Timely Access to Surgical Care in Indonesia

Workforce, volume, and spatial coverage — ARC8 Global Surgery Institute

ARC Institute · 8.3 Timely Access

2026-04-23

Study objectives

  1. Describe the national distribution of surgical workforce capability (SAO teams) across Indonesian hospitals
  2. Quantify bellwether procedure volumes from BPJS claims (2016–2024)
  3. Link workforce capability to surgical output at district level
  4. Map population spatial access to SAO-capable hospitals (ArcGIS network analysis)
  5. Test whether poor spatial access drives patient cross-district outflow

Data sources

Source Content Years
SIRS National hospital registry — class, beds, location 2025 snapshot
DREAMS Specialist workforce — headcount (HC) + practice permits (SIP) 2025 snapshot
BPJS Kesehatan Inpatient claims — ICD-9-CM procedures, member weights 2016–2024
ArcGIS Network Analyst Hospital service-area polygons (2 hr drive-time) 2025
WorldPop Gridded population raster — district catchment population 2020

SAO capability framework

SAO = Surgeon (gen surgery + orthopaedics) · Anesthesiologist · Obstetrician/Gynaecologist

Five criteria assessed in parallel — from most permissive to most demanding:

Criterion Definition N hospitals % eligible
SAO ≥1 (HC) S, A, O each ≥1 by primary-workplace headcount 1175 36%
SAO ≥1 (SIP) S, A, O each ≥1 by practice permit (all sites) 2680 82%
SO ≥1 (HC) S ≥1, O ≥1 by headcount — anaesthesia gap allowed 1489 45%
SAO ≥2 (SIP) S, A, O each ≥2 by practice permit — full minimum team 1593 49%
SO ≥2 (SIP) S ≥2, O ≥2 by practice permit — anaesthesia gap allowed 1881 57%

Bellwether procedures: C-section · Open fracture fixation · Laparotomy

Study sample — hospitals

Study sample — BPJS claims

Year Claims Wt claims Any BW C-sec Frac Lap
2016 103,259 8,740,371 8,844 6,842 745 1,261
2017 103,058 8,808,812 9,983 7,851 885 1,248
2018 111,693 9,587,176 11,268 8,976 944 1,352
2019 123,835 10,409,536 12,047 9,661 976 1,411
2020 97,945 7,678,694 12,402 10,480 839 1,089
2021 96,950 7,198,189 12,916 10,839 946 1,135
2022 137,099 10,715,926 14,690 11,719 1,449 1,523
2023 170,773 13,894,948 16,330 12,693 1,805 1,844
2024 194,201 16,185,911 17,039 12,696 2,155 2,193
Total 1,138,813 93,219,564 115,519 91,757 10,744 13,056

Claims = inpatient episodes after age/sex/date filters. Wt claims = sum of PSTV15 member weights. 2018 excluded (incomplete).

National SAO capability

(Fig 01)

Only 82% of eligible hospitals have at least one SAO team by practice permit (SAO ≥1 SIP). Full teams (SAO ≥2 SIP) cover 49%.

HC vs SIP gap — why two counts matter

(Fig 02)

SIP > HC reflects multi-site practice — the same specialist holds permits at multiple hospitals. Using SIP alone inflates apparent coverage; HC alone misses specialists active at secondary sites.

SAO capability by island (Fig 03)

Jawa concentrates SAO ≥2 hospitals. Papua, Maluku, and Kalimantan show the highest proportions of non-capable hospitals.

Bellwether procedure rates — national trend (Fig 04)

C-section dominates and grew from ~1,009 to ~1,333 per 100k WRA (2016–2024). Fracture fixation doubled. Laparotomy stagnant — possible ceiling or data artefact.

BW rates by hospital type (Fig 05)

Type C (mid-tier general) hospitals carry the largest share of national BW volume — consistent with JKN referral design where most insured patients access district-level hospitals.

Province-level BW rates (Fig 06)

SAO tier × hospital class (Fig 09)

Tiers are SIP-based (practice permit). Class C hospitals make up the largest volume; class A hospitals are the most likely to be SAO ≥2.

Hospital capability by class × criterion (Fig 10)

All five criteria shown simultaneously. Class A/B are consistently more capable; gap between HC and SIP rates is widest in class C/D.

Does workforce capability predict surgical output? (Fig 08)

District-level SAO capability (% hospitals meeting each criterion) vs pooled 2022–2024 crude BW rates.

Positive association across all five criteria, strongest for C-section. SAO ≥2 (SIP) shows the clearest signal.

Case fatality & cost (Fig 11, 12)

Fig 11 — Case fatality rate

Fig 12 — Mean reimbursement per case

Laparotomy CFR fell from 4.5% → 2.4%; C-section CFR remains <0.2%. Cost per case rose in real terms across all procedures.

Patient drift — national trend (Fig 13)

~30–35% of bellwether procedures are delivered in a different district from the patient’s home district. Drift has been stable since 2019; it is not growing despite increased overall volume.

Net importer and exporter districts (Fig 14)

Fig 14 — Top importers/exporters

Jakarta, Surabaya, Makassar = dominant referral hubs (net importers). Surrounding peri-urban and rural districts are the largest exporters.

Population coverage by SAO-capable hospital (Fig 15)

Coverage by island group (Fig 16)

Jawa >90% coverage even under the strictest criterion. Papua <30% under SAO ≥1 (HC) and <15% under SAO ≥2 (SIP).

Coverage × surgical volume (Fig 17)

Weak positive association — many high-coverage districts still have low volume (demand-side or data-attribution issues).

Spatial coverage as a driver of patient drift (Fig 19, 20)

Fig 19 — Coverage vs outflow

Fig 20 — Island-level medians

Districts with lower catchment coverage export more patients. Spearman ρ is negative (more coverage → less outflow), though effect is moderate.

Population burden per hospital (Fig 21)

Districts with higher population-per-qualifying-hospital ratios tend to have higher patient outflow — a supply-constraint signal distinct from geographic reach.

Key findings

  1. Full SAO teams are rare — only ~49% of eligible hospitals have ≥2 each of S, A, O by SIP.
  2. HC–SIP gap is largest for anaesthesia — multi-site anesthesiologists inflate apparent SIP coverage but don’t add team capacity at any single hospital.
  3. Workforce predicts output — districts with more SAO-capable hospitals consistently show higher BW procedure rates across all five criteria.
  4. Type C hospitals carry most volume — JKN referral structure funnels most BW cases to mid-tier general hospitals, not tertiary centres.
  5. Remote regions face a double gap — lowest SAO capability AND lowest spatial population coverage (Papua, Maluku, parts of Kalimantan).
  6. Patient drift is substantial and stable — ~1 in 3 BW procedures delivered outside the patient’s home district; low spatial coverage predicts higher outflow.
  7. Population burden matters independently — districts with high population-per-hospital ratios export patients even when geographic catchment looks adequate.

Limitations

  • Hospital-level linkage not possible — SIRS/DREAMS and BPJS use different hospital ID systems; all linkage is at district level.
  • DREAMS is a cross-sectional snapshot — workforce data reflects 2025; rate trends span 2016–2024. Workforce changes over this period are not captured.
  • BPJS rates reflect delivery, not need — districts that export patients record lower rates than true surgical incidence suggests.
  • ArcGIS service areas are static — 2-hour drive-time polygons do not account for road conditions, seasonality, or actual patient behaviour.
  • 2018 excluded — incomplete year of BPJS visit tracking.

Thank you

Full report: Master_Report.html — figures labelled Fig 01–21 correspond directly to slide references.

Analysis pipeline: 8.1 Workforce (SIRS × DREAMS) → 8.2 Volume (BPJS claims) → 8.3 Timely Access (this report + ArcGIS integration)

ARC8 Global Surgery Institute · ARC Institute · 2026