Emergency Medicine Australasia (2020) doi: 10.1111/1742-6723.

13570

GLOBAL EMERGENCY CARE

Snapshot of emergency departments in Jakarta,

Indonesia
Liga YUSVIRAZI ,1,2 Septo SULISTIO,3 Andi Ade WIJAYA RAMLAN4,5 and Carlos A CAMARGO JR6
1
Division of Emergency Critical Care Medicine, Department of Emergency Medicine, Brigham and Women’s Hospital, Boston, MA, USA,
2
Indonesia Medical Education and Research Institute, Faculty of Medicine Universitas Indonesia, Jakarta, Indonesia, 3Emergency Department,
Cipto Mangunkusumo General Hospital, Jakarta, Indonesia, 4Department of Anesthesiology and Intensive Care, Cipto Mangunkusumo General
Hospital, Jakarta, Indonesia, 5Faculty of Medicine Universitas Indonesia, Jakarta, Indonesia, and 6Emergency Medicine, Massachusetts General
Hospital, Boston, MA, USA

Abstract maxillofacial and plastic surgery

Key findings
cases. Consultant coverage varied
Objective: EDs in Indonesia face an across hospitals and by hospital type • Jakarta had high ED visits
unprecedented increase in patient (P < 0.05), except for general surgery and density in 2017.
influx after the expansion of national and obstetrics and gynaecology con- • The ED contributed as the
health insurance system coverage. sultants who were available in most primary point of inpatient
The present study aims to describe hospitals (74%). admission.
EDs’ characteristics and capabilities Conclusion: Physicians with limited • Heterogeneity of consultant
utilisation in Jakarta. experience and EDs with heteroge-
Methods: An ED inventory was coverage among EDs in
neous emergency care capabilities
created from the Jakarta Provincial likely threatened the consistency of Jakarta might impair emer-
Health Office and the Indonesian quality emergency care, particularly gency care delivery.
Hospital Association registries. The for time-sensitive conditions. Our
EDs that were accessible to the gen- study provides a benchmark for incidence of cerebrovascular diseases
eral public 24/7 were surveyed about future improvements in emer- and motor vehicle accidents are ris-
their characteristics during the calen- gency care. ing and are now the leading causes
dar year 2017. For further ED analy- of premature death and disabilities.2
sis, we stratified the hospitals into Key words: emergency department Concurrently, the communicable dis-
four types (A, B, C and D) based on classification, health policy, Indone- ease rate is still high. The Indonesian
their size and capabilities, with type sia, international emergency medi- government increased its health bud-
A being the largest. cine, Jakarta. get to 5% of total national expendi-
Results: From the 118 (81%) out of ture and expanded coverage of its
146 EDs that responded, there were national healthcare system, which
2 million ED visits or 202 per 1000 Background covered approximately 70% of the
people. The median annual visit vol- Indonesian population.3 These sig-
Emergency medicine (EM) in Indone-
ume was 11 200 (interquartile range nificant changes led to an influx of
4233–18 000). Further stratification sia has not robustly developed as a
field. Despite its recognition as a spe- increased ED utilisation and hos-
highlights the annual visit difference
cialty in 2017,1 there is only one EM pitalised patients.4
among hospital types where type A
training programme throughout the The Indonesia Ministry of Health
hospitals reported the most with
nation. Thus, most EDs are staffed (MOH) regulates and mandates that
32 000 (interquartile range 13 459–
by general practitioners (GPs). With an ED be located in a hospital, and
38 873). Almost half of the EDs
remain open 24 h per day and 7 days
(47%) answered that ≥60% of the recent socioeconomic changes and
the increment of life expectancy, per week (24/7). The hospitals are
inpatient census came from the ED.
Less than half of the EDs (44%) can Indonesia’s healthcare system is fac- classified primarily based on pro-
manage psychiatry, oral- ing new challenges. For example, the vided-services and human resources.5
Per the classification, hospitals are
divided into five types: A, B, C, D and
Correspondence: Dr Liga Yusvirazi, St. Joseph Healthcare, Room 248, 360 Broadway, specialty hospital (Table 1). Type A
Bangor, ME 04401, USA. Email: liga.yusvirazi@live.com hospitals are the largest, most
Liga Yusvirazi, MD, Emergency Medicine Research Fellow, Internal Medicine Physi- equipped, provide most subspecialty
cian; Septo Sulistio, MD, Anesthesiologist, Lecturer; Andi Ade Wijaya Ramlan, MD, services and serve as a top referral
PhD, Pediatric Anesthesiology Consultant, Anesthesiology Staff Specialist, Lecturer; centre. In contrast, type D hospitals
Carlos A Camargo Jr, MD, DrPH, Professor, Conn Chair in Emergency Medicine. are the smallest among those classes
Accepted 1 June 2020 and have the least resources.

© 2020 Australasian College for Emergency Medicine

2 L YUSVIRAZI ET AL.

TABLE 1. Hospital and ED classification

Hospital classification A B C D

Subspecialty All subspecialty Limited subspecialty coverage +/− −

Specialty All specialty Type C + limited specialty Paediatrics, internal medicine, Paediatrics, internal
coverage general surgery, and medicine, general
OBGYN surgery, and
OBGYN
General practitioner + + + +

ED classification Level IV Level III Level II Level I

Capability Same as level III Level II + provide higher Level I + provide Provide basic
level resuscitation, advance life emergency
observation, provide support + diagnosis and
high care unit/ICU emergency treatment, such as
surgery perform BLS
Human resources
Subspecialty All − − −
subspecialty
Specialty All specialty  EM Level II + more selected Paediatrics, −
trained physician specialty coverage internal
(depend on each medicine,
hospital) general surgery,
and OBGYN
Physician in Available 24/7 Available 24/7 only in − −
specialty academic centre
training
General Available 24/7 Available 24/7 Available 24/7 Available 24/7
practitioner
Equipment
Operating room + + +/− −
Isolation room + +/− − −
Mobile X-ray + + + +/−
Mobile USG + + − −
CT scan + + − −
MRI + − − −
Laboratory
CBC + + + +
Basic chemistry + + + +
Blood gas + + +/− −
Cardiac marker + +/− − −

CBC, complete blood count; OBGYN, obstetrics and gynaecology; USG, ultrasonography.

Furthermore, each hospital type must level IV ED is a minimum require- function is to filter low acuity condi-
meet the minimum ED requirements, ment for type A hospitals while level I tions, perform stabilisation and then
which are divided into four different ED is an essential requirement for refer to a higher level of care.
levels of ED according to capability type D hospitals. This type of ED has Currently, there are no studies that
and services provided.6,7 For example, the least resources, and their primary have assessed and described EDs in

© 2020 Australasian College for Emergency Medicine

EMERGENCY DEPARTMENT UTILISATION IN JAKARTA, INDONESIA 3

Indonesia. The Emergency Medicine utilised by 95% of respondents. Sur- Categorical data were analysed with
Network’s National ED Inventory vey questions were drawn from sur- the χ 2 test or Fisher’s exact test.
(NEDI)-International project has been veys that were administered to Because some EDs lack robust
conducted in several countries and thousands of US EDs and more than data collection management systems,
found heterogeneity of EDs within the 10 other countries. Questions were we expected missing data for ED
same city and in different parts of each subdivided into four categories: ED annual visits, resulting in a loss of
country.8–13 Jakarta, Indonesia’s most characteristics, patient experiences in statistical power and potential
populous and diverse city, is at the bias.14 Indeed, 18 (15%) EDs lacked
the ED, capacity and resources and
front line of emergency care develop- annual visit volume data. To assess
capabilities.8 The questionnaire was
ment and has some of the most the impact of missing data, we
advanced EDs in the nation. In the modified based on structural and imputed these values based on the
present study, we sought to describe organisational recommendations for following ED characteristics: hospi-
the characteristics, resources, capacity EDs in Indonesia.6,7 We did not tal ownership, hospital type, hospital
and capabilities of EDs in Jakarta. include questions regarding ED loca- beds, ED beds, ED capacity and ED
tion because all EDs in Indonesia are length of stay (LOS). We used condi-
located within a hospital and are tional specification methods in the R
Methods package multivariate imputation by
independent departments. Addition-
This is a cross-sectional descriptive ally, all EDs in Jakarta are required chained equations (‘mice’).15 In this
study that utilises a survey that was to treat both adults and children.6 approach, each incomplete variable
sent to ED leadership, either a physi- ‘Triage to service’ refers to the pro- is regressed in turn on all other
cian or a nurse at every ED in cess in which the patient arrives at covariates, and the process is iterated
Jakarta. Using terminology from the the ED and is directed to emergency until the regression estimate sta-
NEDI-USA study,12 ED is defined as care by a specialty team for emer- bilises. The imputed value will be
an emergency care facility that is gency care, for example surgical vs chosen using predicted mean
open 24/7. Furthermore, the ED is medical team.8 In Indonesia, the matching, which selects a value at
open to the public, regardless of dis- healthcare provider triaging to ser- random from among the observed
ease state and socioeconomic status. vice is a GP. For further characteri- data for EDs in which predicted
Specialty hospitals (such as cardio- sation, we added hospital type and values are closest to those of the one
vascular hospitals and psychiatric ownership (public vs private) in the with a missing value. Statistical anal-
institutions) are excluded because of survey. The survey was translated ysis was performed using R (R Foun-
their particular structures and techni- into the Indonesian language by dation for Statistical Computing,
cal capabilities. The list of hospitals interpreter services and indepen- Vienna, Austria). A two-sided
was compiled from the data dently checked for accuracy by three P < 0.05 was considered statistically
obtained from the MOH, the Indo- bilingual Indonesian physicians significant. Using ArcGIS 10.6 soft-
nesian Hospital Association and (Appendix S1). Responses were ware (Redlands, WA, USA), EDs
information collected independently entered into LimeSurvey (www. were geocoded (i.e. assigned map
by researchers. Two local physicians limesurvey.org) and downloaded coordinates based on zip code) and
(SS, AAWR) verified the hospital list onto an Excel spreadsheet (Micro- then were mapped according to
for completeness. Research assistants soft Corp., Redmond, WA, USA). annual visit volume (annual visit
contacted and surveyed 199 ED hos- We started the survey in mid-2018 quartiles).
pitals from the list. The study was and specifically asked the participant
endorsed by the Emergency Medi- about their ED characteristics in ref-
cine Network (www.emnet-nedi. erence to the calendar year 2017.
Results
org). The Ethics Committee of the Out of 199 hospitals in Jakarta, 146
Faculty of Medicine Universitas hospitals were eligible, and 118 EDs
Indonesia (Jakarta, Indonesia)
Statistical analysis participated in the survey (81%
approved the study (approval num- Continuous variables are presented response rate). Research assistants
ber 18-08-0904). It was determined with an average and standard devia- re-contacted ED administrators who
to be exempt by the Partners Institu- tion (SD), while the data with non- returned partially filled forms. A
tional Review Board (Boston, normal distributions are presented total of 53 EDs were not eligible for
MA, USA). with a median and interquartile the study because of various reasons
In the present study, we used four range (IQR). Categorical variables (Fig. 1). Surveyed ED hospital loca-
varieties of the NEDI-International are presented as percentages. Subse- tions can be seen in Appendix S2.
questionnaires: contiguous without quently, we further stratified the data As shown in Table 2, more than
triage to service, contiguous with tri- according to hospital type. Inter- half (67%) of the 118 hospital-based
age to service, non-contiguous with- group comparisons were analysed EDs are owned by private entities.
out triage to service and non- using analysis of variance for contin- The EDs were equally distributed
contiguous with triage to service.8 uous variables with normal distribu- among five districts in Jakarta. The
The contiguous without triage to ser- tions and the Kruskal–Wallis test for majority of ED layouts were contigu-
vice type of questionnaire was data with non-normal distributions. ous (95%), where medical and

© 2020 Australasian College for Emergency Medicine

4 L YUSVIRAZI ET AL.

reported an average LOS of >6 h, Indonesia, with a focus on Jakarta.

Hospitals screened = 199
two of which were in type A hospi- With >80% response rate and data
tals. Almost half (47%) of EDs from 118 EDs, we found that there
reported that they operated at the was a significant ED utilisation, high
right balance. Only 9% reported hospitalisation rate from ED, a diverse
Excluded:
13 hospitals closed
that their EDs exceeded capacity, range of ED capabilities and variation
2 hospitals merged mainly EDs in type B and D hospi- of consultant coverage.
33 specialty hospitals
1 hospital without 24 hours ED
tals. Almost half (48%) of the There are similarities and differ-
1 hospital mostly for inmates respondents reported ≥40% of ED ences between EDs in Jakarta com-
3 hospitals opened in 2017
visits led to hospitalisation, with pared to others in different
almost half of EDs (47%) answering countries. From 81% EDs, there
that ≥60% of the inpatient census were approximately 2 million ED
Hospitals surveyed = 146
came from the ED. There was no sig- visits in 2017, which might underes-
Completed survey = 118 (81%) nificant difference among hospital timate the actual number of ED
Abbreviation: ED, Emergency Department.
types in terms of the percentage of visits. Compared to the 2007 MOH
inpatients that were admitted from national public hospital report, the
Figure 1. Screening and survey the ED. Technology support was present study showed half of the
implementation. high, where most EDs had access to national ED visits.5 This finding
cardiac monitors (97%), internet emphasises the use of ED as a signifi-
(86%), and 24-h clinical laboratory cant contributor to the health service
surgical care was provided in one support (73%). CT scans, mechani- since Jakarta only represents approx-
area, with a few (14%) using triage cal ventilators and computer system imately 4% of Indonesia’s popula-
to service. The ED has a median bed access, however, were variable and tion.16 Moreover, compared to
number of 8 (IQR 5–13). In 2017, were mostly located in type A and B NEDI studies in Beijing and Singa-
there were approximately 2 million hospitals. pore, Jakarta has a slightly higher
ED visits with 11 200 (IQR 4233– Most of the emergency cases could ED visit rate per 1000 people. This
18 000) median annual visits. There be treated 24/7 in the ED (Table 3). finding may be explained by
was no significant difference from However, only 44% of EDs were Jakarta’s ED density, which is
the non-imputed data (P = 0.76). able to manage psychiatric, oral-max- approximately 10 times higher than
Based on a population of 10.3 mil- illofacial and plastic surgery emer- in Singapore and 19 times higher
lion,16 the estimated total number of gency cases. There was significant than in Beijing.9,10
ED visits in Jakarta was 202 per variability in treating emergency cases Among EDs in Jakarta, the charac-
1000 people. The EDs were further among hospital types (P < 0.05), teristics varied according to hospital
stratified according to hospital types: except ENT, psychiatry and toxicol- types determined by the MOH. Type
A, B, C and D. Type B (41%) and C ogy. Additionally, there was a distinct A hospitals have the most ED annual
(35%) hospitals comprise the major- variability of consultant coverage in visits among all hospital types.
ity of hospitals in Jakarta. The gov- each hospital type (P < 0.05), except Despite being more equipped, type C
ernment mostly owns type A (86%) for general surgery, obstetrics and hospitals have less visit volume com-
and D (96%) hospitals, where pri- gynaecology and psychiatry (Table 4). pared to type D hospitals. The fact
vate entities mostly own type B Limitations to consultant access that type D hospitals are mostly pub-
(77%) and type C (95%) hospitals. affected mostly type C and D hospi- lic-owned and accept national
All hospital types were distributed tals. Type D hospitals did not have healthcare system could partially
evenly among five districts. Type A coverage for neurology, neurosur- explain the difference. Another pos-
hospitals have a higher percentage of gery, orthopaedics, plastic surgery sible explanation is that type D hos-
non-contiguous EDs (29%) com- and psychiatry cases. Consultant pitals are located in areas with
pared to the rest and have the most availability appeared to correlate higher population densities and more
ED annual visits with a median of with hospital type and the type of sparse ED coverage. Hence for con-
32 000 (IQR 13 459–38 873). Type emergency cases that respective EDs venience, most people may elect to
C hospitals had fewer ED annual were capable of treating (Fig. 3). go to the closest hospital for emer-
visits (7200 [IQR 3066–11 200]) gency conditions, which may par-
compared to EDs in type D hospitals tially explain the higher number of
(12 000 [IQR 2725–16 672]) (Fig. 2).
Discussion type D hospital ED visits. In addi-
Most respondents (58%) answered Emergency care has not received much tion, 47 respondents reported to
that <20% of their patient popula- attention in Indonesia, and it con- have ≥60% inpatients admitted from
tion arrived by ambulance. Type D tinues to develop at a slow pace. the ED, and 48% described an ED
hospitals reported a large number of Although minimum requirements are admission rate of ≥40%, which out-
patients who came without an regulated,7 no studies have been con- lines the importance of ED as the
ambulance (86%). In the majority of ducted which assesses the capabilities primary route of hospital admission.
EDs (86%), the average LOS was and characteristics of the EDs. This Consequently, ED plays a fundamen-
between 1 and 6 h. Only four EDs survey is the first to depict EDs in tal role in overall hospital function

© 2020 Australasian College for Emergency Medicine

EMERGENCY DEPARTMENT UTILISATION IN JAKARTA, INDONESIA 5

TABLE 2. Characteristic ED in Jakarta, Indonesia, based on 118 sampled EDs†

Hospital type

All, n = 118 A, n = 7 (6%) B, n = 48 (41%) C, n = 41 (35%) D, n = 22 (19%) P-value

Ownership
Public 39 (33) 6 (86) 11 (23) 2 (5) 20 (91) <0.001
Private 79 (67) 1 (14) 37 (77) 39 (95) 2 (9)
Location
North Jakarta 23 (19) 1 (14) 7 (15) 9 (22) 6 (27) 0.53
Centre Jakarta 22 (19) 2 (29) 11 (23) 4 (10) 5 (23)
West Jakarta 15 (13) 0 (0) 9 (19) 3 (7) 3 (14)
East Jakarta 31 (26) 2 (29) 11 (23) 15 (37) 3 (14)
South Jakarta 27 (23) 2 (29) 10 (21) 10 (24) 5 (23)
ED contiguous 112 (95) 5 (71) 45 (94) 40 (98) 22 (100) 0.02
ED annual visits, 11 200 32 000 14 168 7200 12 000 0.002
median (IQR) (4233–18 000) (13 459–38 873) (5829–24 000) (3066–11 200) (2725–16 672)
ED beds, median 8 (5–13) 40 (23–54) 12 (8–16) 6 (4–8) 5 (4–6) <0.001
(IQR)
Hospital beds, 90 (43–182) 600 (188–775) 181 (124–304) 60 (35–91) 38 (30–50) <0.001
median (IQR)
ED triage to service 17 (14) 5 (71) 9 (19) 3 (7) 0 (0) <0.001
Come to ED by ambulance
<20% 68 (58) 3 (43) 20 (42) 26 (63) 19 (86) 0.003
20–39% 28 (24) 1 (14) 17 (35) 8 (20) 2 (9) 0.07
40–59% 6 (5) 0 (0) 3 (6) 3 (7) 0 (0) 0.55
≥60% 2 (2) 0 (0) 2 (4) 0 (0) 0 (0) 0.40
No response 14 (11) 3 (43) 6 (13) 4 (10) 1 (5)
ED length of stay
<1 h 12 (10) 1 (14) 5 (10) 4 (10) 2 (9) 0.98
1–6 h 102 (86) 4 (57) 42 (88) 37 (90) 19 (86) 0.13
>6 h 4 (4) 2 (29) 1 (2) 0 (0) 1 (5) 0.002
ED capacity
Under capacity 16 (14) 0 (0) 4 (8) 9 (22) 3 (14) 0.20
Balance 56 (47) 3 (43) 24 (50) 21 (51) 8 (36) 0.68
At right capacity 34 (29) 3 (43) 14 (29) 11 (27) 6 (27) 0.85
Over capacity 11 (9) 1 (14) 5 (10) 0 (0) 5 (23) 0.03
Percentage inpatient admitted from ED
<20% 16 (14) 1 (14) 7 (15) 5 (12) 3 (14) 0.99
20–39% 18 (15) 2 (29) 7 (15) 8 (20) 1 (5) 0.32
40–59% 21 (18) 0 (0) 10 (21) 8 (20) 3 (14) 0.54
60–79% 32 (27) 2 (29) 13 (27) 7 (17) 10 (45) 0.12
≥80% 24 (20) 2 (29) 8 (17) 10 (24) 4 (17) 0.76
No response 7 (6) 0 (0) 3 (5) 3 (7) 1 (5)

(Continues)

© 2020 Australasian College for Emergency Medicine

6 L YUSVIRAZI ET AL.

TABLE 2. Continued
Hospital type

All, n = 118 A, n = 7 (6%) B, n = 48 (41%) C, n = 41 (35%) D, n = 22 (19%) P-value

ED visit lead to admission

<20% 20 (17) 0 (0) 9 (19) 6 (14) 5 (23) 0.53
20–39% 26 (22) 2 (29) 13 (27) 8 (20) 3 (14) 0.59
40–59% 34 (29) 3 (43) 14 (29) 11 (27) 6 (27) 0.85
60–79% 12 (10) 1 (14) 6 (13) 2 (5) 3 (14) 0.58
≥80% 11 (9) 1 (14) 1 (2) 7 (17) 2 (9) 0.11
No response 15 (13) 0 (0) 5 (10) 7 (17) 3 (14)
ED resource
CT scan 59 (50) 7 (100) 39 (81) 12 (29) 1 (5) <0.001
Cardiac monitor 114 (97) 7 (100) 48 (100) 39 (95) 20 (91) 0.22
Mechanical 58 (49) 6 (86) 35 (73) 10 (24) 7 (32) <0.001
ventilator
Negative pressure 42 (36) 5 (71) 28 (58) 6 (15) 3 (14) <0.001
room
Electronic medical 74 (63) 6 (86) 37 (77) 20 (49) 11 (50) 0.01
record
Internet 101 (86) 7 (100) 42 (88) 31 (76) 21 (96) 0.10
24-h laboratory 86 (73) 6 (86) 41 (85) 27 (66) 12 (55) 0.03
service

†Results are n (%) unless otherwise noted. Percentages may sum to >100% because of rounding. The total annual visit in
2017 was 2 344 205. The total hospital bed was 17 972. The total ED bed was 1383. IQR, interquartile range.

and the economic status of most robust data to support this, we for emergency care. Based on one
hospitals. believe that most patients use private study at Cipto Mangunkusumo Gen-
The National Command Centre automobiles or taxis to get to the eral Hospital, two main reasons
coordinated the pre-hospital system ED.4 To increase emergency access, appeared to be responsible for pro-
from MOH office.17 The dispatcher the government needs to inform and longed LOS.18 Firstly, consultation
in National Command Centre will educate citizens regarding the impor- process in an academic hospital
inform the nearest local public safety tance of using the emergency call involves several physicians in spe-
centre for the emergency, then direct centre, besides, increasing pre-hospi- cialty training (from junior to chief)
them to the healthcare facility. Nev- tal services funding to add more before it reaches the consultant for a
ertheless, the present study showed ambulance coverage. disposition, consuming a large
low ambulance utilisation in Jakarta. Most EDs in Jakarta were located amount of time. Second, lack of
Based on unpublished data from in non-academic hospitals and staffed hospital bed availability and ineffi-
municipal health officials, most of by GPs, which may explain the low cient management worsens the
the ambulance usage is primarily for numbers of triage to service. Special- placement of ED patients, especially
inter-hospital transportation or ists are usually consulted if there is a in type A hospitals or academic cen-
transferring patients from home to complicated case or if a patient tres. Most EDs reported a typical
ambulatory clinics or haemodialysis requires admission. Most GPs are stay between 1 and 6 h, explained
centres. Aside from ambulance short- new graduates and might not have by the two following reasons: a
age, minimal public information adequate clinical education in the more straightforward consulting
regarding the emergency call centre management of patients in an emer- process and low patient acuity
and the fear of financial burden for gency setting.4 In academic hospitals, because of ED dual functionality as
the ambulance user might explain triage to service occurs when a GP a walk-in clinic after hours. Some of
the low usage. Although there is no directs a patient to a specialty team the hospitals, located in areas that

© 2020 Australasian College for Emergency Medicine

EMERGENCY DEPARTMENT UTILISATION IN JAKARTA, INDONESIA 7

across the country, which aggravated

consultant availability and coverage
among hospitals in Jakarta.3 Com-
pared to the private hospitals, physi-
cian reimbursement in the public
hospitals is generally lower, and most
of the physicians also work in three
different hospitals to meet a living
standard which partly explains the gap
in physician coverage. Training and
distributing more consultant might
lead to optimisation of the delivery of
emergency patient care. Lastly,
resource disparities among hospital
types could worsen the delivery of
emergency care and might harm
patients who need specific workup or
intervention. Well-equipped public
hospitals operated as a safety net and
gave easy access to the population,
especially the middle-low-income com-
munity. We need more study to dem-
onstrate ED utilisation and how far
emergency care quality impacted by
the variation of resources. Current ED
categorisation has not accommodated
the need for an individual or pre-hos-
pital emergency medical system to
direct them to the appropriate ED.
Assigning hospitals as trauma centres,
appointing stroke centres and designat-
ing 24/7 heart catheterisation labs will
provide more valuable information
and increase the quality of emergency
care, especially for time-sensitive con-
ditions. Hence, policymakers will be
able to establish emergency care plans/
Figure 2. Emergency department visits and population density. disaster planning at the regional or
national level with more specific
information.23
are sparsely ED-covered or high- another way to optimise emergency
density sub-districts in Jakarta, have care, especially in small ED.
a LOS >6 h. To optimise the quality Our study highlights the significant
Limitations
of emergency care, we could train differences among hospital types for We recognise this is a pilot study
more EM physicians. Several studies ED resources and consultant coverage. with descriptive statistics, but it pro-
justify the benefits of EM-trained As we expected, type A hospital EDs vides new insights to guide efforts to
physicians to take care of critically were more equipped compared to the advance emergency care in Indone-
ill patients.19–22 The MOH has rec- other EDs. Furthermore, the capability sia. To our knowledge, a validated
ommended level IV ED staffed by of EDs to handle cases varied instrument to assess ED worldwide
the EM-trained physician.7 Type A depending on their hospital type and does not exist. Questions from our
and B hospitals, usually serve as a consultant coverage. Certain emer- survey have been used in US studies
referral, need to staff their ED with gency cases (i.e. oncology, neurology, and in several other countries,9–13
EM-trained physician. To staff low urology, ophthalmology and ENT) which certifies usability and ensures
census EDs with EM-trained physi- can be managed only in ≤75% of the the wording of questions to be
cians might not be cost-effective. EDs. The smaller the hospital, the appropriate in diverse contexts.
Besides, it is only one EM training lesser and limited access to consultants. Another potential limitation is that
programme in Indonesia to date.4 The ability of EDs in those hospital the present study relies on self-
Improving EM training curriculum types to manage time-sensitive condi- reported data. While hospital docu-
for medical students or GPs who tions was therefore limited. One possi- mented data would be ideal, almost
work in the ED could provide ble explanation is consultant shortage all of the surveyed hospitals have

© 2020 Australasian College for Emergency Medicine

8 L YUSVIRAZI ET AL.

Figure 3. Association between ED case and consultant. ( ) Case; ( ) consultant coverage.

lacklustre record-keeping. Hence, responding sites (19%) did not sig- play a significant role in the delivery of
ED administrators were only able to nificantly differ in key parameters acute healthcare, specifically for the
provide their best estimates. Finally, from those who did (81%). Jakarta population. It also identifies
the response rate in the study was the wide variability of ED resources,
81%, with 28 EDs choosing not to consultant coverages and the provision
participate. If their experiences or
Conclusion of emergency care by non-EM trained
responses markedly differed from The present study demonstrates the physicians who have potentially hin-
those studied, this would alter our immense volume of ED visits in dered the universal delivery of qual-
findings; however, the non- Jakarta in 2017. It shows that EDs ity emergency care, particularly for

© 2020 Australasian College for Emergency Medicine

EMERGENCY DEPARTMENT UTILISATION IN JAKARTA, INDONESIA 9

TABLE 3. Percentage of EDs able to treat 24/7, overall and by hospital type

Emergency type Example of emergency All, n = 118 A, n = 7 B, n = 48 C, n = 41 D, n = 22 P-value

Medical-Cardiology Arrhythmia, AMI 89% 100 96 76 91 0.02

Medical-Oncology Neutropenic fever 70% 86 79 67 46 0.03
Medical-Other UTI, acute asthma 93% 71 100 86 96 0.01
Trauma Motor vehicle accident, stab wound 91% 100 98 78 95 0.008
Neurological and Intracranial haemorrhage, acute 67% 86 90 54 41 <0.001
neurosurgery cerebrovascular accident
Urological Kidney stone 74% 86 85 68 52 0.02
Obstetrical Peripartum haemorrhage 94% 10 98 95 81 0.049
Gynaecological Ruptured ovarian cyst, yeast infection 86% 71 92 92 64 0.007
Ear, nose, throat Severe epistaxis 65% 71 71 76 50 0.21
Ophthalmological Acute glaucoma, eye injury 55% 57 63 62 27 0.03
Toxicological Overdose, carbon monoxide 71% 71 71 73 68 0.98
poisoning
Psychiatric Psychosis 44% 57 48 44 32 0.55
Dental Tooth extraction 46% 57 53 49 18 0.04
Surgical-Oral Jaw fracture, oral abscess 44% 57 63 32 14 0.001
maxillofacial
Surgical-Plastic Severe lip laceration 44% 71 66 30 14 <0.001
Surgical-Hand Tendon injury 52% 86 72 43 9 <0.002
Surgical-Orthopaedic Long bone fracture 68% 86 87 64 27 <0.003
Surgical-General Acute appendicitis, pneumothorax 80% 86 92 81 46 <0.004

Martin Lukmanto, Siseana Gabriela,

TABLE 4. Percentage of EDs with consultant availability, overall and by hos- Florentina Priscillia and Sarah
pital type Levita, for collecting data and relent-
lessly contacting ED administrators
Consultant
to complete the survey. We thank
specialty All, n = 118 A, n = 7 B, n = 48 C, n = 41 D, n = 22 P-value
Dian Kusuma, MPH, PhD, and
Cardiology 52 83 72 40 7 <0.001 Venkat Vuddandda, MD, MS, in
assisting with statistical analysis. We
General surgery 74 83 84 72 56 0.16
thank Yufi Priadi, MSc for helping
Orthopaedics 51 83 72 40 0 <0.001 generated the geographical analysis.
Plastic surgery 31 50 54 11 0 <0.001
Neurology and 59 67 63 16 0 <0.001
Author contributions
neurosurgery
Obstetrics and 74 83 84 69 59 0.18 LY, CAC conceived the study. LY,
gynaecology SS, AAWR, CAC contributed to
Psychiatry 25 33 38 19 0 0.02 the creation of the questionnaire.
SS and AAWR collected the data.
LY generated the geographical
analyses. LY performed the statis-
time-sensitive conditions. We hope Acknowledgements tical analyses. LY, SS, AAWR dra-
that our study can be a catalyst for
We thank the survey participants for fted the manuscript. All authors
the development of EM and serve as
a tool to monitor the advancement their essential contributions to emer- revised the manuscript. All
of emergency care both regionally gency care in Jakarta, Indonesia. We authors read and approved the
and nationally. also thank the research assistants, final manuscript.

© 2020 Australasian College for Emergency Medicine

10 L YUSVIRAZI ET AL.

Competing interests Nomor 47 Tahun 2018. 2018. sp2010.bps.go.id/index.php/site?id=

[Cited 2 Sep 2019.] Available from 3100000000&wilayah=DKI-Jakarta
None declared. URL: http://ditjenpp.kemenkumham. 17. Peraturan Kementrian Kesehatan.
go.id/arsip/bn/2018/bn1799-2018.pdf Permenkes Nomor 19 Tahun 2016.
Data availability statement 8. Steptoe AB, Corel B, Sullivan AF, 2016. [Cited 5 May 2020.] Avail-
Camargo CA Jr. Characterizing able from URL: http://hukor.
Data available on request due to pri- emergency departments to improve kemkes.go.id/uploads/produk_
vacy/ethical restrictions. understanding of emergency care hukum/PMK_No._19_ttg_Sistem_
systems. Int. J. Emerg. Med. 2011; Penanggulangan_Gawat_Darurat_
4: 42. Terpadu_.pdf
References 9. Wen LS, Xu J, Steptoe AP et al. 18. Habib H. Patient Flow IGD RSCM
1. Indonesian Association of Emer- Emergency department characteristics (Lecture). Faculty of Medicine, Uni-
gency Medicine. Tentang Perdamsi. and capabilities in Beijing, China. J. versity of Indonesia Rumah Sakit
[Cited 8 Jul 2019.] Available from Emerg. Med. 2013; 44: 1174–9.e4. Cipto Mangunkusumo. (12 Feb
URL: https://perdamsi.or.id/about-us 10. Wen LS, Venkataraman A, 2019).
2. Institute of Health Metrics and Sullivan AF, Camargo CA. 19. Nelson M, Waldrop RD, Jones J,
Evaluation. Global Burden of Dis- National inventory of emergency Randall Z. Critical care provided in
ease: Indonesia. [Cited 1 Sep 2019.] departments in Singapore. Int. J. an urban emergency department.
Available from URL: http://www. Emerg. Med. 2012; 5: 38. Am. J. Emerg. Med. 1998; 16:
healthdata.org/sites/default/files/files 11. Bustos Y, Castro J, Wen LS, 56–9.
/country_profiles/GBD/ihme_gbd_c Sullivan AF, Chen DK, 20. Panacek EA. Intensive care unit
ountry_report_indonesia.pdf Camargo CA Jr. Emergency depart- care in the emergency department.
3. Agustina R, Dartanto T, ment characteristics and capabilities Acad. Emerg. Med. 1999; 6:
Sitompul R et al. Universal health in Bogotá Colombia. Int. J. Emerg. 675–7.
coverage in Indonesia: concept, pro- Med. 2015; 8: 30. 21. Overton DT. Emergency medicine
gress, and challenges. Lancet 2019; 12. Sullivan AF, Richman IB, Ahn CJ and critical care medicine: have the
393: 75–102. et al. A profile of US emergency stars (finally) aligned? Ann. Emerg.
4. Yusvirazi L, Ramlan AAW, departments in 2001. Ann. Emerg. Med. 2005; 46: 225–7.
Hou PC. State of emergency in Med. 2006; 48: 694–701. 22. Holliman CJ, Mulligan TM,
Indonesia. Emerg. Med. Australas. 13. Sanchez B, Hirzel AH, Bingisser R Suter RE et al. The efficacy and
2018; 30: 820–6. et al. State of emergency medicine value of emergency medicine: a sup-
5. Kementerian Kesehatan Republik in Switzerland: a national profile of portive literature review. Int. J.
Indonesia. Permenkes Nomor 56 emergency departments in 2006. Emerg. Med. 2011; 4: 44.
Tahun 2014. 2014. [Cited 8 May Int. J. Emerg. Med. 2013; 6: 23. 23. Bernstein SL, Haukoos JS. Public
2019.] Available from URL: http:// 14. Bell ML, Fairclough DL. Practical health, prevention, and emergency
yankes.kemkes.go.id/assets/downlo and statistical issues in missing data medicine: a critical juxtaposition.
ads/PMK%20No.%2056%20ttg% for longitudinal patient-reported Acad. Emerg. Med. 2008; 15: 190–3.
20Klasifikasi%20dan%20Perizinan outcomes. Stat. Methods Med. Res.
%20Rumah%20Sakit.pdf 2014; 23: 440–59.
6. Kementerian Kesehatan Republik 15. van Buuren S, Groothuis- Supporting information
Indonesia. Kepmenkes Nomor 856 Oudshoorn K. mice: Multivariate
Tahun 2009. 2009. [Cited 25 May Imputation by Chained Equations Additional supporting informa-
2019.] Available from URL: http:// in R. 2011. [Cited 15 Aug 2019.] tion may be found in the online ver-
sardjitohospital.co.id/sardjitowp/wp- Available from URL: http://doc. sion of this article at the publisher’s
content/uploads/2015/12/kepmenkes- utwente.nl/78938/ web site:
856-thn-2009-standar-IGD.pdf 16. Badan Pusat Statistik. Sensus
7. Peraturan Kementrian Kesehatan Penduduk 2010. 2010. [Cited 1 Jul Appendix S1. Survey questionnaire.
Republik Indonesia. Permenkes 2019.] Available from URL: https:// Appendix S2. Hospital response map.

© 2020 Australasian College for Emergency Medicine