Sensitivity of Mental Health Professional Shortage Estimates to Assumptions about Patterns of Care

Background and Purpose: Mental health workforce shortages in the US remain a persistent barrier to care, particularly in rural and underserved communities. Effective policy and planning to address these shortages require accurate workforce shortage estimates. Much prior research has produced shortage estimates, yet the sensitivity of those estimates to underlying assumptions has received little attention. We examined how variation in key assumptions about mental health need, workforce supply, and service delivery affects national county-level estimates of mental health workforce shortage. The primary research question is: to what extent do changes in assumptions about treatment requirements, workforce composition, and the role of primary care alter mental health workforce shortage estimates?

Methods: We applied sensitivity analyses to a previously published methodology for estimating mental health shortage. In this methodology, we used synthetic estimation and national surveys (National Comorbidity Survey-Replication, Medical Expenditure Panel Survey, Census, workforce surveys) to estimate need. We used licensure and professional association data to estimate supply. After smoothing the need and supply estimates to account for travel between counties, we estimated shortage as the difference between estimated need and supply, generating separate shortage estimates for prescribers and non-prescribers. In the current study, we applied a sensitivity analysis to this methodology by systematically varying five key parameters, one at a time: (1) the maximum proportion of mental health treatment that can be delivered by primary care providers, (2) the total annual provider time required for individuals with serious mental illness (SMI), (3) the proportion of total provider time for individuals with SMI allocated to prescribers versus non-prescribers, (4) the total annual provider time required for individuals without SMI, and (5) the proportion of total provider time for individuals without SMI allocated to prescribers versus non-prescribers.

Results: Using default parameters, the model estimated a national shortage of approximately 19,000 prescriber FTEs. Sensitivity analyses demonstrated substantial variability in prescriber shortage estimates, ranging from fewer than 2,000 FTEs to more than 79,000 FTEs. Assumptions about provider time requirements and the distribution of care between prescribers and non-prescribers produced the largest effects, while assumptions regarding the role of primary care had smaller effects. Across all sensitivity analyses, estimates consistently indicated a national surplus of non-prescribers, with no scenarios producing a national shortage of non-prescribers.

Conclusions and Implications: Mental health workforce shortage estimates reflect decisions–sometimes implicit decisions–about care delivery, including treatment requirements, workforce composition, and the role of primary care. Shortage estimates are highly sensitive to these underlying assumptions and should be carefully interpreted in that context, preferably through explicit sensitivity analyses. These findings highlight the need for greater transparency in workforce modeling and caution in using shortage estimates to guide policy decisions.