Zombie Jobs: 163 Days Later

01 · Headline Numbers

The shortage stopped being a number. It became a duration.

Every figure below is derived from one thing nobody else measures: the same US radiology opening, watched every day from the moment it first appeared until it came down — or didn't.

4,319

Distinct radiologist roles advertised in the US right now, after removing 2.8× average re-post inflation

52.4%

Of them have been advertised 90 days or longer — 2,262 chairs

804

Have never come down since tracking began on Mar 9, 2026 — 163 days and counting

82.0%

Of long-stuck roles never changed the advertised salary. Median change across 694 roles: $0

33×

Times a single emergency-radiology role has been re-listed since March. It is still open

1,108

Role-years of unfilled reading capacity advertised on the US market today

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

What a job posting can and cannot tell you

A posting disappearing does not prove a hire. That is precisely why this report is built on the roles that did not disappear. A chair advertised for the 33rd time is unambiguous: nobody sat down in it. Full limitations in Section 12.

02 · Find Your Chair

Before the national tables — here is your own number.

Pick your state, your subspecialty and your group type. No typing, no upload, no spreadsheet: every figure is already in the data below. Everything after this section is the market-wide version of the same question.

State

Subspecialty

Group type

—

Your selection

—

US national average

52.4%

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9 – August 20, 2026, de-duplicated to 13,555 distinct roles · Analysis: xAID.ai

This section is interactive in the online edition

The web version of this report at RadBoard.io carries a live lookup here: choose a state, a subspecialty and a group type, and it returns the stuck rate, the median days advertised, the median advertised maximum and the average number of re-listings for that exact cell — 468 combinations in all. In this printed edition the same cuts appear as full tables in Sections 07 through 10.

If your cell is above 52.4%

That is the national share of live US radiology roles advertised 90 days or longer. A cell above it means chairs like yours sit longer than the average American radiology chair — and the sections that follow show that neither pay, nor remote work, nor a disclosed salary reliably moves that number.

About xAID

Ready-to-sign CT reports, powered by AI with humans in the loop.

xAID helps U.S. radiology groups scale CT volume without hiring additional radiologists. If a chair on the tables in this report has been open since spring, this is the lever that does not depend on a candidate existing.

100%

Of reports are validated by our European radiologists before they reach yours

<0.1%

Discrepancy rate

Full report

Templated and ready to sign — not a list of flags

24/7

Coverage

The difference

Unlike competitors who only flag pathologies, xAID delivers a full report — so your radiologists review instead of dictating. That lifts CT reporting capacity by up to 50%, not just detection rates. If you are turning away volume because you cannot hire, that is the revenue this recaptures.

And you get the full report at the price of narrow AI — the same budget line as a single tool that only flags, so the comparison is like-for-like.

Visit xAID.ai →

03 · The May Cohort, Revisited

We went back to the 2,111 roles that were already stuck in May.

The May 2026 edition of this report closed with one row it could not fill: what happens to an opening past 90 days. The tracking window simply ended first. 86 days later, here is the answer.

2,111

Roles advertised 60+ days as of May 26, 2026

1,226

Still being advertised today — 58.1% never filled

163

Median days on the market for the survivors — every day since counting began

The May 2026 cohort	Roles	Share
Advertised 60+ days as of May 26, 2026	2,111	100%
Still advertised on August 20, 2026	1,226	58.1%
Came off the market during the 86-day gap	885	41.9%
Average times a survivor has been re-listed	4.1×	—

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai · Cohort defined May 26, 2026; status re-checked August 20, 2026

Why this table does not exist anywhere else

Job boards publish a snapshot: how many roles are open today. Recruiting benchmarks publish an average time-to-fill from the employer's side. Neither can follow one advertised chair across five months. RadBoard.io is the only longitudinal record of the US radiology job market, and this is the first time it has been long enough to answer the question.

04 · The Calcification Curve

The number of open chairs barely moved. The share that cannot close nearly doubled.

Between mid-June and today the live inventory is flat. What changed is its age. This is the single most important chart in the report: the market is not growing, it is hardening.

Mar 15, 2026

window too short

Mar 31, 2026

window too short

Apr 15, 2026

window too short

Apr 30, 2026

window too short

May 15, 2026

window too short

May 31, 2026

window too short

Jun 15, 2026

40.6%

Jun 30, 2026

38.7%

Jul 15, 2026

51.9%

Jul 31, 2026

51.0%

Aug 20, 2026

52.4%

Share of live roles advertised 90+ days. Points before Jun 15 are blank because the tracking window was not yet 90 days old — the same window limit the May report ran into.

As of	Live roles	Advertised 90+ days	Share stuck
Mar 15, 2026	3,659	—	not yet observable
Mar 31, 2026	3,704	—	not yet observable
Apr 15, 2026	3,857	—	not yet observable
Apr 30, 2026	4,967	—	not yet observable
May 15, 2026	4,886	—	not yet observable
May 31, 2026	4,618	—	not yet observable
Jun 15, 2026	4,411	1,789	40.6%
Jun 30, 2026	4,360	1,686	38.7%
Jul 15, 2026	4,188	2,172	51.9%
Jul 31, 2026	4,545	2,317	51.0%
Aug 20, 2026	4,319	2,262	52.4%

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

Read it this way

If the market were simply busy, live inventory would rise and the age mix would stay flat. Instead inventory is flat at roughly 4,319 and the stuck share went from 40.6% in June to 52.4% today. New roles are still arriving. The old ones just aren't leaving.

05 · The Frozen Offer

82.0% of long-stuck roles never changed the salary. Not by a dollar.

For 694 roles advertised past 120 days we can line up the first ad against the most recent one and compare the number. Five months of no candidate produced almost no repricing.

Salary unchanged

569 · 82.0%

Salary raised

68 · 9.8%

Salary cut

57 · 8.2%

Median change across all 694 comparable roles: $0. Mean: $-1,627. Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

The uncomfortable reading

Two explanations fit. Either these groups have concluded that money is not the binding constraint — or nobody re-opened the requisition after the first posting went up. Both are worth knowing before the 2027 budget is set. Neither is consistent with a market where a bigger number solves the problem.

And the stuck chairs are not the cheap ones

Days advertised	Live roles	Median advertised max	Remote offered	Avg. re-listings
Under 30 days	1,102	$700,000	27.8%	1.3×
30–59 days	432	$650,000	30.6%	2.0×
60–89 days	523	$640,000	22.9%	2.4×
90–119 days	388	$700,000	26.8%	3.4×
120–162 days	1,070	$650,000	25.1%	3.1×
163+ days	804	$650,000	16.9%	5.1×

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai · Advertised maximum of the posted range, $150K–$2.5M, latest ad per role

A chair open since day one advertises $650,000. A chair posted this month advertises $700,000. The whole spread is $50,000 on a $650,000 salary — far too small to explain a five-month vacancy.

06 · The Levers

Every recruiting lever in the market shows up more often on the chairs that stay empty.

Remote work, disclosed pay, partnership track, loan forgiveness, visa sponsorship. Each is supposed to make a job easier to fill. Sorted by how often they appear on stuck roles versus everything else, five of the six point the wrong way.

Lever present in the ad	Roles with it	Stuck 90+ days	Roles without it	Stuck 90+ days	Gap
Loan forgiveness offered	654	60.7%	3,665	50.9%	+9.8 pts
Visa sponsorship offered	231	61.5%	4,088	51.9%	+9.6 pts
Partnership track mentioned	1,284	56.4%	3,035	50.7%	+5.7 pts
Salary disclosed in the ad	2,344	54.3%	1,975	50.1%	+4.2 pts
AI tools mentioned	771	49.2%	3,548	53.1%	-3.9 pts
Remote work offered	1,067	47.7%	3,252	53.9%	-6.2 pts

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

This is association, not cause — and the direction still matters

Loan forgiveness does not make a job harder to fill. It appears on hard-to-fill jobs, because that is when a group reaches for it. Read the table as a diagnostic, not a prescription: these levers are markers of difficulty, and the fact that deploying them has not cleared the chair after five months is the finding. One line runs the other way — roles that mention AI tooling in the workday sit stuck 3.9 points less often than roles that don't. Association again. Worth a question in your next candidate debrief.

07 · Where the Chairs Stay Empty

The shortage is local. It is also not where the money is thin.

States ranked by the share of their live radiologist roles that have been advertised 90 days or longer. Only states with at least 40 live roles are ranked; all 51 appear in the appendix on the next page with their sample size shown.

Pennsylvania

78.3%

Louisiana

68.5%

South Carolina

65.5%

Oregon

64.7%

Tennessee

64.2%

Michigan

63.8%

Virginia

63.3%

Wisconsin

62.6%

California

59.8%

Minnesota

58.8%

Kentucky

56.0%

Massachusetts

53.7%

Top 12 of 29 ranked states. Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

State	Live roles	90+ days	Share stuck	Since day one	Median advertised max
Pennsylvania	207	162	78.3%	71	$700,000
Louisiana	54	37	68.5%	13	$675,000
South Carolina	87	57	65.5%	18	$700,000
Oregon	51	33	64.7%	9	$571,000
Tennessee	95	61	64.2%	28	$699,000
Michigan	58	37	63.8%	12	$900,000
Virginia	90	57	63.3%	25	$650,000
Wisconsin	91	57	62.6%	31	$700,000
California	244	146	59.8%	59	$600,000
Minnesota	68	40	58.8%	20	$703,584
Kentucky	75	42	56.0%	16	$650,000
Massachusetts	123	66	53.7%	29	$549,000
Colorado	86	45	52.3%	15	$635,000
New Jersey	101	52	51.5%	22	$700,000
Florida	409	209	51.1%	50	$699,000
Arizona	96	49	51.0%	14	$600,000
Missouri	63	32	50.8%	11	$700,000
Illinois	166	83	50.0%	31	$700,000
Connecticut	52	26	50.0%	8	$725,000
Georgia	128	63	49.2%	25	$660,000
New York	214	104	48.6%	34	$550,000
Nebraska	50	24	48.0%	2	$549,000
Texas	267	128	47.9%	39	$750,000
Maryland	74	35	47.3%	12	$550,000
North Carolina	202	89	44.1%	31	$850,000
Washington	107	47	43.9%	15	$616,032
Ohio	222	97	43.7%	36	$730,500
Indiana	114	49	43.0%	18	$675,000
West Virginia	49	18	36.7%	9	$600,000

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai · states with 40+ live roles

Pennsylvania is the counter-example that kills the pay theory

162 of Pennsylvania's 207 live radiologist roles have been advertised past 90 days — the worst rate of any state with real volume — at a median advertised maximum of $700,000. That is above the national median of $650,000. Whatever is holding those chairs open, it is not the number on the page.

07b · Appendix — Every State

All 51 jurisdictions with a live radiologist opening

Sorted by share stuck. Rows shaded grey have fewer than 40 live roles and are shown for completeness only — the percentage is not stable at that sample size and should not be quoted.

#	State	Live roles	90+ days	Share stuck	Since day one	Avg. re-listings	Median advertised max
1	Wyoming ‡	2	2	100.0%	0	1.0×	$699,000
2	South Dakota ‡	16	13	81.2%	7	1.9×	$620,800
3	Pennsylvania	207	162	78.3%	71	3.0×	$700,000
4	Montana ‡	32	25	78.1%	4	2.3×	$639,634
5	Oklahoma ‡	17	12	70.6%	2	3.4×	$900,000
6	Utah ‡	17	12	70.6%	2	2.9×	$725,000
7	Louisiana	54	37	68.5%	13	5.3×	$675,000
8	Vermont ‡	19	13	68.4%	5	1.9×	$516,600
9	South Carolina	87	57	65.5%	18	2.4×	$700,000
10	Oregon	51	33	64.7%	9	3.5×	$571,000
11	Tennessee	95	61	64.2%	28	3.9×	$699,000
12	Michigan	58	37	63.8%	12	2.4×	$900,000
13	Virginia	90	57	63.3%	25	2.5×	$650,000
14	Wisconsin	91	57	62.6%	31	3.0×	$700,000
15	Maine ‡	23	14	60.9%	4	2.6×	$625,000
16	Iowa ‡	25	15	60.0%	4	2.0×	$600,000
17	California	244	146	59.8%	59	2.9×	$600,000
18	Minnesota	68	40	58.8%	20	2.1×	$703,584
19	New Hampshire ‡	26	15	57.7%	9	4.6×	$606,000
20	Delaware ‡	23	13	56.5%	5	3.5×	$635,045
21	Kentucky	75	42	56.0%	16	3.8×	$650,000
22	Nevada ‡	20	11	55.0%	2	2.0×	$650,000
23	Mississippi ‡	11	6	54.5%	3	2.7×	$800,000
24	Hawaii ‡	24	13	54.2%	5	2.2×	$636,264
25	Arkansas ‡	13	7	53.8%	2	1.4×	$799,000
26	Massachusetts	123	66	53.7%	29	2.6×	$549,000
27	Alabama ‡	19	10	52.6%	4	1.6×	$699,000
28	Colorado	86	45	52.3%	15	3.2×	$635,000
29	New Jersey	101	52	51.5%	22	3.1×	$700,000
30	Florida	409	209	51.1%	50	3.6×	$699,000
31	Arizona	96	49	51.0%	14	4.2×	$600,000
32	Missouri	63	32	50.8%	11	2.8×	$700,000
33	Illinois	166	83	50.0%	31	2.9×	$700,000
34	Connecticut	52	26	50.0%	8	2.9×	$725,000
35	Georgia	128	63	49.2%	25	2.7×	$660,000
36	New York	214	104	48.6%	34	2.2×	$550,000
37	Nebraska	50	24	48.0%	2	1.2×	$549,000
38	Texas	267	128	47.9%	39	3.4×	$750,000
39	Maryland	74	35	47.3%	12	2.5×	$550,000
40	North Carolina	202	89	44.1%	31	2.2×	$850,000
41	North Dakota ‡	25	11	44.0%	3	1.8×	$587,607
42	Washington	107	47	43.9%	15	2.1×	$616,032
43	Ohio	222	97	43.7%	36	2.4×	$730,500
44	Indiana	114	49	43.0%	18	3.1×	$675,000
45	Alaska ‡	15	6	40.0%	2	1.7×	$800,000
46	Rhode Island ‡	8	3	37.5%	1	1.2×	$500,000
47	West Virginia	49	18	36.7%	9	1.7×	$600,000
48	New Mexico ‡	24	8	33.3%	3	2.7×	$550,000
49	Idaho ‡	22	7	31.8%	4	4.3×	$700,000
50	Kansas ‡	39	12	30.8%	3	1.4×	$400,000
51	District of Columbia ‡	17	3	17.6%	0	1.3×	$750,000

‡ fewer than 40 live roles — sample too small to rank. Puerto Rico had no live radiologist role at snapshot and is omitted. Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

08 · By Subspecialty

The hardest chairs to fill are the ones that read the most volume.

Subspecialty assigned from the job title, since the structured field is populated on roughly a fifth of postings. Roles carrying more than one label are counted in each.

Subspecialty	Live roles	90+ days	Share stuck	Median advertised max	Avg. re-listings
Pediatric	207	131	63.3%	$600,000	2.7×
Body / Abdominal	237	136	57.4%	$650,000	2.8×
Chest / Cardiothoracic	121	68	56.2%	$600,000	2.6×
General / not specified	1,885	1,055	56.0%	$650,000	2.8×
Nuclear / PET	135	75	55.6%	$650,000	3.2×
MSK	194	96	49.5%	$700,000	3.1×
Neuroradiology	181	88	48.6%	$700,000	2.4×
Emergency	106	50	47.2%	$650,000	2.8×
Teleradiology	248	116	46.8%	$700,000	2.3×
Interventional	647	290	44.8%	$675,000	3.1×
Breast / Mammography	466	206	44.2%	$650,000	2.9×

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai · subspecialty parsed from job title; 40+ live roles required

Body, chest and general are 57% stuck — and they are the CT workhorses

Pediatric sits at the top on a small, structurally scarce pool. But the volume story is underneath it: body/abdominal, chest and general radiology together account for 2,243 live roles, and more than half of each has been advertised for three months. These are the reads that arrive every day whether the chair is filled or not.

09 · By Employer Type

Same market, same candidates, twenty-one points of difference.

Employer type	Live roles	Share stuck 90+ d	Median advertised max	Remote offered	AI mentioned	Avg. re-listings
Academic medical center	677	59.1%	$557,000	8.1%	14.2%	2.5×
Hospital system	1,244	57.9%	$650,000	12.6%	14.0%	2.9×
Independent practice	1,133	53.6%	$700,000	29.7%	21.5%	2.1×
PE-backed platform	610	45.2%	$700,000	52.8%	35.1%	5.5×
Staffing / locums agency	552	39.3%	$699,000	34.4%	7.8%	1.8×
Government (VA, DoD, IHS)	99	38.4%	$400,000	6.1%	0.0%	1.5×

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

Academic medical center

59.1%

Hospital system

57.9%

Independent practice

53.6%

PE-backed platform

45.2%

Staffing / locums agency

39.3%

Government (VA, DoD, IHS)

38.4%

Share of live roles advertised 90+ days, by employer type. Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

It is not the salary. Look at the workday columns.

Academic medical centres advertise $557,000 and leave 59.1% of their chairs open past three months. PE-backed platforms advertise $700,000 — $143,000 more — and sit at 45.2%. But the two columns that separate them are not money: remote work at 52.8% versus 8.1%, and AI tooling named in the ad at 35.1% versus 14.2%. The gap is what the day looks like, not what it pays.

10 · The Re-Post Ledger

217 chairs have been advertised ten times or more.

The same opening, re-listed again and again. This is the cleanest evidence in the dataset that a role is genuinely unfilled: a posting that disappears is ambiguous, a posting that keeps coming back is not.

Listed once

2,101 · 48.6%

2–4 times

1,455 · 33.7%

5–9 times

546 · 12.6%

10–19 times

203 · 4.7%

20+ times

14 · 0.3%

Distribution of re-listings across 4,319 live roles. Median 2, mean 2.8, maximum 33. Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai

33×

Most re-listed live role in the country — an emergency radiologist post, first seen Mar 9, 2026, still open

217

Live roles re-listed 10+ times

5.1×

Average re-listings for a chair open since day one, against 1.3× for one posted this month

Why the national "open jobs" number is inflated

Across the full window 29,541 physician postings collapse to 13,555 distinct roles. Any count taken straight off aggregators needs roughly that deflation before it means anything — which is the correction we published in May and the reason the headline count here is 4,319 rather than a number twice as large.

11 · What 1,108 Role-Years Means

The chairs currently advertised in America represent 1,108 radiologist-years of reading nobody is doing.

4,319 live roles, each advertised for an average of 94 days, sum to 404,402 advertised days — 13,285 role-months, or 1,108 role-years. That is the measured part, and it comes entirely from RadBoard.io.

107

Days the median live chair has already been advertised

130

Days the AAPPR benchmark reports it takes to fill a full-time radiology position

1,108

Role-years of advertised, unfilled reading capacity on the market today

Source: RadBoard.io — live US radiology job database, 34,522 postings tracked Mar 9, 2026–Aug 20, 2026 · Analysis: xAID.ai · AAPPR figure: Association for Advancing Physician and Provider Recruitment Benchmarking Report, as reported in trade coverage — see Section 12

The conversion we are deliberately not making

Multiplying role-years by a median wRVU load and a dollar-per-wRVU rate produces a very large number. We are not printing it. The productivity and collections benchmarks that would anchor it sit behind paid industry reports, and the widely circulated "$150,000–$250,000 per month of physician vacancy" figure has no traceable primary source. Every number in this report traces to data we collected ourselves. That one would not.

What the duration actually forecloses

A chair that has been advertised 163 days at an unchanged salary, in a state paying above the national median, is not waiting on a better offer. The 2026 fellowship match produced 721 fellows against thousands of open searches; the candidate the plan assumes was not trained. Recruiting harder does not create one. What remains under a leadership team's control is a different variable entirely: how many studies the radiologists already on payroll can clear in a day.

12 · Method & Sources

How this was built, and what it cannot tell you

Data

Every figure comes from the RadBoard.io live job database: 34,522 US radiology postings collected from 20+ sources, each row carrying the timestamp it was first seen, the timestamp it was last seen, and the timestamp it was deactivated. Those three fields are what make duration measurable. Snapshot taken August 20, 2026; observation window Mar 9, 2026 to Aug 20, 2026, 163 days.

From postings to roles

Non-physician listings were removed by title — technologists, medical physicists, dosimetrists, NP/PA, administrative and trainee posts — leaving 29,541 physician postings. Subspecialty society boards that re-post on behalf of members under generic titles, and listings with no named employer, were excluded. The remainder was de-duplicated to 13,555 distinct roles on employer + title + city + state. A role's age runs from the first appearance of any of its postings to today if live, or to its last appearance if not.

Definitions

Live role — at least one posting for that role was active at snapshot. Advertised 90+ days — first appearance was 90 or more days before snapshot and the role is still live. Advertised max — the top of the posted salary range, filtered to $150,000–$2,500,000, taken from the most recent posting for that role. Salary comparisons over time use the first and latest postings of the same role where both carry a figure.

Known limits — stated plainly

A posting is not a seat. We measure a continuously advertised opening, not headcount. A group with five chairs under one job title collapses into one role here, which makes every count in this report a conservative floor rather than a ceiling.

The window is a floor, not a ceiling. The 804 roles live since Mar 9, 2026 were already being advertised when tracking began. 163 days is the minimum we can prove, not their true age.

Disappearing is not filling. A posting can come down because it was filled, pulled, expired, or rotated to another source. That ambiguity is exactly why this report is built on roles that stayed up or kept returning, and why the re-post ledger in Section 10 carries as much weight as the age tables.

Advertised is not paid. Salary figures are the top of a posted range, not realised compensation. Their value is comparability: every employer is measured with the same ruler.

Association is not cause. Section 06 shows which levers appear alongside stuck roles. It does not show that any lever caused, or would cure, a vacancy.

External figures cited

One number in this report is not ours: the AAPPR benchmark of approximately 130 days to fill a full-time radiology position, cited here from trade coverage of the Association for Advancing Physician and Provider Recruitment Benchmarking Report. The 2026 fellowship output figure (721 matched fellows) comes from the NRMP Results and Data 2026 specialties matching service. Prior RadBoard findings referenced in Section 03 were published in the US Radiology Job Market Dynamics Report, May 2026, at RadBoard.io, and covered by Radiology Business, AuntMinnie and The Imaging Wire.

Reproducibility

Every table recomputes from the snapshot; no figure in this document was entered by hand. The next edition extends the window and re-checks this cohort the same way. Questions on method: kirill@xaid.ai.

Zombie Jobs:163 Days Later

The shortage stopped being a number. It became a duration.

Before the national tables — here is your own number.

Ready-to-sign CT reports, powered by AI with humans in the loop.

We went back to the 2,111 roles that were already stuck in May.

The number of open chairs barely moved. The share that cannot close nearly doubled.

82.0% of long-stuck roles never changed the salary. Not by a dollar.

And the stuck chairs are not the cheap ones

Every recruiting lever in the market shows up more often on the chairs that stay empty.

The shortage is local. It is also not where the money is thin.

All 51 jurisdictions with a live radiologist opening

The hardest chairs to fill are the ones that read the most volume.

Same market, same candidates, twenty-one points of difference.

217 chairs have been advertised ten times or more.

The chairs currently advertised in America represent 1,108 radiologist-years of reading nobody is doing.

How this was built, and what it cannot tell you

Data

From postings to roles

Definitions

Known limits — stated plainly

External figures cited

Reproducibility

Zombie Jobs:
163 Days Later