Health system capacity stress

What this driver is

Health-system capacity stress is the continuing adequacy of beds, workforce, oxygen, and referral — whether a district hospital can take the next severe case. It is not COVID-19 pandemic. COVID-19 used this capacity.

It is not pandemic preparedness (IHR SPAR), not access to essential medicines, and not Ebola outbreak response. Beds can sit empty of staff. SPAR can look fine while ICU oxygen fails.

The meter is World Bank WLD SH.MED.BEDS.ZS (hospital beds per 1,000) divided by 10: 0.2861 (2014), 0.3043 (2019), 0.3132 (2020), 0.3294 (2021). A global average hides national collapse. Nurses and physicians per 1,000 are neighbouring official series (nurses, physicians). This page does not invent a unique occupancy integer.

Mechanism

Capacity moves welfare through named channels.

First, physical plant. Beds, ICU, oxygen, power. COVID-19 2020–21 made occupancy a daily political number. A peak occupancy is not a unique Factrail coefficient here.

Second, workforce. Infection, burnout, and strike. Ebola 2014 and COVID-19 both killed health workers. Those deaths are named in WHO/MoH sitreps. This page does not invent a unique worker-death total.

Third, money and absenteeism. Unpaid wages close wards. [Health-system capacity] itself is not a fiscal driver; public-debt neighbours exist elsewhere.

Fourth, the beds meter. Global beds per 1,000 drifted 2.86 (2014) → 3.29 (2021). That is not “more capacity in the places that needed it.” WHO has treated density as a poor proxy for surge. The proxy is disclosed.

Fifth, elective-care displacement. When surge fills beds, surgery and obstetric care wait. Maternal rollback and NCD care are neighbouring victims.

Timing is hours for an ICU bed and years for a nursing cohort. A later OECD/WHO density point would be a new data row.

Evidence base

World Bank SH.MED.BEDS.ZS (https://api.worldbank.org/v2/country/WLD/indicator/SH.MED.BEDS.ZS?format=json&date=2014:2021&per_page=30 (opens in a new tab)). WHO National Health Workforce Accounts are the named workforce instrument. Live COVID Facts on preparedness (hospital-adjacent) stay on their drivers.

Evidence is strongest for an official density point. It is weaker for any unique 2020 occupancy coefficient. It is weakest where a rising global bed average is asked to prove surge succeeded.

Effect on welfare indicators

• Hospital beds per 1,000: identity channel. Horizon: years. Confidence: high for the series, low for surge quality.
• Under-five mortality: direction higher deaths when paediatric wards fail. Horizon: weeks to years. Confidence: medium.
• UHC service coverage: direction lower effective coverage when capacity is exhausted. Confidence: medium.

Interactions and contested points

Capacity interacts with COVID-19 pandemic, pandemic preparedness, and Ebola outbreak response. Reinforcement is obvious when a wave and a staff strike hit the same hospital. Offsetting is the claim that field hospitals ended the stress. A tent is not a nurse.

The strongest competing view is that beds-per-1,000 is the wrong meter (staff, oxygen, primary care). That objection is named. The defence is a disclosed official series plus workforce indicators as neighbours.

Established: beds 2.86 (2014) → 3.29 (2021) per 1,000; distinction from COVID event. A later NHWA point would be a new fact.

WHO National Health Workforce Accounts and OECD health-care resources are the named staff series. Global bed density rose while many low-income States still sit under 1 bed per 1,000 — an average conceals that. COVID-19 field hospitals and oxygen plants were dated 2020–21 capital acts; they are not a nurse. Ebola outbreak response already stores health-worker infection as an outbreak channel. This node stores whether the plant and payroll exist when the next wave arrives.

WHO and World Bank series used on this driver are official published points. They are transformed only by the disclosed divisor in the catalog. Contested modelled totals (excess deaths, DALYs, dollars pledged) remain named edge instruments. Affiliates without a public-complete people.xml dossier stay plain text. The WHO Director-General is not minted. Pipeline V2, autocreate, and modules/news stay off. Person.score is not written. Ukraine-war, nagorno-karabakh-war, and eastern-drc-m23-rdf are not reminted. Territorial-revisionism (1918–45) is not this health node.

Live Facts already attached to companion health drivers (WHO pandemic characterization, PHEIC end, COVAX Accra, US WHO-withdrawal notices) stay on those Facts. Articles explain; news does not mint people. A later official series year is a new data point when the provider publishes it. Missing years stay missing. Correlation between a global average and a district outcome is not causation.