BlogAI adoption

In my map of 44 healthcare problems, 30 need no patient data

Tyler Kent · October 9, 2026 · Updated October 14, 2026 · 3-minute read

I mapped the problems that cost six kinds of healthcare organizations the most money or carry the most risk: large health systems, community and rural hospitals, health centers, independent practices, post-acute and senior care, and the health-tech firms and associations that serve them. Each kind got seven or eight problems, 44 in all, and for each problem I checked one number, such as a cost or a rate, against the source I cite for it. Fig. 4 lists all 44, each with its number and source.

Two areas lead the list. Thirteen of the 44 problems are about money and payers, and ten are about rules, surveys and audits. Every one of the six kinds of organizations has at least one problem in each of those two areas.

Fig. 1

Of the 44 problems, 13 are about money and payers and 10 about rules, surveys and audits

Top problems of six kinds of US healthcare organizations, grouped by area. Kent Applied analysis, October 2026.

Show the numbers as a table
Of the 44 problems, 13 are about money and payers and 10 about rules, surveys and audits
CategoryValue
Money and payers13 problems
Rules, surveys and audits10 problems
Patients, referrals and customers8 problems
Paperwork and back office5 problems
Staffing5 problems
Daily operations3 problems

The numbers behind the list

Every problem on the list comes with a hard number. Hospital drug expenses grew 13.6% in 2025, against 7.5% for total expenses, AHA reports. Chartis puts 41.2% of rural hospitals in the red. Physicians complete an average of 40 prior authorizations a week, which take 13 hours of physician and staff time. In the median skilled nursing facility, 46% of nursing staff turned over in a year.

41.2%

of US rural hospitals operate in the red

41.2% of US rural hospitals operate in the red, and the median rural hospital's operating margin is 2.0%, Chartis reported in its 2026 study of rural hospitals.

Source: 2026 Rural Health State of the State.

KentAppliedkentapplied.com/facts/rural-hospitals-41-percent-in-the-red-2026

Most of the ten problems about rules, surveys and audits have a deadline or a penalty behind them. Hospital price files, website accessibility, community health needs assessments, health center site visits and nursing home surveys each come with one, and much of the work is matching documents to a checklist.

Healthcare uses AI at about the average US rate

In the Census Bureau's business survey for September 7 to 20, 2026, 24.8% of health care and social assistance businesses said they had used AI somewhere in their business. That is close to the 23.8% for all US businesses, and half the rate of information firms.

Fig. 2

Health care and social assistance firms use AI at about the US rate, half the top sector's

Share of US employer businesses that said they had used AI somewhere in their business between September 7 and 20, 2026: the four sectors with the highest use, health care and social assistance, all businesses, and the two sectors with the lowest use

Show the numbers as a table
Health care and social assistance firms use AI at about the US rate, half the top sector's
CategoryValue
Information50.2%
Professional, scientific and technical services42.7%
Finance and insurance40.1%
Educational services39.9%
Health care and social assistance24.8%
All US businesses23.8%
Accommodation and food services10.1%
Agriculture, forestry, fishing and hunting9.5%

How healthcare organizations buy AI matters as much as how many of them use it. In an HFMA survey of 233 health systems, nearly 80% said an existing vendor would have a significant advantage over a new one in an AI pilot.

For most of these problems, an AI tool needs no patient records

For each of the 44 problems I asked a narrow question: does any tool I'd advise for it have to read patient records? For 30, none does. Contracts, policies, staff records, public data, websites and summary reports are enough. A contract register that pulls claim filing deadlines and notice dates out of payer and vendor contracts never sees a patient's name. The same is true of a check of a hospital's public price file against its chargemaster (its internal list of charges), a site visit checklist matched to a health center's own documents, and a weekly brief built from summary reports.

Fig. 3

30 of 44 top healthcare problems need no patient data, in Kent Applied's analysis

Top problems of six kinds of US healthcare organizations, by whether an AI tool for them needs patient data. Kent Applied analysis, October 2026.

Show the numbers as a table
30 of 44 top healthcare problems need no patient data, in Kent Applied's analysis
CategoryNeeds no patient dataNeeds patient data for part of the workNeeds patient data from the startNo AI tool advisedTotal
Health systems62008
Community hospitals61007
Health centers52007
Practices40217
Post-acute and senior care43108
Health-tech and associations52007

Ten more have a first step that needs no patient data and a later step that does. With denials, for example, a tool can track contract terms and appeal windows before anyone opens a denied claim. With referrals, it can count referrals by source before anyone reads a referral packet. Three need patient data from the start: prior authorization at practices, practice claim denials and the hospice cap. Even a plain list of admission and discharge dates is patient data, since HIPAA's Safe Harbor method counts every part of those dates except the year as an identifier.

For one problem, hiring medical assistants and hygienists at practices, I'd leave AI out. AI that ranks job applicants needs a yearly bias audit in New York City and falls under California's rules on automated decision systems. A practice's applicant tracking system already posts jobs and schedules interviews, so an AI tool would bring new legal requirements and add little that system can't already do.

Independent practices and post-acute and senior care organizations have all three of the problems that need patient data from the start, and they have the fewest problems that need none: four of seven at practices and four of eight in post-acute and senior care. Health systems and community hospitals each have six problems that need no patient data, and health centers and health-tech firms each have five.

All 44 problems

Here is every problem on the map. You can group the list by area, by kind of organization or by the patient data a tool would need, and you can show one kind of organization at a time. Tap a problem to see its number and where that number comes from, who feels it, and the tools that can help.

Fig. 4

All 44 problems, with the number behind each one

Search them, or tap any square for the problem's figure and source, who feels it, and the tools that address it. Kent Applied analysis, October 2026.

  • Needs no patient data 30
  • Needs patient data for part of the work 10
  • Needs patient data from the start 3
  • No AI tool advised 1

Tap a square for the problem's figure, source and tools. Tap a row or column name to show only that kind or area.

Group the list by

All 44 problems, grouped by area.

Money and payers 13

Rules, surveys and audits 10

Patients, referrals and customers 8

Paperwork and back office 5

Staffing 5

Daily operations 3

Each figure was checked at its source in October 2026. Free to reuse with credit to Kent Applied.

Where to start

Start with the problem on your own list that costs the most and needs the least patient data. Fig. 4 tags the tools that read patient records from the start, and I don't start work on any of them until a business associate agreement is signed. When a later step would use patient-level data, such as UDS reporting or a discharge-delay study, the note on that problem says so and names the agreement it needs. The research page goes through the same problems one kind of organization at a time and explains how I chose them.

Sources

  1. Kent Applied, Where AI tools fit in healthcare, by kind of organization.
  2. AHA, Costs of Caring 2026.
  3. Chartis, 2026 Rural Health State of the State.
  4. AMA, 2025 prior authorization physician survey.
  5. MedPAC, March 2026 report, Chapter 7.
  6. US Census Bureau, Business Trends and Outlook Survey, sector estimates.
  7. US Census Bureau, Business Trends and Outlook Survey, national estimates.
  8. HFMA, Health system adoption of AI outpaces internal governance and strategy.
  9. 45 CFR 164.514.
  10. NYC Department of Consumer and Worker Protection, automated employment decision tools.

Corrected October 14, 2026. The summary now says the problems were chosen because they cost the most money or carry the most risk. The paragraph on problems about rules, surveys and audits now says most of them, not all, have a deadline or a penalty behind them, and the note on patient data now says which tools Fig. 4 tags. In Fig. 4, several figures now match their sources more closely, notes that state a rule now link to it, and two vendor prices whose source could not be linked were removed. Three problems also moved from needing no patient data to needing it for part of the work, because a later step for each reads patient records: automating the UDS report at health centers, checking charts before a Medicare claim in post-acute and senior care, and a results study on claims or visit records for health-tech firms. So 30 of the 44 problems need no patient data, not 33, and the title now says so.

Related service

Related articles

More on AI adoption: 16 charts and 16 sourced facts.

Insights by email

Does this look like your organization?

Book a 20-minute call with Tyler to talk through your version of this problem and what fixing it would cost.

Prefer to skip the call? Get a written proposal