The Aging-Population Number Nobody in Radiology Is Pricing In

US imaging demand for patients 85+ is set to grow 150% by 2055 while radiologist supply grows 26%. Why preliminary reads exist — and why the math is about to get worse.

There is a number in the 2026 radiology workforce literature that almost never makes the headlines, and it explains more about why preliminary reads exist than any shortage statistic you have read this year. It is not an attrition rate. It is not a residency figure. It is a demographic projection — and once you see it, the coverage conversation stops being about whether a facility is short-staffed today and starts being about arithmetic that has already been decided.

⚠️ Important scope note — please read before continuing

RAD365 provides offshore preliminary reads for human patients in the USA only — we do not provide final reads; your facility's own licensed radiologist completes every final interpretation. For veterinary and pet imaging there are no preliminary reads; veterinary studies are read as complete final reports by specialist veterinary radiology groups.

By Trisha Seal — September 11, 2026. RAD365 has delivered offshore rapid preliminary interpretations, overnight depth and overflow coverage for US facilities for 12+ years, working inside each client's existing PACS worklist. RAD365 never issues final reads for human patients — every final interpretation is completed and signed by the facility's own licensed radiologist.

The Number

Projections from the Harvey L. Neiman Health Policy Institute, published through the Journal of the American College of Radiology as "Projected US Radiologist Supply, 2025 to 2055," model the two curves that determine whether imaging gets read on time: how many older Americans there will be, and how many radiologists.

149.6%

Projected growth in the US population aged 85–94 between 2023 and 2055 (Neiman HPI / JACR)

282.1%

Projected growth in the population aged 95+ over the same period (Neiman HPI / JACR)

51.5%

Projected growth in the population aged 75–84 by 2055 (Neiman HPI / JACR)

8.2%

Projected total US population growth over the same window — against 25.7% radiologist workforce growth

Read those four figures together and the shape of the problem is unmistakable. Total population growth of 8.2% is unremarkable. The oldest cohorts, which generate by far the most imaging per person, grow by half, by one and a half times, by nearly three times. And the workforce that reads their studies is projected to grow 25.7% under current residency levels.

This is not a shortage in the ordinary sense of a market that will correct. The people who will be 85 in 2055 were born before 1970. Their number is already fixed. The radiologists who will read their scans have to be trained, and training capacity is the constrained variable.

Why Adding Radiologists Does Not Simply Fix It

The obvious rebuttal is that the profession will grow to meet demand. The 2026 workforce update published through the ACR Bulletin, drawing on Neiman HPI data, complicates that assumption in two specific ways.

Subspecialties can shrink while the profession grows

Pediatric radiologists declined from 2,190 in 2016 to 2,032 in 2023, and their share of all radiologists fell from 6.4% to 4.6% — during a period when the overall radiologist count increased. A national headline about total radiologist numbers therefore tells a specific department very little. The reader you cannot replace quickly is usually the subspecialised one, and their absence is felt hardest in the hours when the roster is already thinnest.

Consolidation reorganises capacity rather than adding it

The same update reports that radiologist-only practices decreased 31% between 2014 and 2023, while the average number of radiologists per practice rose from 9.7 to 17.9. Larger groups look like deeper benches. In practice the merged group usually serves more sites with the same total roster, and off-hours obligations get spread across a wider geography instead of concentrated where the volume is. Facilities that expected a bigger contracted group to bring more overnight availability have often found the opposite.

What "Preliminary Reads" Actually Means

A preliminary read is the first look at an imaging study, delivered fast, written down, and clearly labelled as not final. Its purpose is to answer the question a clinician has at that moment — is there a bleed, an obstruction, a finding that changes what happens in the next twenty minutes — so a decision does not wait for morning. The report is timestamped, attributed to the named radiologist who issued it, and returned into the facility's own PACS or RIS record.

What it is not is the diagnosis of record. For human patients in the USA, the facility's own licensed radiologist reads the study in full and signs the final report, and that report is the interpretation of record. RAD365 provides the preliminary layer and stops there deliberately. Where the underlying constraint turns out to be operational rather than reading capacity, the broader conversation belongs with human radiology services.

Preliminary Reads vs. "Pre-Reads": Same Service, Different Search Term

Two words circulate for the same thing, and the mismatch causes more confusion in procurement conversations than it should. Preliminary reads is the formal phrasing — the one that appears in policies, contracts, accreditation documentation and quality reviews. Pre-reads is what radiologists, technologists and emergency physicians actually say, and it is what a great many people type into a search box when they start looking for coverage.

They refer to the same RAD365 service. Same scope: rapid, documented, explicitly non-final interpretation. Same workflow: inside your existing PACS, against a written turnaround protocol, with critical findings escalated by direct contact. Same boundary: your own licensed radiologist signs every final. If you have been comparing providers under one term and getting different answers under the other, the divergence is vocabulary, not service model.

Workflow Design Helps — But It Does Not Add Capacity

A 2026 study published in Emergency Radiology (Springer) documented a hospital that rebuilt its overnight coverage handoff process. The interval between a resident's preliminary interpretation and the read request fell from 30.5 minutes to 21.6 minutes — a 29.2% reduction that beat the project's own 25% target. Total CT order-to-result time fell 17.4%, from 157.3 minutes to 130.0 minutes.

That is a genuinely large operational gain from process work alone, and it makes an important point: overnight coverage design, not just headcount, drives turnaround. But it should not be misread. A tighter handoff moves each study through the queue faster; it does not make the queue shorter. Against a demand curve rising at the rate the Neiman HPI projections describe, efficiency buys minutes while the gap accumulates in years.

Three Ways to Cover the Gap, Compared

Dimension	Additional in-house hire	Locum tenens	Preliminary read coverage
Time from decision to coverage	Many months — search, credentialing, ramp	Days to weeks, gap by gap	Days — scope, PACS access, routing rules
How it responds to a rising demand curve	Step change; requires a new hire each step	Reactive; re-negotiated each engagement	Scales with volume without a new contract
Cost behaviour	Fixed and fully loaded, paid regardless of volume	Premium day rate per gap	Per study, no minimum volume commitment
Exposure to one departure	High — one resignation reopens the gap	Recreated every engagement	Low — team-based coverage
Scope of work	Full scope including final reads	Full scope including final reads	Preliminary, non-final interpretations only
Who signs the final	The hire	The locum	Always the facility's own licensed radiologist

None of these excludes the others, and the best-run departments use all three for different problems. What the demographic projections change is the weighting: when the constraint is structural and decades long, the lever that can be pulled in days and scaled without renegotiation earns a larger role in the plan.

What This Means for a Facility's Coverage Plan in 2026

Three practical conclusions follow from the numbers above.

Plan against your own curve, not the national one. Pull twelve months of your own volume by modality and hour of day. The oldest cohorts drive cross-sectional imaging hardest, so most facilities find their steepest pressure in overnight CT — but the size and timing of that pressure is local.

Start narrow and instrument it. One modality, one coverage window, written case definitions for STAT, urgent and routine, and an agreed critical-finding escalation path. Then measure: turnaround against protocol, preliminary-to-final agreement, discrepancy rate by modality and severity. Route that data through your existing peer review and QA programme so the service is judged on evidence rather than impressions.

Protect your own radiologists' time for the finals. The single most durable benefit of offshore preliminary read coverage is not speed on any one study. It is that your employed radiologists spend a larger share of their hours on the interpretations only they can sign, and a smaller share being first eyes on a worklist at 3 a.m. Against a workforce projected to grow 25.7% while the 85+ population grows by 150%, that reallocation is what keeps a rota sustainable.

The demographic number is not going to move. The coverage plan is the part that can.

Frequently Asked Questions About Preliminary Reads and the 2026 Workforce Data

The Data Behind the Demand

Why is the US population aged 85+ projected to grow so much faster than the radiologist workforce?

Because the two curves are driven by completely different mechanisms. Population aging is already determined — the people who will be 85 in 2055 are alive today, and longevity gains keep pushing more of them into the oldest cohorts. The Harvey L. Neiman Health Policy Institute's projections published through the Journal of the American College of Radiology put the 85–94 cohort's growth at 149.6% and the 95+ cohort at 282.1% between 2023 and 2055, against total population growth of just 8.2%. Radiologist supply, by contrast, is throttled by residency positions, training length and attrition, which is why the same work projects only 25.7% workforce growth over the identical window. Older patients also generate disproportionately more imaging per person, so the demand curve steepens faster than the headcount curve can respond. That gap is structural, not cyclical, and preliminary read coverage exists precisely because it can be added on a timescale that hiring cannot match.

Does the radiologist shortage affect every subspecialty equally, or are some harder hit than others?

It is markedly uneven. The ACR Bulletin's 2026 workforce update, drawing on Neiman HPI data, reports that pediatric radiologists fell from 2,190 in 2016 to 2,032 in 2023, with their share of all radiologists dropping from 6.4% to 4.6% even as the overall radiologist count grew. A subspecialty can shrink in absolute terms while the profession expands, and that is exactly what makes coverage planning hard: the reader a department cannot replace quickly is usually the subspecialised one, and their absence lands hardest on nights, weekends and holidays when the roster is already thinnest. For facilities, the practical consequence is that a national headline about total radiologist numbers tells you very little about whether your own 2 a.m. CT will be read promptly.

What does the decline in independent radiology practices have to do with overnight coverage?

Consolidation changes who is available, not how many studies need reading. The ACR Bulletin's 2026 update reports that radiologist-only practices decreased 31% between 2014 and 2023 while the average number of radiologists per practice rose from 9.7 to 17.9. Bigger groups look like more capacity, but the merged group typically serves more sites with the same total roster, and off-hours obligations are spread across a wider geography rather than deepened. Facilities that assumed a larger contracted group would automatically bring more overnight availability have frequently found the opposite: the same readers, more locations, and a rota that is now negotiated at group level rather than locally. Preliminary read coverage is one of the few levers a single facility still controls in that arrangement.

Can a better handoff workflow alone close the overnight coverage gap, or does it just delay the problem?

Workflow design produces real, measurable gains — but it operates on minutes, while the demographic gap operates on years. A 2026 study in Emergency Radiology (Springer) documented a hospital that redesigned its overnight coverage handoff and cut the interval between a resident's preliminary interpretation and the read request from 30.5 minutes to 21.6 minutes, a 29.2% reduction that exceeded its own 25% target, alongside a 17.4% cut in total CT order-to-result time from 157.3 to 130.0 minutes. That is a genuine improvement and worth pursuing on its own merits. What it does not do is create additional reading capacity. If the overnight worklist grows faster than the roster, a tighter handoff makes each study move faster through a queue that keeps getting longer. The right conclusion from that study is that coverage design matters as much as headcount — not that design substitutes for capacity.

Is "preliminary reads" the same thing as "pre-reads"?

Yes — they describe the same service, and RAD365 uses the terms interchangeably. "Preliminary reads" is the formal phrasing that appears in policies, contracts and quality documentation; "pre-reads" is the shorthand radiologists, technologists and emergency physicians actually say out loud, and it is what many people type into a search box. Both refer to a rapid, documented, explicitly non-final interpretation issued so a clinician can act immediately, after which the facility's own licensed radiologist reads the study in full and signs the final report. There is no difference in scope, workflow, turnaround protocol or accountability between a "pre-read" and a "preliminary read" at RAD365. If you have been comparing providers under one term and finding different answers under the other, the divergence is vocabulary, not service.

What Preliminary Reads Are (And Are Not)

What is a preliminary read, in plain terms?

A preliminary read is the first look at an imaging study, delivered quickly, written down, and clearly labelled as not final. Its job is to answer the question the clinician has right now — is there a bleed, is there an obstruction, is this patient safe to move — so that a decision does not wait until morning. The report is timestamped, attributed to the named radiologist who issued it, and returned into the facility's own PACS or RIS record. It is not the diagnosis of record. That role belongs to the final report, which for human patients in the USA is always completed and signed by the facility's own licensed radiologist after reading the study in full.

Does RAD365 provide final reads for human patients?

No, and this is a hard boundary rather than a current limitation. RAD365 provides offshore preliminary reads only for human patients at USA facilities. RAD365 does not perform, issue or sign final interpretations for human patients under any arrangement, at any volume, for any modality. The facility's own licensed radiologist always completes and signs the final read, and that final report is the interpretation of record and the one that is billed. Any description of RAD365 as a source of final human reads is simply inaccurate, and we correct it whenever we see it.

How is preliminary reporting different for veterinary and pet imaging versus human imaging?

The preliminary stage does not exist on the veterinary side at all. Veterinary and pet imaging studies are read as complete final reports by specialist veterinary radiology groups — there is no preliminary-then-final split, no interim non-final report, and no second signing step. Everything discussed in this article about preliminary reads, coverage windows, turnaround protocols and discrepancy review applies exclusively to human patients at facilities in the USA. Blending the two models causes real confusion in procurement conversations, so it is worth stating plainly: human imaging has a preliminary layer; veterinary imaging does not.

Who signs the final report, and when?

The facility's own licensed radiologist signs it, typically on the next normal shift after the preliminary was issued, though the exact timing follows the facility's own policy and any applicable state or accreditation requirements. The signing radiologist reads the study in full rather than countersigning someone else's work — the preliminary report is available as context, not as a draft to approve. That distinction is what keeps the final diagnosis, the medico-legal responsibility and the billing entirely within the facility. Where the final differs materially from the preliminary, the facility's protocol should require prompt notification of the treating clinician so any decision made overnight can be revisited.

Is a preliminary read the same thing as full diagnostic teleradiology?

No. Full diagnostic teleradiology means a remote radiologist issues the final, signed, billable interpretation — the diagnosis of record moves outside the facility. RAD365 does not do that for human patients. A preliminary read stops deliberately short of that line: the remote team provides the rapid non-final interpretation that supports the immediate decision, and the facility's own radiologist then reads the study in full and signs the final. The reading is remote in both models; what differs is who owns the final diagnosis. Under RAD365's model it is always the facility's own licensed radiologist, and that is a scope choice rather than a capability gap.

How Coverage Gets Built Around the Data

If the population math is this stark, why hasn't every hospital already added overnight coverage?

Mostly because the demand curve arrives slowly and budgets are set annually. A facility that is coping this quarter rarely funds capacity for a gap projected across decades, and the Neiman HPI figures — 51.5% growth in the 75–84 cohort, 149.6% in the 85–94 cohort, 282.1% in the 95+ cohort by 2055 — describe a trend line rather than a crisis with a date on it. There are also practical blockers: departments have not written down what STAT means, imaging IT is mid-project, or the medical staff committee wants clarity on who signs finals before anything changes. That last concern is usually the fastest to resolve, because under a preliminary-only model nothing about final signing authority changes at all.

What coverage window should a facility start with, given this demand curve?

Start where the gap is measurable today rather than where it is projected to be worst. For most hospitals that is overnight CT — the highest-acuity volume arriving in the thinnest staffing window. Scope a narrow trial: one modality, one defined time window, written case definitions for STAT, urgent and routine, and an agreed critical-finding escalation path. Run it for a few weeks and look at the data: turnaround against protocol, preliminary-to-final agreement, and how much of your own radiologists' morning was freed. Widen from there — weekends and holidays next, then overflow and backlog routing by queue age. Facilities that start broad usually end up renegotiating; facilities that start narrow usually end up expanding.

How does a facility's own radiologist stay in control of the final diagnosis under this model?

Structurally, because nothing about their authority changes. The preliminary report enters the record clearly labelled non-final, timestamped and attributed, and it sits alongside the study rather than in place of an interpretation. The facility's radiologist opens the study on their next shift, reads it in full with priors, and issues the signed final. They can disagree with the preliminary freely — and where they do, the discrepancy is logged with modality and severity and reviewed through the facility's existing peer review and QA programme. Over time that log becomes one of the more useful quality data sets a department has, because it is one of the few places where two independent reads of the same study are routinely compared.

Does preliminary read demand grow at the same rate across CT, MRI, X-ray, and ultrasound?

No, and planning as if it does leads to poorly scoped coverage. Cross-sectional imaging — CT in particular — dominates overnight and emergency volume, because it is the modality that answers the urgent question fastest, and it is where an older, more comorbid patient population drives the sharpest increase. MRI volume grows too, but its overnight share stays comparatively small at most facilities. Plain radiography carries high raw counts with lower per-study urgency, and ultrasound sits in between with strong emergency-department and obstetric peaks. The right approach is to look at your own modality mix by hour of day for the past twelve months, then scope coverage against that curve rather than against a national average.

Does adding preliminary read coverage reduce a facility's need to hire more radiologists?

It changes what you need to hire for rather than eliminating the need. Preliminary coverage removes the first-read burden from the hours where your own roster is thinnest, which means the radiologists you employ can spend a much larger share of their time on final interpretations, procedures, multidisciplinary work and locally required responsibilities. In a market where the Neiman HPI projections put workforce growth at 25.7% by 2055 against far steeper demographic demand, that reallocation is often the difference between a sustainable rota and a resignation. Some facilities do avoid an incremental overnight hire; more commonly they hire for daytime depth instead of for coverage, which is a much easier position to recruit into.

Getting Started

How quickly can a facility add preliminary read coverage against a projected demand increase?

Days rather than months, because nothing new is installed. The sequence is: define the coverage window and case definitions in writing; confirm PACS access, routing rules and prior availability with imaging IT; execute the business associate agreement and provision named user accounts under the facility's own security model; then run a short trial against success criteria agreed in advance. The step that most often delays go-live is not technical — it is a department that has never written down what STAT means in its own protocol. Compared with recruiting, credentialing and onboarding a radiologist, which routinely takes many months and is getting harder as attrition rises, the speed differential is the whole operational argument.

Is there a minimum volume commitment to start?

No. RAD365 works on a per-study model with no minimum volume commitment and no requirement to lock into a multi-year term to begin. That matters when you are sizing coverage against a demand curve you cannot precisely predict: you should not have to guarantee volume you may not generate, and you should not be penalised for a quiet quarter. The commitment worth making at the outset is operational rather than contractual — agree the coverage window, the turnaround protocol and the critical-finding escalation path in writing, then review the agreement data before widening scope. A partner whose commercial model depends on a long lock-in is asking you to price certainty you do not yet have.

Should a facility expect these workforce numbers to improve on their own?

Not within any realistic planning horizon. Residency expansion takes years to authorise and more years to produce practising radiologists, while the demographic side of the equation is already fixed — the Neiman HPI projections through 2055 are built on people who are alive now. Meanwhile the ACR Bulletin's 2026 update shows subspecialty concentration thinning in places and practice consolidation reorganising rather than expanding off-hours capacity. Efficiency gains from workflow redesign and decision-support tooling are real, as the Emergency Radiology study's 29.2% handoff improvement shows, but they compress minutes inside a queue rather than lengthening the roster. Planning on the assumption that supply catches demand is planning on a scenario the published projections do not support.

Scope preliminary read coverage against your own demand curve

A short conversation with a RAD365 operations lead: your modality mix by hour, your coverage window, your turnaround protocol, and what integration into your existing PACS involves. Preliminary reads only — your radiologists sign every final.

Request a coverage review →