California External Appeal Outcome Demographics

Background

When U.S. health insurers deny coverage for a treatment or service after it's already been rendered, consumers often have the right to appeal the decision. This typically begins with some form of internal appeal, which is reviewed by employees of the insurer, followed by an external appeal or Independent Medical Review (IMR), which is reviewed by a third party. We've discussed background on these rights in detail in previous posts, but those who are unfamiliar can peruse this for a brief primer.

In California, external appeals for many commercial health plans are regulated and monitored by the California Department of Insurance (CDI) and the Department of Managed Healthcare (DMHC).

Both of these departments release data about the outcomes of external appeals that they oversee:

• The CDI maintains this database of IMR outcomes.

  This data corresponds to 4,719 independent medical reviews conducted from 2011 through 2023 for individual and group health plans regulated by the California Department of Insurance.

• The DMHC maintains this database of IMR outcomes.

  This data corresponds to 36,615 independent medical reviews conducted from 2001 through 2023 for individual and group health plans regulated by the California Department of Managed Healthcare.

Collectively, these datasets correspond to the external appeals submitted among a population of roughly 30 million¹ insured individuals in California.

We previously analyzed various aspects of this data as well as a large swath of the entire landscape of publicly available U.S. claims denial data in our claims denial article.

Demographic Breakdowns

In this post (and generally) we're interested in understanding how claims adjudication outcomes vary based on social determinants of health, and other factors that have implications for equity in adjudication processes. Public claims metadata that includes such information is currently highly limited, but has the potential to help ensure vulnerable and disadvantaged populations are not facing inequitable or inappropriate treatment (or discrimination).

In the California data we consider here, we are limited to considering how external appeal outcomes vary according to a small and sparsely recorded set of demographic metadata fields recorded with each appeal.

We consider first how the external appeal outcomes vary to according age and gender. Demographic data of this sort is recorded, albeit inconsistently, in both datasets. We also consider how outcomes vary according to race; demographic data necessary to make this consideration is only provided in the CDI dataset.

We then consider how outcomes vary according to diagnoses and treatments associated with the claims. In both datasets each external appeal record contains two levels of diagnosis indications, as well as two levels of treatment indications; in the DMHC data, all of these records are further qualified by detailed medical/contractual summaries pertaining to the case at hand.

Age, Gender and Race

There is no conclusive indication in the data that an appellant's age, gender or race alone affects their likelihood of external appeal success. However, there are a few facts that might merit more investigation:

• Historically younger appellants have achieved considerably more favorable external appeal outcomes than older appellants in aggregate.

• Historically appellants whose reported gender is Female have achieved minimally favorable external appeal outcomes compared to those whose reported gender is Male, while those whose reported gender is Other have achieved the best external appeal outcomes, but the data in that category is extremely sparse, so it's unclear whether it is meaningful to compare to directly.

• Overall historical overturn rates vary considerably based on reported race, but the the data for which race is reported is also heavily skewed towards a few categories². As a result, it's very difficult to draw conclusions from the reported data.

The interactive components below can be used to compare outcomes based on reported age, gender, race and expedited status of each external appeal.

You can view the data used to generate the findings presented in these widgets here.

Disease States

We're also interested in understanding how denial and appeal patterns vary across different populations of patients, particularly those which already face inequities in other aspects of our health systems.

You can use the widgets below to compare outcomes based on the specific diagnoses (labeled as Diagnosis Subcategory in the database) associated with each external appeal.

You can view the data used to generate the findings presented in these widgets here.

We highlight some outcome statistics associated with a few particular diagnoses. We focus on these only because, anecdotally, we've personally witnessed a disproportionate number of seemingly inappropriate denials among such patient populations, and were curious to what extent these patterns actually exist in the data. We have no reason to believe these particular diagnoses are the only diagnoses worth carefully investigating in this data.

As the table above shows, overturn rates among appeals labeled with these diagnosis subcategories are mostly higher than the average across the dataset (the average being 49% for each database). In the case of gender dysphoria, the overturn rate is as high as 86% across the DMHC data.

We're interested in understanding why there are large disparities between the determinations of internal and external reviewers among certain demographics of patients (e.g. those who reach external appeals for care related to gender dysphoria), while there are not in other cases (e.g. those who reach external appeals for care related to back pain)?

We can't answer such questions from this data alone, but it is an important question whose answer may have implications for health equity in appeals processes.

Review Times

Finally, we want to highlight that some of the individual and average review times for expedited external appeals throughout these datasets are troubling (at least on paper, without further explanation).

In at least a subset of plans in California to which this data corresponds, the maximum allowed expedited review time barring extenuating circumstances ought to be 72 hours from the time of receipt. However, this data displays many instances that exceed these allowable review times (for example, the average expedited review time in 2021 for records in the CDI dataset with Diagnosis Subcategory Crohn's Disease was 6 days).

This is troubling. The argument is often made that those who risk health consequences as a result of inappropriate denials can always file an expedited appeal; what's often left out is that in practice this time limit may or may not be followed, and if it is not, all remaining recourse available to patients only adds additional delay to coverage of their care. Furthermore, recourse involving regulators or attorneys cost additional time or money, and more advantaged patients are of course more easily able to pursue such recourse.