pediatric --- 912 370 67 1 16 1, 366
Surgery -792 291 50 240 93 1, 466
Medicine -438 261 37 26 12 774
Obstetricsand gynecology -98 169 16 3 2 288
Unabletoclassify -5 9 --- 14
Total -2,245 1,100 170 270 123 13,908 Source of referral notavailablefor49 patients.
Table 6. Urgency rating of patients' conditions, by source of referral
Physician's
Self Physician staff or Employer Police
answering
Urgency rating service Total
Num- Per- Num- Per- Num- Per- Num- Per- Num-
Per-ber cent ber cent ber cent ber cent ber cent
Emergent -64 2.85 39 3.54 2 1.17 11 4. 07 11 &94 127 Urgent -1, 369 60. 97 470 42. 72 74 43. 52 207 76. 66 105 85.36 2, 225 Nonurgent -804 35. 81 458 41. 63 90 52. 94 51 18. 88 7 5.69 1,410 Scheduled -3 .13 122 11.09 4 2.35 -129 Notrated -5 . 22 11 1. 00 1 . 37 17 Total-2, 245 99. 98 1, 100 99. 98 170 99. 98 270 99. 98 123 99.99 13, 908
1 Sourceofreferralnotrecorded for49 patients.
NOTE: Percentages donottotal100 because ofrounding.
tients according to urgency and source of referral is given in table 6. The percentages
refer to thedistributions according to urgency
level forall
patients
from eachreferralsource.Considering
emergent andurgentcasestogether,
employers and
police
made more of theirre-ferrals in this combined category than other
sources.
Of
the self-referredpatients,
63.8 per-centhademergent or urgentconditions.Physi-cians,
theiroffice staff,oranswering service
hadthe lowest
percent
(about 45percent)
oftheir referrals inthe emergent-urgentgroup.Discussion
Data
obtained
in thisstudy
groupdid
notidentify
patterns ofmisuse
of the emergencyroom.
Comparing
the study data with the 1960 censusdata,
thestudy
grouphad ahigher
per-centageofNegro
patients
andaslightly higher
percentage
unemployed
or in the labor group.However, the educational level of the study
group wassimilarto theeducational levelinthe
1960 population, and the
age
andsex distribu-tions for thestudygroup were similar to thosefoundin the 1960 census.
Ofthe
study
group,86percent claimedhaving
aprivatephysicianastheirusual source of
med-ical care, in contrast to the 36.6 percent found inthe YaJe-New Haven
Hospital
by Weinermanand co-workers (1). In the Yale-New Haven
Hospital study13.2percentwereclinicpatients, while in the Saginawhospital only 6.6percent of the emergency room
patients
claimed theclinicastheusual sourceofmedicalcare. Almost three-quarters of the
Saginaw patients
hadpri-vatehealthinsuranceorwere
receiving
compen-sation.
Only
10.9 percent of thesepatients
reported
using
the emergency room three or more times in theprevious
year. These data suggest that themajority
of thestudy patients
received privatemedical careandrelatively few
used theemergency roomasthemain sourceof
medical care.
Thecharacteristicsof theservices providedin
thisstudysuggest that theseservicesare
appro-priate to an emergency room. Sixtypercent of the cases were rated emergent orurgent, while
in
Weinerman's
study (1) 42percentwerecon-sidered emergent or urgent. One-half of the
study cases were trauma cases. Only about 2 percent of all patients seen had no diagnosed illness.Almost 10 percent of the patients, how-ever, hadsymptomsofsenility, ill defined. Prob-ablythis groupofpatients should beexamined
moreclosely.Werethesepatientsseeking treat-mentforanessentially untreatable conditionor
seeking attention or were families perhaps
try-ing to have the patient hospitalized to escape burdensome
care?
Thetimedistributionofpatients inthe emer-gency room was expected. The patient load increased on weekends and on the day when most physicians in the community close their
offices.
Thedistributionofurgent casesaccordingto source of referral indicates that referrals by
police and employers were almost always for
emergent or urgent cases. Of the self-referred patients, 36 percent had nonurgent conditions.
However,physicians' referralswere41.6percent nonurgent,andoffice-answeringservicereferrals
were52.9percent nonurgent.
Considering
that it isnot alwayseasyforlaymen todecidewhether an illness is urgent or forphysicians to decide this withoutexamination,somenonurgent cases should beexpected inanyemergencyroom.Pa-tients in this study didaswellas
physicians
inestimatingurgency.
In the sample of 214 patients interviewed,
there were no
significant discrepancies
in infor-mation obtainedby
the clerkcompared
within-formation elicited
by
the interviewer.Inter-viewing
did
not result inidentifying
instances of grossmisuse
of theemergency room. Therewereamongtheinterviewedgroupaswellasin
thelarger group,however,caseswhichwere not
emergentorurgent.Alternatewaysof
handling
such casesinclude hoasecalls,flexibilityin
phy-sicians' office hours, and service in
physicians'
officesandhospital clinics during regular office
hours.Education ofthepublicand ofphysicians and provision of acceptable alternates should minimize the use ofemergency rooms for non-urgentconditions.
Summary
A study of emergency room services at
Sag-inawGeneral Hospital,a voluntary, nonprofit,
short-stay hospital in amedium-sized city, was
conducted in 1966-67.The emergency room was
foundtoserveacrosssectionof thecommunity. Analysesof conditions diagnosed, sources of
referral, and urgencyof theconditions handled
suggest that the demands for services in this
emergency room were generally appropriate.
There was no evidence inthis study of specific patternsofmisuseof theemergency room.
Themethod used inthis studyisapplicableto any emergency room. Such a study is helpful
inanalyzingservices providedand inplanning programs. Emergency room service patterns will differaccordingtohospital size,
character-istics of
patients
served by that hospital, andsize and
characteristics
of the community in whichthehospital is located.REFERENCES
(1) Weinerman, E. R., Ratner, R. S., Robbins, A., and
Lavenhar, M. A.: Yale studies in ambulatory
medical care. V. Determinants ofuseofhospital
emergency services. Amer J Public Health 56:
1037-1056, July 1966.
(2) Wingert, W. A., Friedman, D. B., and Larson, W. R.: The demographical and ecological char-acteristics of a large urban pediatric outpatient
populationand implications for improving
com-munity pediatric care. Amer J Public Health
58:859-876,May1968.
(3) Hospital emergency room utilization study. A
re-port on emergency room visits in 22 Michigan
hospitals. Research and Statistical Division,
Michigan Blue Cross, Detroit, 1965.
Tearsheet Requests
Dr. Patricia A. O'Connor, University of Michigan
School of Public Health, 1111 E. Catherine St., Ann Arbor, Mich. 48104