The VA: what the measures show

1The exhibits

Both claims are stated in their common form. Each circulates in many wordings, and each is rephrased here accurately and in the same way, so that no one person's exact words are singled out. The first claim, in full, argues against Medicare for All by pointing to government run care in Canada, the United Kingdom, the veterans medical system, and the Indian Health Service, and describes that care as poor throughout. This brief takes up the two American systems named in that claim, the veterans system and the Indian Health Service. The second claim is stated in its common form, because it circulates in many wordings and no single person's sentence can honestly be quoted for it. Both claims are repeated constantly in arguments over government health care. This brief checks them against the record, figure by figure, and then leaves the part that is genuinely debatable where it belongs, which is with the reader.

2What is being decided

• The question the record can settle: how good the care is once a patient is inside the system, on the measures that exist, because RAND and CMS have measured inpatient quality, outpatient quality, and patient experience with stated methods and published the results.
• The question the record cannot settle: how to weigh measured care quality against documented access failures, and what an enrolled, defined-population delivery system teaches about a national insurance plan with private delivery, because those ask how evidence should be weighed and how far evidence from one model carries to another, which are matters of judgment.
• The definitions that bound the topic: Public delivery means the government runs the facilities and employs the providers, as the VA and the Indian Health Service do; public insurance means the government pays for care that private providers deliver, which is the structure the Medicare for All Act proposes, and the two are not the same arrangement (the exhibit's glossary). Patient-experience stars are CMS ratings built from patient surveys covering communication, responsiveness, and cleanliness; they measure the experience of care and are not a pure grade of clinical quality (CMS). Outcome measures are clinical results, such as mortality, complications, and readmissions for defined conditions, and they are different evidence from satisfaction measures (the exhibit's glossary).

3What this argument is really about

Underneath both claims sit two questions that are usually graded as one. The first question is how good the care is once a patient is inside the system. The second question is what it takes to get inside, and whether the system reports that honestly. The VA has a documented record on both questions, and the two records point in different directions. There is also an unstated step in the argument. The systems named in the first claim are systems where the government delivers the care itself, in its own facilities, to a defined population. The Medicare for All Act proposes something structurally different, which is government insurance paying for care that private providers deliver. What the VA's record can and cannot say about that proposal depends on keeping the two models separate.

4The verified record

• Quality, measured independently: RAND's comparison, published in 2018, found VA hospitals performed the same as or significantly better than non-VA hospitals on 21 of 26 inpatient measures. On outpatient care, RAND found the VA better than commercial and Medicaid managed-care plans on 28 of 30 measures, with no significant difference on the remaining two. (Corroborated, RAND.)
• Patient experience: CMS patient-experience figures reported in September 2024, from survey data collected between October 2022 and September 2023, place 79% of VA facilities at four or five stars against 40% of non-VA hospitals, with the VA ahead on all ten core satisfaction metrics. (Corroborated, CMS data as reported.)
• The access record: the VA has real and documented problems in access and wait-time reporting, including a national scandal over manipulated wait lists in 2014, and outcomes vary considerably across its facilities. (Verified as a matter of public record and the VA's own inspector general reporting.)
• The funding comparison: the Indian Health Service is government-delivered care for a defined population, on the same model as the VA. A GAO comparison of 2017 spending puts spending per person at $4,078 for the Indian Health Service, $8,109 for Medicaid, $10,692 for the VA, and $13,185 for Medicare. American Indian and Alaska Native life expectancy was 67.9 years in 2022, against 77.5 years for the United States overall. The two systems share the delivery model. Their funding differs by roughly a factor of three, and their measured outcomes differ with it. (Corroborated, GAO figures; life expectancy, CDC.)

5Where exhibit 1 holds, and where it breaks

The first claim holds on access, because the VA's wait-time failures are documented, and they include an institution reporting waits it knew were false. The claim also holds on the Indian Health Service's outcomes, because a life expectancy of 67.9 years against a national figure of 77.5 years is a poor result by any measure. The claim breaks where it describes the quality of VA care, because the measured record on inpatient quality, outpatient quality, and patient experience runs the other way on most indicators. The claim also reaches past what its examples can test, because the VA and the Indian Health Service are systems where the government delivers care, and the proposal they are cited against would have the government insure care that private providers deliver. A delivery system and an insurance system are different machines, and the record of one is not a measurement of the other. One further point belongs to the record. The system in the claim with the poor outcomes is funded at roughly a third of the level of the system whose care measures well, and the difference between the two systems that the figures show is the funding.

6Where exhibit 2 holds, and where it breaks

The second claim holds on the measured quality of care, because the RAND comparison and the CMS patient-experience figures match or favor the VA on most of the measures that exist. The claim breaks where it treats those measures as an answer to the access complaint, because a veteran who cannot get an appointment is describing a real and documented failure, and a star rating does not answer it. The claim also reaches past its evidence in one respect, because national averages conceal VA facilities that perform badly, and "the VA" is not one hospital. There is a limit on what the VA's record can show about covering everyone, because its patients are enrolled in advance and its budget is appropriated as a whole, and a national plan would have to reproduce those design features deliberately rather than inherit them.

7What is true in both

Strip both claims down and they share a skeleton. Both treat the VA as a single thing that can be graded with a single word. Both also point at real parts of the record. The care delivered inside the system measures well, and the front door of the system has failed in documented ways. Both claims also stand on government-delivered systems, one funded at $10,692 per person and the other at $4,078, in the 2017 GAO comparison, with outcomes that differ accordingly. The record holds all of those parts at once.

8The number in the middle

79% vs 40%. The first claim needs the VA's measured record to be poor across the board. The second claim needs the star distribution to settle the argument on its own. CMS measured the distribution, as reported, and the measured distribution is 79% of VA facilities at four or five stars for patient experience, against 40% of non-VA hospitals. It is a measure of the experience of care inside the system. It is not a measure of getting in.

9Projections and analogues

No projection filed. Too many other decisions sit between this subject and any outcome we could measure, so the record does not support a projection.

10Debatable / not debatable

The factual part about quality is the part the record can settle, and RAND and CMS have measured it, with the VA matching or ahead on most measures. The factual part about access is also documented, and it runs against the system. What the record cannot settle is how to weigh measured care quality against documented access failures, and reasonable people weigh those differently. It also cannot settle what an enrolled, defined-population delivery system teaches about a national insurance plan with private delivery, because that is a judgment about how far evidence from one model carries to another. Those questions belong to the reader, and they are real ones.

11What we simplified

RAND's measures are process and outcome indicators for comparable conditions, not a grade on every service the VA provides, and mental-health and specialty access are measured less completely than the inpatient set.

12Words in this brief

Patient-experience stars

CMS ratings built from patient surveys covering communication, responsiveness, and cleanliness. Commonly misused as: a pure grade of clinical quality. They measure the experience of care, which is real, and partial.

Outcome measures

clinical results, such as mortality, complications, and readmissions for defined conditions. Commonly misused as: interchangeable with satisfaction. This brief keeps the two in separate paragraphs because they are different evidence.

Public delivery and public insurance

two different arrangements. In public delivery, the government runs the facilities and employs the providers, as the VA and the Indian Health Service do. In public insurance, the government pays for care that private providers deliver, which is the structure the Medicare for All Act proposes. Commonly misused as: the same thing, so that the record of one is treated as a verdict on the other.

13Sources

RAND, Comparing Quality of Care in VA and Non-VA Settings (2018) · CMS Hospital Compare / HCAHPS patient-experience data, as reported in 2024 · GAO-19-74R per-capita spending comparison, 2017 data (IHS, Medicaid, VHA, Medicare) · VA Office of Inspector General reporting on wait-time practices. Receipt Pack MC-0005 available on request.

14Filed updates

No filed updates.