Newly available data show that the number of 53+ minute sessions (CPT code 90837) provided to Medicare beneficiaries increased by 48% from 2013 to 2015, to nearly 3.2 million of these sessions. Estimated Medicare payments for these sessions nearly doubled in that same period, to more than $263 million. The increase in total payments is not accounted for by the 8% increase in the reimbursement amount for 90837 during that time, and is largely attributable to increased utilization of this code.

Is it possible that the increased provision of 53+ minute sessions is attributable to an increase in the number of Medicare beneficiaries for whom treatment is necessary? CMS reports there were 52.5 million enrollees in 2013, and nearly 55.6 million in 2015, a 5.9% increase. It seems doubtful that a 6% increase in the Medicare population would account for a 48% increase in the need for such services.

Were psychotherapy services to Medicare beneficiaries increasing overall during this time period? To answer this question, it is useful to look at rates of change in the second most commonly provided psychotherapy session length, the 45-minute 90834. Here is what the data show.

The number of 45 minute sessions decreased among all provider groups from 2013 to 2015. Use decreased by 11.9% among psychologists, by 2.3% among social workers, by 17.8% among psychiatrists, and by 23.2% among all other Medicare providers of this service. Use of 90834 decreased by 8.8% overall among all Medicare providers who billed this service. Changes in total amount of — and total expenditures for — psychotherapy provided to Medicare beneficiaries from 2013 to 2015 are not attributable to an increase in therapy in 45 minute sessions, but rather almost entirely due to the the increase in use of 53+ minute sessions. Here is the pattern from 2013 to 2015.

Relative changes in use of these codes by provider group may be seen in the following.

Given the pattern, it is difficult to escape the conclusion that sessions of 53+ minutes were being substituted for at least some of the 45 minute sessions. The impact on Medicare expenditures for these two modes of psychotherapy may be shown as follows.

Who was providing the therapy?

From 2013 to 2015, clinical psychologists provided about 47% of all sessions (45 minutes and 53+ minutes combined), licensed clinical social workers about 43%, psychiatrists about 7%, and all other health professionals about 3%.

The pattern is slightly different depending on the session length. From 2013 to 2015, psychologists provided 50% of the 45 minute sessions, social workers about 40%, with the remaining 10% provided by psychiatry and other providers.

In contrast, from 2013 to 2015, social workers provided about 45% of the 53+ minute sessions, psychologists 42%, psychiatrists about 10%, and other provider groups 4%.

It is noteworthy that, while psychiatrists consistently provided about 7% of all 45 minute sessions, and 4% of sessions overall, they provided nearly 10% of the 53+ minute sessions. In contrast, “all other” provider types provided about 4% of the 53+ minute sessions, and 3% of all therapy sessions during this time period.

It appears psychiatrists have increasingly been providing therapy in 53+ minute increments, while decreasing their use of the 45 minute session (by about 18%). However, their rate of growth in the use of this format (5% from 2013 to 2015) does not come close to matching the rate of growth among social workers (57% increase from 2013 – 2015), or among psychologists (51% increase from 2013 – 2015).

Also of note, it appears other providers are increasingly abandoning 45 minute sessions to provide therapy in 53+ minute sessions. Other providers showed a decrease of 23% in the number of 45 minute sessions provided, and a corresponding increase of about 27% in the number of 53+ minute sessions, from 2013 to 2015.

This does seem somewhat surprising given that few among the “other provider” group are specifically trained in mental health treatment or primarily serve those with mental health conditions. While it is understandable that these providers would encounter and treat those with mental health conditions, it is difficult to understand that they would have increased availability in busy specialty practices to provide psychotherapy in 53+ minute increments rather than 45 minute increments. It seems possible that the economic incentive is playing a role. It will be interesting to see if the increasing use of the 53+ minute session among other providers draws increased scrutiny.

The same certainly could be said for the “mental health specific” professionals, particularly if increased use of the 53+ minute modality — and associated expenditures by Medicare — continue at rates seen from 2013 to 2015, along with corresponding declines in the use of the 45 minute code. There certainly could be non-economic incentive explanations for this pattern. But it would not be surprising if “medical necessity” reviews become more common for the 90837 code, especially now that base rates — and outliers — can be well defined in the very large Medicare patient population and provider panel.