Simulating the Impact of Psychologist Prescribing Authority Policies on Mental Health Prescriber Shortages

Abstract

Objective:

To simulate the impact of granting prescriptive authority to licensed psychologists on shortages of mental health providers with prescriptive authority.

Methods:

We used national and state-level secondary data to construct a policy simulation. Mental health prescribing need and the number of mental health prescribers were estimated at the national and state levels, and the provider shortage was calculated as the difference between these two values. The simulated policy intervention added 10% of psychologists to the prescriber group to estimate the associated reduction in prescriber shortages. Probabilistic uncertainty and sensitivity analyses were conducted using 10,000 Markov trials in which all model parameters varied randomly based on their associated uncertainty. The simulated policy value was allowed to vary between 5% and 15%.

Results:

The prescriber shortage was predicted to fall by 4.34% (95% confidence interval = [0.75%, 16.58%]) nationally, though this varied widely by state, ranging from 1.10% [0.20%, 3.82%] in North Dakota to 26.44% [2.89%, 570.29%] in Washington, D.C. Uncertainty and sensitivity analyses demonstrated that variability in the provider shortage reduction was primarily driven by the percentage of psychologists becoming licensed to prescribe.

Conclusions:

Our results suggest that granting prescriptive authority to licensed psychologists would reduce the shortage of mental health professionals with prescriptive authority. Further work is needed to examine the potential implications for other mental health providers.

Mental health problems are highly prevalent in the United States, affecting 20.6% of all adults in 2019 (Center for Behavioral Health Statistics and Quality, 2020). In the same year, only 36.7% of those with a mental health problem reported receiving a psychotropic prescription, and over half (55.2%) reported not receiving any treatment (Center for Behavioral Health Statistics and Quality, 2020). The underutilization of psychotropic medications is concerning, as they are often part of the first-line treatment for some patients with many of the most prevalent mental health problems, such as depression and anxiety (American Psychological Association, 2017, 2019; Locke, 2015; Wise & Swartz, 2017). The well-documented shortage of mental health providers (e.g., psychiatrists, psychologists, and psychiatric nurse practitioners) is one potential source of this underutilization, especially the shortage of mental health specialists with prescriptive authority (hereafter, prescribers) (Andrilla et al., 2018; Beck et al., 2018; Konrad et al., 2009; Office of Health Equity, 2017; Olfson, 2016; Thomas et al., 2009). A foundational study of mental health workforce shortages found that 96% of counties had unmet need for mental health prescribers, and an average of 67% of prescribing need was unmet per county (Thomas et al., 2009). A more recent study suggests that not much has changed, finding that over half (51%) of counties have no psychiatrists, and a full two-thirds (67%) have no psychiatric nurse practitioners (Andrilla et al., 2018). Adjusting for population, the same study found only 15.6 psychiatrists and 2.1 psychiatric nurse practitioners per 100,000 people, demonstrating a severe shortage of prescribers trained in mental health (Andrilla et al., 2018). Solutions to address this shortage are needed.

One proposed solution to improve prescriber access is the passage of policies that allow for the licensure of prescribing psychologists (RxP) (Baird, 2007; DeLeon et al., 1991; Long Jr, 2005; Norfleet, 2002). Psychologists are much more plentiful than psychiatrists and psychiatric nurse practitioners combined, with only 37% of counties having no psychologists and a rate of 30.0 psychologists per 100,000 population (Andrilla et al., 2018). Such policies have been implemented in six states since 2002: New Mexico (2002), Louisiana (2004), Illinois (2014), Iowa (2016), Idaho (2017), and Colorado (2023); however, only the policies in New Mexico and Louisiana have been in place long enough for more than a handful of psychologists to obtain the necessary education and credentials required by their statutes (Robiner et al., 2020). In New Mexico, licensed psychologists are required to complete 450 credit-hours in graduate-level psychopharmacology, a 400-hour practicum supervised by a licensed physician, and a national psychopharmacology exam to be granted conditional RxP licensure in which they must continue to be supervised by a physician and complete 20 hours of continuing education annually (Prescriptive Authority to Psychologists, 2002). Psychologists may apply for full RxP status after 2 years of conditional RxP contingent upon the endorsement of their supervising physician (Prescriptive Authority to Psychologists, 2002). Louisiana requires licensed psychologists seeking RxP to complete a post-doctoral master’s degree in psychopharmacology and a national psychopharmacology exam, at which point they are eligible for RxP under the condition that all prescriptions be made in consultation and consensus with the patient’s primary care physician (Act No. 11, 2004). After at least three years of practice and having met several other conditions, prescribers in Louisiana can apply for advanced practice certification, which removes the consensus requirement. In both New Mexico and Louisiana, psychologists with RxP are limited to medications associated with the treatment of mental illness (Act No. 11, 2004; Prescriptive Authority to Psychologists, 2002). Under these regulations, as of 2019, 9.0% of psychologists in New Mexico and 13.6% in Louisiana had obtained their RxP licensure (Robiner et al., 2020).

Although the policies in New Mexico and Louisiana have been in effect for a number of years, the merits of expanding the scope of practice for psychologists is still hotly debated across the mental health field (McGrath, 2020; Robiner et al., 2006, 2020). One of the driving factors of this continued debate, as highlighted by McGrath (2020), is that there has been minimal research to date on the patient or workforce outcomes associated with these policies. To date, two studies have demonstrated a decrease in mental health-related mortality following the implementation of these policies (Choudhury & Plemmons, 2023; Hughes et al., 2022) and a third found that this is a cost-effective policy strategy for reducing suicide rates (Hughes et al., 2023); however, the mechanisms underlying those decreases remain unexplored. This study serves as an early step towards empirically estimating the impact of this policy by simulating the changes in the prescriber shortages at the state and national level under policies similar to those enacted in Louisiana and New Mexico. This study does not, however, attempt to assess unmet mental health need, nor does it assume that psychotropic medications are an ideal or appropriate treatment option for all patients with mental health concerns, but rather that some patients may benefit from psychotropic medications. This aligns with current perspectives on public health, especially given that increasing access to mental health treatment is a goal of HealthyPeople 2030 (National Academies of Sciences, Engineering, and Medicine (U.S.) & Committee on Informing the Selection of Leading Health Indicators for HealthyPeople 2030, 2020).

Methods

Data & Study Design

This was a simulation study based on secondary data with a Markov probabilistic uncertainty analysis (PUA) to estimate model variability (Figure 1). We approached this simulation from the perspective of state policymakers seeking to improve population mental health; however, given the lack of data on the implementation timeframe for these policies (e.g., psychologist prescriptive authority licensure rate changes over time), we designed this model to provide a cross-sectional simulation describing outcomes once the policy has been fully implemented. We used combined data from the 2018 and 2019 National Survey on Drug Use and Health (NSDUH) to estimate the number of prescribers needed to treat patients age 12 years and older with mental health needs using the Restricted Data Analysis System provided online by SAMHSA (Substance Abuse and Mental Health Services Administration, n.d.). The NSDUH is an annual survey with a complex sampling design conducted by the Substance Abuse and Mental Health Services Agency (SAMHSA) through a computer-assisted interview process designed to track substance use and mental health in the general, non-institutionalized population. SAMHSA generates state-level estimates using small area estimate methods (Substance Abuse and Mental Health Services Administration, 2017). To estimate the available mental health workforce, we used provider count estimates from the Fitzhugh Mullan Institute for Health Workforce Equity’s Behavioral Health Workforce Tracker (Fitzhugh Mullan Institute for Health Workforce Equity, 2022), which includes counts of psychiatrists and psychologists compiled using a combination of state licensure data and national prescribing data.

Figure 1. Unmet Prescribing FTE Calculation.

Abbreviations: Rx = prescription; FTE = full-time equivalents. Patients with any mental illness were classified into groups based on the type of mental health care they did or did not receive. Providers were classified into prescribers and non-prescribers by scope of practice. Estimated prescribing need and available prescribers were converted to FTE. *People with no mental health care were weighted to match the prescription distribution of those who had mental health care.

Measures

Prescribing FTE Needed.

To identify the total mental health prescribing need per state, we used a five-step process. First, we identified all individuals in the NSDUH with any mental illness. The NSDUH uses a prediction model to flag individuals with any mental illness based on a combination of multiple screening tools and age (Substance Abuse and Mental Health Services Administration, 2017). Second, we used the recoded mental health services utilization indicator to categorize individuals with any mental illness into three categories: did not receive any mental healthcare in the past year, received mental healthcare without a prescription (i.e., therapy only), and received mental healthcare including a prescription (i.e., a prescription with or without therapy). SAMHSA generates these counts using survey weights appropriate for state-level estimates and provides output as a table of the estimated population count of individuals with any mental illness in each mental health service category for each state. Third, under the assumption that all people with a mental illness would benefit from mental health services and a proportion would choose medication, we estimated the need for prescription medication among those not receiving care by applying the proportion of service users who had a medication to nonusers. In this way, need was measured on the basis of actual utilization among users and attempts to conservatively account for patient preferences by acknowledging that only a portion of people would seek medication-based treatment. This approach is consistent with prior work on mental health workforce shortages (Konrad et al., 2009; Thomas et al., 2009) and the characterization of unmet mental health need by the Health Resources and Services Administration (Office of Health Equity, 2017). Note, however, that this approach is not intended to account for patient-level barriers (e.g., insurance coverage) in our need estimate, but rather to provide an estimated count of individuals desiring a mental health medication. Fourth, the count of people with prescription needs among those not using services was combined with the count of people with any mental illness currently receiving treatment with a prescription, forming the total count of people with prescribing needs. All state-level measures of unmet need were allowed to vary in the PUA using a log-normal distribution based on the standard error for each estimate. A log-normal distribution was used to prevent negative counts of people with mental illness in small states while still approximating a normal distribution in larger states. Finally, the visits needed to provide medication to all patients was converted into full time equivalents (FTE) per visit assuming a 30-minute appointment. The FTE conversion was allowed to vary between 15, 30, and 60 minutes in the PUA. Given that we calculated the FTE needed for a singular visit for all patients, this value can be viewed as an instantaneous prevalence estimate of workforce demand, providing a measure of how many full-time prescribers would be needed to treat all patients at once.

Available Provider FTE.

To estimate the mental health workforce available per state, we used the total number of psychologists and psychiatrists. Psychiatrists were used as the total count of available prescribers, while psychologists were counted as non-prescribers in the base model. Providers were converted to FTEs based on the average percentage of time spent in direct patient care by licensure as reported in prior workforce studies (psychiatrists: 0.604 FTEs; psychologists: 0.640 FTEs) (Konrad et al., 2009; Thomas et al., 2009) and were allowed to vary based on a BetaPERT distribution using the base estimates as the likeliest value. By calculating available FTEs in this manner, they can be viewed as an instantaneous prevalence estimate of workforce supply, providing a measure of how many full-time prescribers are available to treat all patients at once.

Prescribing FTE Shortage.

Prescribing FTE shortage was defined as the difference between the Prescribing FTE needed and the Prescriber FTE available and can be viewed as an instantaneous prevalence estimate of workforce shortage, providing a measure of the difference between how many full-time prescribers are needed and how many are available to treat all patients at once. In order to adjust for mental health prescribing done by providers other than psychiatrists and psychologists, we deflated the prescribing FTE needed by the percentage of psychotropic medications prescribed by general practitioners (43.5%) and nurse practitioners (12.9%) prior to calculating unmet need (Hughes et al., In Press). These deflation parameters were allowed to vary using a normal distribution based on the standard errors of the base-case parameters. This deflation approach is consistent with prior mental health workforce simulations based on visits. However, it may overstate the portion of mental health prescribing that would be met by primary care physicians in areas where there are primary care shortages, and thus yields a conservative estimate of the policy impact (Konrad et al., 2009; Thomas et al., 2009).

Proportional Prescriber Shortage.

In order to compare changes in prescriber shortages across states, we examined the proportional prescriber shortage. This proportion was calculated using the prescribing FTE shortage as the numerator and prescribing FTE needed as the denominator.

Analytic Approach

We estimated the newly available prescriber FTE per state by converting a percentage of the available psychologist FTE into prescriber FTE. Based on the percentage of psychologists who became licensed prescribers in New Mexico (9.0%) and Louisiana (13.6%) (Robiner et al., 2020), we used a base-case value of 10% of psychologists becoming prescribers and allowed this to vary uniformly between 5% and 15% of psychologists becoming prescribers in uncertainty analyses. This simulated proportion of psychologists becoming prescribers is then added to the prescriber FTE estimate. The newly available prescriber FTE was then used to calculate the absolute and relative reductions in the proportional prescriber shortage after adoption of the RxP policy. We included both absolute and relative reductions for transparency, as both metrics may be useful to policymakers and future researchers. Louisiana and New Mexico were excluded from all analyses in order to avoid doubling the impact of their existing policies, while Idaho, Illinois, Iowa, and Colorado were included given that their policies are relatively new and a minimal number of psychologists have become licensed to prescribe in those states.

The PUA was conducted using a 10,000-trial Markov simulation in which all parameters were varied randomly across their specified distributions. The results of the PUA were summarized in several ways. First, the 2.5^th and 97.5^th percentiles of the PUA were presented as the 95% confidence interval for the base-case estimates, a common approach used in simulation modeling (Butani et al., 2021; Prinja et al., 2016). Second, the median absolute and relative reductions per state per percentage of psychologists licensed to prescribe were presented to describe the association between psychologist prescribing licensure rates and the reduction in prescriber shortages. Finally, the results of the PUA were used in distributional one-way sensitivity analyses describing which model parameters had the greatest influence on the reduction in prescriber shortages (Vreman et al., 2021). All analyses were conducted using Microsoft Excel 2016 (Redmond, WA) and Crystal Ball (Oracle Crystal Ball v 11.1.2.4, Oracle Software, 2020). This study was determined to be exempt by the institutional review board of the University of North Carolina at Chapel Hill due to the use of publicly available secondary data. This study adheres to the JARS reporting standards (Kazak, 2018). Materials for this study are available by emailing the corresponding author. This study was not preregistered.

Results

Nationally, we found 29,135.1 prescribing FTEs and 82,090.9 psychologist FTEs currently available. After adjusting for prescribing done in primary care, there were 218,112.7 FTEs needed to provide and manage a prescription for all individuals with any mental illness who currently have or would receive a medication. The difference between the needed FTE and the available prescriber FTE produced a prescriber shortage of 188,977.5 FTE, or a shortage of 86.6% needed prescribers. The proportional prescriber shortage ranged from a low of 43.1% in Washington, D.C. to 95.7% in Idaho. State-level estimates of available FTE, needed FTE, and provider FTE shortage can be found in Table S1.

In the base-case model, there was a 4.34% (95% Confidence Interval [0.75%, 16.58%]) reduction in prescriber shortages corresponding to an absolute reduction of 3.76 percentage points [0.73, 11.96] (Table 1). Washington, D.C. (26.44% [2.89%, 570.29%]), Hawaii (10.69% [1.73%, 62.63%]), and Rhode Island (10.61% [1.76%, 51.45%]) had the largest relative reductions. Oklahoma (1.10% [0.20%, 3.82%]), West Viginia (1.13% [0.20%, 4.07%]), and Wisconsin (1.17% [0.21%, 7.11%]) had the smallest relative reductions. All relative reductions and the corresponding absolute reductions for all states can be seen in Table 1.

Table 1.

Simulated Change in Unmet Mental Health Prescribing Need Under RxP

	Absolute Reduction			Relative Reduction
		95% CI			95% CI
	Base	LL	UL	Base	LL	UL
Overall	3.76	0.71	11.96	4.34%	0.75%	16.58%
Alabama	1.68	0.31	5.40	1.80%	0.33%	6.24%
Alaska	3.71	0.56	42.80	4.17%	0.59%	183.17%
Arizona	2.50	0.46	8.03	2.73%	0.48%	9.74%
Arkansas	1.91	0.34	7.08	2.08%	0.35%	8.99%
California	5.14	0.96	16.33	6.16%	1.05%	24.96%
Colorado	5.28	0.98	16.96	5.91%	1.04%	21.86%
Connecticut	7.20	1.30	23.56	10.49%	1.56%	152.47%
Delaware	5.24	0.96	16.99	5.88%	1.02%	22.38%
District Of Columbia	11.40	2.12	37.13	26.44%	2.89%	570.29%
Florida	2.99	0.56	9.60	3.31%	0.58%	12.00%
Georgia	2.77	0.51	9.00	3.10%	0.54%	11.60%
Hawaii	8.17	1.52	26.77	10.69%	1.73%	62.63%
Idaho	2.11	0.39	6.90	2.21%	0.40%	7.55%
Illinois	4.61	0.86	14.95	5.41%	0.93%	21.80%
Indiana	2.39	0.44	7.67	2.57%	0.46%	8.93%
Iowa	3.06	0.56	10.00	3.47%	0.60%	13.37%
Kansas	3.10	0.58	9.94	3.45%	0.61%	12.59%
Kentucky	1.38	0.25	4.46	1.49%	0.26%	5.30%
Maine	4.39	0.81	14.24	5.09%	0.88%	20.09%
Maryland	6.80	1.26	22.09	8.79%	1.42%	45.77%
Massachusetts	7.51	1.38	24.46	10.39%	1.60%	79.46%
Michigan	3.38	0.63	10.80	3.87%	0.67%	14.71%
Minnesota	5.43	1.01	17.51	6.14%	1.08%	23.17%
Mississippi	1.10	0.21	3.54	1.17%	0.21%	4.03%
Missouri	2.27	0.42	7.43	2.56%	0.45%	9.77%
Montana	2.21	0.41	7.28	2.38%	0.43%	8.48%
Nebraska	2.88	0.54	9.26	3.27%	0.57%	12.25%
Nevada	1.86	0.34	6.13	2.00%	0.36%	7.11%
New Hampshire	3.35	0.62	10.83	3.85%	0.66%	14.82%
New Jersey	4.81	0.89	15.52	5.70%	0.97%	23.69%
New York	7.49	1.38	23.90	10.58%	1.62%	87.91%
North Carolina	2.86	0.53	9.20	3.23%	0.57%	12.16%
North Dakota	3.19	0.59	10.31	3.68%	0.63%	14.31%
Ohio	2.23	0.41	7.06	2.47%	0.43%	8.88%
Oklahoma	1.02	0.19	3.30	1.10%	0.20%	3.82%
Oregon	3.92	0.73	12.65	4.36%	0.77%	15.94%
Pennsylvania	4.43	0.83	14.24	5.33%	0.90%	22.37%
Rhode Island	8.38	1.56	26.79	10.61%	1.76%	51.45%
South Carolina	1.52	0.28	4.98	1.71%	0.30%	6.56%
South Dakota	2.06	0.38	6.72	2.35%	0.41%	9.10%
Tennessee	2.22	0.41	7.21	2.40%	0.43%	8.53%
Texas	2.26	0.42	7.26	2.51%	0.44%	9.12%
Utah	2.80	0.52	9.01	3.01%	0.54%	10.49%
Vermont	6.51	1.21	21.33	8.03%	1.33%	36.65%
Virginia	5.41	1.00	17.42	6.28%	1.07%	24.58%
Washington	2.80	0.52	9.09	3.10%	0.54%	11.43%
West Virginia	1.03	0.19	3.33	1.13%	0.20%	4.07%
Wisconsin	3.01	0.56	9.89	3.39%	0.59%	12.81%
Wyoming	3.37	0.62	10.87	3.54%	0.63%	12.00%

Note: The absolute reduction presents the percentage point difference between the current unmet need and the modeled unmet need under an RxP policy. The relative reduction represents that difference as a percentage of the current unmet need. Confidence intervals represent the 2.5th and 97.5th percentile of a probabilistic uncertainty analysis using 10,000 Markov simulations. LL = Lower Limit; UL = Upper Limit.

The median reduction per state by simulated psychologist prescribing licensure percentage can be seen in Figure 2a (relative reduction) and Figure 2b (absolute reduction). On both the absolute and relative scale, there was a dose-response relationship between the percentage of psychologists who become licensed to prescribe and the size of the reduction in prescriber shortages. At the maximum of 15% of psychologists becoming prescribers, Connecticut (12.5%), Washington, D.C. (30.0%), Hawaii (13.5%), Maryland (10.6%), Massachusetts (12.8%), New York (12.9%), Rhode Island (13.4%), and Vermont (10.2%) all had relative reductions of 10% or greater (Figure 2a). Conversely, at the minimum of 5% of psychologists becoming prescribers, Alabama (0.9%), Kentucky (0.7%), Mississippi (0.6%), Oklahoma (0.5%), South Carolina (0.8%), and West Virginia (0.6%) all had relative reductions of less than 1%. On the absolute scale, only Washington, D.C. (14.4), Hawaii (10.2), and Rhode Island (10.4) had reductions of greater than 10 percentage points at the maximum psychologist prescribing percentage, while 9 states had reductions of less than 1 percentage point at the minimum psychologist prescribing percentage (Figure 2b).

Figure 2. Median Reduction in Prescriber Shortage by State and Psychologist Prescribing Licensure Percentage.

Note: New Mexico and Louisiana have been excluded, as these policies are already in place in those states.

The results of the distributional one-way sensitivity analysis are presented as tornado plots in Figure 3 (relative reduction) and Figure 4 (absolute reduction). These analyses describe the proportion of variation in shortage reduction attributable to the variation in each component of the model, identifying which components of the model the outcomes are most sensitive to changes in. The variation in median relative reduction was explained primarily by variation in the calculations for FTE per visit (62.18% of the variation), followed by psychologist prescribing licensure percentage (19.71% of the variation), FTE per psychologist (17.85% of the variation), and FTE per psychiatrist (0.27% of the variation). For example, 62.18% of the variation in median relative reduction was explained by how long patient visits were assumed to be in the calculation of FTE needed (i.e., 15, 30, or 60 minutes). The variation in median absolute reduction was explained similarly, with the FTE per visit explaining over half (54.14%) of the variation in median absolute reduction, followed by the psychologist prescribing licensure percentage (24.06%) and the FTE per psychologist (21.79%).

Figure 3. Distributional One-Way Sensitivity Analysis for the Median Relative Reduction in Prescriber Shortage.

Note: New Mexico and Louisiana have been excluded, as these policies are already in place in those states. Upside and Downside represent the 95^th and 5^th distributional value from 10,000 Markov trials. FTE = full-time equivalents; Rx = Prescription.

Figure 4. Distributional One-Way Sensitivity Analysis for the Median Absolute Reduction in Prescriber Shortage.

Note: New Mexico and Louisiana have been excluded, as these policies are already in place in those states. Upside and Downside represent the 95^th and 5^th distributional value from 10,000 Markov trials.

Discussion

This study used a simple model of provider availability and prescribing need to give a first-look estimate of the potential for policies granting prescribing authority to psychologists to reduce prescriber shortages. While all states saw a reduction in prescriber shortages, we found that the psychologist prescribing licensure percentage was a critical component driving the size of the reduction. Additionally, regardless of licensure percentage, there was considerable variation in the policy impact between states. As such, adoption of this policy should be considered within the specific context of each state. Our findings appeared to be relatively robust to uncertainty regarding the parameters underlying the model, as the one-way sensitivity analyses demonstrated that the combination of FTE per appointment, FTE per psychologist, and psychologist prescribing licensure percentage explained nearly all variation in the model. These findings suggest that granting psychologists the ability to prescribe psychotropic medications would significantly reduce prescriber shortages nationally, especially if efforts are undertaken to increase the percentage of psychologists who become licensed to prescribe. At the state level, our findings suggest there is considerable variability in the impact of this policy, and the potential impact of adopting this policy should be examined on a per-state basis. The discussion below explores interpretation and the future research needed to develop a more nuanced understanding of impacts by state, small area, and the individual characteristics of both providers and patients.

Prior to contextualizing these findings, it is worth reiterating some of the boundaries and assumptions made in the framing of this simulation. Our focus was on exploring potential prescriber shortages with the understanding that psychotropic medications are considered a first-line treatment modality for people with many common mental health conditions and that the mental health workforce shortage is far greater among prescribers than non-prescribers (Thomas et al., 2009); Importantly, psychotherapy is also a first-line treatment modality that may be impacted by these prescriptive authority policies, and we discuss the need for future research on this in further detail below.

As this is a relatively new policy area, there are myriad directions for future research. However, two areas seem particularly important to examine based on our findings. First, while we demonstrated the potential increased availability of mental health prescribers, these policies do not increase the overall mental health provider workforce. It is not yet understood how granting prescribing authority to psychologists may impact the availability of the psychotherapy and testing services they normally provide. While the ideal scenario arising from these policies would be that psychologists integrate prescribing and medication management into their current practice, it is possible that these policies may result in a corresponding decline in the availability of therapy and testing services. Future studies should seek to examine how psychologist prescribing may impact the availability of therapy and testing services, as well as opportunities for synergistic policies that could bolster other available therapy providers, such as Master’s-level psychologists, counselors, and social workers.

Second, while these policies may improve the availability of psychotropic prescribing, there is currently a dearth of research empirically examining the efficacy and quality of care provided by psychologists with prescribing authority. The existing policies in New Mexico and Louisiana sought to ensure this quality of care by requiring post-doctoral training in psychopharmacology and varying degrees of oversight and collaboration with a medical doctor; however, there is some disagreement about whether these measures are sufficient (McGrath, 2020; Robiner et al., 2006, 2020). Future studies should seek to empirically examine the safety and effectiveness of the prescribing occurring under these policies in New Mexico and Louisiana. Additionally, future work is needed to identify which patients may benefit from the added medical knowledge of a psychiatrist versus the advanced diagnostic and psychotherapy knowledge of a psychologist. This is particularly important as it relates to patients with complex medical histories who may require more intricate medication management. If more policies are enacted granting psychologists prescribing authority, these issues will be important to address.

Policy Implications

Based on our findings, state-level governing bodies should re-examine the policies outlining the scope of practice for psychologists and consider the addition of prescribing authority in a manner similar to New Mexico and Louisiana. Our results suggest that codifying prescribing authority for psychologists would be a step towards achieving the “Triple Aim” of healthcare (Berwick et al., 2008), especially given that recent studies found that the passage of RxP policies produced cost-effective reductions in mental health-related mortality (Choudhury & Plemmons, 2023; Hughes et al., 2022, 2023). Such a policy would promote an improvement in the health of the population via more widely available mental health care without necessarily increasing the cost per capita. In New Mexico for instance, the Medicaid fee schedule for psychotherapy with medicine evaluation and management services is identical for both psychiatrists and psychologists with prescribing authority (New Mexico Human Services Department, 2020). Additionally, this would likely improve patient satisfaction as well by increasing patient choice when selecting a provider.

In light of the extensive state-level variation in reductions, states seeking to explore an RxP policy should consider the policy in the context of their state’s existing workforce. In the one-way sensitivity analyses, the FTE per psychologist and FTE per prescription explained a substantial portion of the variability in the estimates. This could suggest that states with more psychologists seem more likely to benefit from this policy. When viewing the results sorted by the number of psychologists (Table S2), this appears to generally be the case. For example, New York has over 14,000 total psychologists and relative reductions of over 10% while Oklahoma has less than 500 psychologists and the lowest relative reduction at 1.10%. It is important to note, however, that many states with less than 1,000 psychologists had substantial reductions in proportional prescriber shortages, such as Hawaii (922 psychologists, 10.69% relative reduction) and Vermont (493 psychologists, 8.03% relative reduction). Additionally, an RxP policy and its implementation should be designed to maximize the percentage of psychologists that become licensed to prescribe in that state given that licensure percentage played a large role in determining the reduction in unmet need. To accomplish that, there remains a need for patient outcomes research to guide policy with respect to the degree of training and oversight needed for prescribing psychologists to provide safe and effective care. Survey results from medical providers that work with prescribing psychologists overwhelmingly (95.5%) agree or strongly agree that prescribing psychologists are adequately trained to prescribe (Linda & McGrath, 2017), suggesting that less-restrictive policies, such as the one in Louisiana, may be sufficient to meet the goal of protecting patient safety while also easing the administrative burden of psychologists seeking prescriptive licensure.

Limitations

This study had several limitations. The data used for this study lacked the granularity needed to assess policy implications in greater detail. While the state and national level data were useful for evaluating the broad implications, the lack of county data limited our ability to describe policy effects in sub-state regions where significant variation in provider shortages exist. While increases in the use of telehealth may mitigate this limitation somewhat, future studies should nonetheless seek to examine how the impact of RxP might differ based on more granular spatial factors. Additionally, any bias in the count of psychologists and psychiatrists per state would have a substantial bias on the results of our study. While the psychologist and psychiatrist data from the Behavioral Health Workforce Tracker appears highly reliable in that it uses a combination of state licensure data and prescriber data to calculate their provider counts, this potential issue is worth noting. Future analyses with more granular data would also be less subject to this issue. Similarly, the estimate of needed FTE used in the calculation of prescriber shortages does not account for patient-level factors and is therefore unable to reflect the complexity of medication access at the patient level. For example, while we attempted to account for patient preference based on medication use among those receiving treatment, we were unable to directly account for individual-level patient preferences in this model despite preference being a highly salient component of patient care. A future analysis with more granular data could address this by facilitating the use of agent-based modeling techniques that can simulate individual-level barriers (e.g., costs), facilitators (e.g., patient references), and treatment decisions (e.g., treatment modality) as part of a holistic system.

We were not able to assess the differential impacts of various components of the RxP policies, such as differences in supervision requirements. The policies in New Mexico and Louisiana overlap considerably in their requirements, but there is enough heterogeneity in the specific implementation of those components that it is difficult to isolate the effects of any specific policy component as it relates to the uptake of licensure. However, we speculate that the degree of independence offered under the two policies may be a motivating factor in Louisiana’s higher rate of uptake. While New Mexico explicitly requires physician supervision and the completion of a practicum, Louisiana offers a pathway to independent prescribing and requires no such practicum. It is worth noting as well that while Louisiana had a higher uptake, the rates were not as different as they could have been, suggesting that interest from psychologists plays a substantial role in determining uptake. Future studies should seek to identify which policy components and environmental factors are likely to result in greater uptake of psychologist RxP licensure.

The limited availability of data and outcomes research regarding the impact of RxP policies had a considerable impact on our modeling approach. Despite the constraints of the available evidence, the current study attempts to fill a knowledge gap now to aid policy makers when considering RxP policies; as such, we had to develop a more complex simulation strategy. This simulation relied on several assumptions about the provision of mental health care that ignore the inherent variability involved in providing care. While these assumptions were based on current practice and available literature where possible, it is important to acknowledge that the estimates of prescriber shortages produced by them are intended for simulation purposes and likely do not reflect the reality of these shortages as experienced by patients. For example, we were not able to include variation in practice patterns and patient preferences across states, though these almost certainly exist. To the extent that providers in each state may be more or less inclined to treat using medications, findings will over or underestimate the potential impacts of the policy. Details on general and individual preferences for psychotropic medications could provide important information to guide policy decisions per state. We attempted to check our results for robustness to assumption modifications through our uncertainty and sensitivity analyses, which demonstrated consistent relationships; however, caution is still warranted. A future study evaluating the impact of the RxP policies in Illinois, Iowa, Idaho, and Colorado and extending those findings to other states would be an ideal step to replicate these findings.

Finally, it is important to note that primary care providers were not included in our estimate of available prescribing FTE in the same way as specialty providers despite their role in providing mental health care. Due to their broad scope of practice, the direct inclusion of primary care providers would have resulted in a significant overestimation of available prescribing FTE and underestimation of prescriber shortages. In contrast, the complete exclusion of primary care providers would have resulted in an overestimation of prescriber shortages, which could have led to larger estimated effect size. We attempted to strike a balance in addressing this by deflating the estimates of prescribing need in order to account for the percent of mental health practice that might be addressed in primary care. Similarly, we did not directly include psychiatric nurse practitioners in our model due to the absence of available data on these providers and their limited prevalence (Andrilla et al., 2018). However, these providers clearly have a growing role in mental health prescribing (Delaney & Vanderhoef, 2019) and should be included in future models when data become available. More broadly, RxP policies likely have a complex impact on the mix of providers prescribing psychotropic medications unable to be accounted for in the current simulation. Future studies should describe any task-shifting between providers that may result from RxP policies.

Conclusion

There is currently an insufficient workforce available to provide psychotropic medications for the individuals with mental illness for whom such a treatment may be beneficial. This is the first study to quantitatively estimate the impact that prescriptive authority for psychologists would have on prescriber shortages. We simulated the potential outcomes of granting prescriptive authority to licensed psychologists at a state level in a manner similar to New Mexico and Louisiana, and our results suggest these policies could substantially increase provider availability and reduce prescriber shortages for some states. The resulting overall increase in prescribing has the potential to greatly improve the availability of much-needed mental health care and offer patients more choice when deciding where to receive treatment. Future studies are needed to explore how other mental health practitioners may be impacted by such a policy, to examine how specific policy requirements may impact licensure rates, and to determine what impact such a policy may have on the quality of care received.

Public Significance Statement:

Prescriptive authority for psychologists could reduce mental health prescriber shortages in the United States by approximately 4%. Policymakers seeking to improve access to mental health care should consider policies granting prescriptive authority to psychologists.

Biographies

PHILLIP M. HUGHES received his MS in clinical psychology from Auburn University at Montgomery. He is currently pursuing his PhD in Pharmaceutical Outcomes and Policy at UNC Chapel Hill, a pre-doctoral fellow at the Cecil G. Sheps Center for Health Services Research, and a research data analyst at UNC Health Sciences at MAHEC. His areas of interest include mental health and substance use policy with a focus on scope-of-practice and improving access to and quality of mental health and substance use care.

ROBERT E. MCGRATH received his MS and PhD in clinical psychology from Auburn University. He is currently a Professor in the School of Psychology and Counseling at Fairleigh Dickinson University in Teaneck NJ. His areas of research interest include character measurement and education and professional issues for psychologists.

KATHLEEN C THOMAS received her MPH from Yale University and her PhD in health economics from the University of North Carolina at Chapel Hill. She is currently Associate Professor in the Division of Pharmaceutical Outcomes and Policy, Eshelman School of Pharmacy, UNC Chapel Hill, where she serves as Vice Chair for Research and Graduate Education and Associate Director of the Cecil G Sheps Center AHRQ training program in health services research. Her work focuses on improving access to and quality of care for children and adults with psychiatric conditions and disabilities.