Industries / Health / Healthcare / Nurse staffing
Matching nurses to shifts
Matching is finding a nurse with the right licence, specialty and dates for a hospital's open shift or contract and getting both sides to agree, work done first by doctors' registries, then by agency recruiters, and now partly by software.
6
Milestones achieved, from 1882 to 2025.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1882
Philadelphia opens a directory
The College of Physicians of Philadelphia opened its Directory for Nurses on 14 May 1882, a referral office where doctors and families could hire qualified nurses. It followed Boston's: the archive's finding aid says S. Weir Mitchell's notes on setting it up were probably written in reply to letters from Edwin Brigham, the Boston directory's registrar. From 1882 to 1936 it registered more than 7,500 nurses and filled more than 150,000 calls, but in one month of the 1918 influenza epidemic it received 2,955 calls for nurses and filled 197.
LessonA matching office cannot fill a surge with nurses who do not exist; in that month of 1918, calls outnumbered fills fifteen to one.
Credit
S. Weir Mitchell, William W. Keen and Albert H. Smith, the founding committee
1914
Doctors' registries lose out
The Boston directory had 412 nurses registered in 1881, and in 1886 it brought the library more money than its members' dues. Its receipts fell as nurses' clubs and other registries multiplied, and it closed on 31 December 1914. Philadelphia's directory, facing the same competition and a threat to the College's tax-exempt status, stopped on 31 March 1936.
LessonFarlow wrote that once everyone had a telephone, registries and nurses' clubs sprang up everywhere, and the doctors' offices lost the near-monopoly that had made them pay.
Credit
John W. Farlow, the library's historian (1918); the College of Physicians of Philadelphia
1970s
Travel nursing begins
During a nursing shortage in the 1970s, nurses from northern US states began spending winters at Florida hospitals whose patient numbers could double in the season, informally and without contracts. Agencies then began recruiting nurses for fixed stints away from home; the most-told account has a businessman, Bruce Male, advertising in the Boston Globe for "traveling nurses" and sending the first group to New Orleans hospitals for Mardi Gras. Cross Country, which bought the travel agency TravCorps in December 1999, told investors in 2001 that it had worked in travel nurse staffing "since the 1970s".
Credit
Bruce Male, by travel nurses' own histories
2001
Recruiters work from a database
In its 2001 prospectus Cross Country, then the largest US healthcare staffing company by revenue, with about 80% of its revenue from travel nursing, set out the method that still dominates. Account managers keep a database of hospitals' open assignments, recruiters match them to each nurse's experience, skills and preferred places, and an account manager checks the match before sending the nurse's file to the hospital. Assignments typically ran 13 weeks, and the company held about 2,900 leased apartments for nurses on assignment.
Credit
Cross Country TravCorps
2022
A surge, then a slump
By SIA's figures, US travel nurse revenue rose from $8.7 billion in 2019 to $44.6 billion in 2022 as hospitals staffed COVID-19 surges. It then fell three years running, to a projected $14.2 billion in 2025, which SIA puts down to lower hospital use and falling bill rates.
LessonThe surge did not last; by 2025 revenue was under a third of the 2022 peak, though still above 2019.
Credit
SIA
2025
Platforms take a fifth
SIA's 2025 market assessment estimated that platform-based staffing, in which clinicians book work through an app or website rather than through a recruiter, accounts for more than 20% of the US healthcare temporary staffing market. It also found per diem and short local contracts holding steadier than travel work, which it expected to fall 12% in 2025.
Credit
SIA (Crystal Fullilove)
Watch for
AI in agency recruiting
AMN Healthcare says it has seen "positive early results" from automation and AI in recruiting, credentialing and contact with candidates, and that AMN Passport, its app where clinicians store credentials, see jobs and message recruiters, has more than 340,000 registered users. Against this, SIA's 2026 survey found that even at the 75th percentile, travel nurse firms filled no orders without a human recruiter.
Achieved whenAMN or an outside party publishes before-and-after figures, such as days from order to start or orders filled per recruiter, with the method.
Claimed by
AMN Healthcare, annual report filed February 2026
Open
01
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.
The briefCredit · unclaimed02
Registry ledgers into data
An open series of the yearly, and where printed the monthly, counts of nurses registered, calls received and calls filled at the Boston and Philadelphia doctors' directories from 1879 to 1936, transcribed from their printed annual reports with no nurses' names, so that the 1918 influenza surge and the arrival of rival registries can be studied as a matching market.
The briefCredit · unclaimed
Open 01 · the brief
Shortages a week ahead
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
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.
Why it is still open
The answers are published as daily counts by state, but searches found no study that scores them; the forecasting work found is about beds and patient numbers, such as Cedars-Sinai's 2020 model for its own hospitals. When the data first came out, NPR noted that each hospital decided for itself what counted as a critical shortage, which is a reason to measure how well the answers predict, not to assume it.
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.
First step
Download the state file, compute for one state the share of reporting hospitals that expected a shortage on each day, and set it beside the share that reported one seven days later.
Done looks like
For every state and every week from 2021 to May 2024, the hospitals' own one-week expectation and at least three forecasts, each scored by its error in percentage points against the share of hospitals reporting a shortage seven days later, with the Delta and Omicron waves of 2021 and 2022 reported separately and every number rebuilt from the public file by one script.
Open 02 · the brief
Registry ledgers into data
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open series of the yearly, and where printed the monthly, counts of nurses registered, calls received and calls filled at the Boston and Philadelphia doctors' directories from 1879 to 1936, transcribed from their printed annual reports with no nurses' names, so that the 1918 influenza surge and the arrival of rival registries can be studied as a matching market.
Why it is still open
The published histories give scattered figures only: 412 nurses registered and an average of 74 placed a month in 1881 in Boston, lifetime totals for Philadelphia, and one month of 1918. The reports sit in the Countway Library at Harvard and at the College of Physicians of Philadelphia, and a search of the Internet Archive's catalog found no digitized series.
Built with what exists
Scans requested from the two archives; an AI model that reads printed tables, checked by double entry; a spreadsheet; GitHub for the data and method.
First step
Ask the Countway Library and the College of Physicians which directory annual reports they hold and what scans would cost.
Done looks like
Every surviving year from 1880 to 1935 filled for both directories, each figure keyed twice by different people with disagreements resolved and logged, and the 1918 months set beside the same months of 1917 and 1919.