BRENNEN

Industries / Health / Healthcare

Nurse staffing

Nurse staffing is the business of supplying hospitals and clinics with nurses they do not employ themselves: travel nurses on contracts of weeks or months, per-diem nurses booked shift by shift, and contract nurses placed through agencies, registries and apps.

1879

In November 1879 the Boston Medical Library opened its Directory for Nurses, started by the physicians Frederick C. Shattuck and Charles P. Putnam. It registered nurses it could recommend and gave their addresses and qualifications to families and doctors who needed one, for a fee from both sides. The library's historian, John W. Farlow, wrote in 1918 that it was "believed to be the first of its kind ever established".

archive.org

39

Milestones achieved across 6 domains.

10

To watch for: announced, not yet confirmed outside.

12

Open deliverables. Nobody has shipped them.

In one minute

Where it started

Doctors' directories in Boston, from 1879, and Philadelphia, from 1882, matched private-duty nurses with families and doctors for a fee. Once telephones spread, nurses' clubs and commercial registries multiplied, and both directories closed, in 1914 and 1936. Travel nursing began in the United States in the 1970s, when agencies started sending nurses to short-staffed hospitals for fixed stints, typically 13 weeks. Hospitals later bought contract nurses through ordering software, and from the mid-2010s nurses could claim single shifts on apps.

Where it stands

By the figures of Staffing Industry Analysts (SIA), a research firm, US travel nurse revenue rose from $8.7 billion in 2019 to $44.6 billion in 2022, then fell to a projected $14.2 billion in 2025, and apps and other platforms now hold more than a fifth of the US market for temporary health care staff. Governments have pushed back on the cost. England capped agency rates in 2015, Quebec set deadlines to end agency use in its public health network by October 2026, and Japan lets agencies supply nurses to hospitals only for leave cover, temp-to-hire placements and, since 2021, remote areas. Studies disagree on whether agency nurses are worse for patients, and an outside rating found that only one of four US nursing apps could show its workers earned the minimum wage.

What is open

Most of the new software is known only from its makers' accounts. AI agents now sold to check licences, read orders, text nurses about open shifts and fill per-diem shifts publish no error rates, and nobody outside has tested them on cases with known answers. Public records that would show where nurses were short and what agencies were paid sit unscored or unpublished: hospitals' own one-week forecasts of staffing shortages, and Quebec's agency hours by region. Tests with made-up nurses, mock websites and simulated markets could be built now with available tools.

Domains

AchievedWatch forOpen

Advances to build with

2 research notes

Open in nurse staffing

3 of 12 open deliverables

Open · Matching nurses to shifts

Shortages a week ahead

A public scoreboard of one-week-ahead forecasts of hospital staffing shortages, the signal that says where contract nurses will be needed. In US federal hospital reporting from 2020 to May 2024, each hospital answered two questions every day: whether it had a critical staffing shortage that day, and whether it expected one within a week. Score the hospitals' own expectations, state by state, against what the same states reported seven days later. Then test whether forecasts built only from public data, such as COVID-19 admissions, beds in use and the season, predict the share of hospitals short of staff better.

Built with what exists

The US health department's state time series of hospital capacity (healthdata.gov dataset g62h-syeh), which holds each day's count of hospitals answering yes, no or not reported to both questions; Python with pandas; a seasonal-naive forecast and a logistic regression as baselines; pretrained forecasting models such as Amazon's Chronos-2 and Google's TimesFM; a public notebook.

Credit · unclaimed

Open · Checking licences and credentials

Agency personnel files tested

1,000 made-up personnel files for made-up travel and per diem nurses, as PDFs and phone photos: licence printouts, life-support cards (BLS, ACLS), tuberculosis and immunization records, skills checklists, competency test results and background-check summaries. Plant known defects in most of them: an expired card, a licence for the wrong state, a name spelled differently between documents, a missing page, a competency test for the wrong specialty, a background check that is out of date. Score AI document tools and agents on whether they mark each file complete or incomplete under the three parts of the Joint Commission's file measure (job-appropriate credentials, evidence of current competency, background check), and on which defect they name.

Built with what exists

Python with the Faker library for made-up people; ReportLab to render documents; a phone camera; AI models such as Claude and Gemini, and open models; Google Document AI or Amazon Textract as plain extraction baselines; Gemini Decision 1, which returns probabilities for up to 1,000 fixed-answer questions in one request, for the checklist items (see the Note "A thousand questions, one pass").

Credit · unclaimed

Open · Scheduling

Float pool sizing tool

An open-source simulator that takes a hospital's daily inpatient census, a staffing target in nurse hours per patient day, and the hourly cost of each kind of cover (permanent staff, overtime, a float pool, per-diem nurses, internal travel nurses and agency nurses), and reports the float pool size and mix that keep staffing shortfalls under a set level at the least cost. Run it on public census series for at least 50 US hospitals.

Built with what exists

The US health department's facility-level dataset of weekly inpatient beds in use at each reporting hospital from 2020 to May 2024 (healthdata.gov dataset anag-cw7u); SimPy, a Python library for simulation; Google OR-Tools for the cost minimization; hourly costs set by the user.

Credit · unclaimed

Where it sits

Broader

Healthcare