42,621 decisions · California · 2001–2025
What independent reviewers actually decided
When a California health plan denies care and the enrollee takes it to Independent Medical Review, a physician who is not the plan’s decides. Every one of those decisions since 2001 is published, and the reasoning behind it is required by statute to be published with it. Nobody had turned that into base rates by class. This does.
- 42,621
- published decisions read
- 65.2%
- overturned, 2021-2025
- 188
- classes with enough decisions to rate
- 11,960
- name the standard applied
Every decision, and how it came out
every decisionOne column per report year. Above the line, decisions that overturned the plan; below it, decisions that upheld it — both on the same scale, so the column heights are comparable. The mass moving from below the line to above it is this dataset’s whole story, and the jump in 2015–16 is when the programme’s volume arrived.
- Overturned — the reviewer did not sustain the denial
- Upheld — the reviewer sustained the denial
42,621 decisions across 26 report years. In 2001 reviewers overturned 25%; in 2025, 72%. Hover a year for its counts. The dimmed final column is a part-year — the published file has not been refreshed since June.
The rate has an era, and the era matters
computedAcross every published decision the overturn rate is 52.5%. That number describes no era. Reviewers overturned 35.1% of denials in 2002 and 72.4% in 2025. A headline that averages across that is not a finding, it is an artefact.
Hover the chart for a year’s counts. Series stops at 2025 — report year 2026 uses a different category vocabulary.
Regime annotation. The dashed line marks SB 1410 (Stats. 2012, ch. 872), operative 1 July 2015, which added the field mandate now at Health & Safety Code § 1374.33(h). Median decision length nearly doubles just after it and the reviewer-credentials block goes from 16% of decisions to 98%. Rates are never averaged across a regime boundary.
The spread between classes is the product
computedA single headline hides the only thing a person holding a denial letter wants to know. Each dot is one treatment class, placed at its overturn rate; the bigger the dot, the more decisions behind it. Every class shown clears n = 30.
22 classes, from skilled nursing & inpatient rehab at 23.9% to neurosurgery at 86.0%. Median class: 53.6%.
- Skilled nursing & inpatient rehab24%
- Alternative treatment30%
- Orthopedic procedures31%
- Outpatient rehabilitation39%
- Home Health Care42%
- Special procedures42%
- Reconstructive & plastic procedures46%
- Diagnostic imaging & screening48%
- OB-GYN procedures50%
- Autism-related treatment51%
- Cancer Care53%
- Diagnostic & physician evaluation54%
- General surgery55%
- Durable medical equipment56%
- Electro, thermal & radiofrequency57%
- Acute inpatient medical62%
- Cardiovascular procedures62%
- Pharmacy75%
- Ear, nose & throat procedures76%
- Pain Management80%
- Mental Health80%
- Neurosurgery86%
All 92 treatment classes · 96 condition classes · Pair a condition with a treatment
The conditions most often disputed
by number of decisionsOrdered by how many decisions the record holds, never by rate — sorting these by outcome would publish a league table of what is most worth fighting for, which is a prediction dressed as a sort order.
What reviewers cite
validated codes onlyCalifornia requires each published decision to carry “a detailed case summary that includes the specific standards, criteria, and medical and scientific evidence, if any, that led to the case decision.” Reviewers name the standard they applied, and extracting that name is nearer to reading a label than to judging a case — which is why it is the most reliable thing on this site.
FDA
Pharmacy
74.2%n=2,151
LOCUS family
Mental Health
82.1%n=820
American Society of Addiction Medicine
Mental Health
79.3%n=584
NCCN
Cancer Care
52.1%n=451
American Academy of Child and Adolescent Psychiatry
Mental Health
77.1%n=315
American College of Rheumatology
Diagnostic imaging & screening
57.9%n=261
All 18 authorities — and what a rate against an authority does and does not mean.
Where this dataset is thin
what the record cannot supportA rate is published only on a cell of at least 30 decisions. Below that the matched decisions still count and no rate is claimed. This is what that costs at each level of detail — and it is the honest half of the product.
Treatment category
98.7%of decisions sit in a cell of 30+
22 of 29 cells
Condition x treatment category
86.9%of decisions sit in a cell of 30+
54 of 353 cells
Condition x treatment x review type
81.4%of decisions sit in a cell of 30+
58 of 479 cells
Condition x treatment subcategory
49.5%of decisions sit in a cell of 30+
67 of 2,165 cells
Known defects in the source, measured rather than estimated: 753 decisions record a partial outcome the published binary field collapses into a full win or loss; 89 (0.21%) contradict their own stated result and are excluded from every rate here rather than resolved in favour of one side.
Who this does not cover
stated before anything is soldThis covers California, and only denials fought all the way to external review. If your plan is a self-funded employer plan, or you are in Texas, Florida or Georgia, or you are on Medicare or Medicaid, it probably cannot cover your case at all — and 65.2% overturned does not mean 65.2% of denials are wrong. Read the coverage limits before anything else.
Free help, first: the California DMHC Help Center files your Independent Medical Review at no cost — 1-888-466-2219, dmhc.ca.gov. It is the same review these figures describe, and you do not need us, a lawyer or any paid service to reach it.