Abstract
BACKGROUND:
Given provider, patient, and health plan administrative burden, there is consensus that the medication prior authorization (PA) process needs to be streamlined and standardized. This includes broader adoption of electronic PA (ePA) solutions, including real-time clinical benefit determination. Insight into provider experiences with ePA will help health plans and pharmacy benefit managers (PBMs) prioritize efforts and investments in PA process improvement.
OBJECTIVE:
To identify provider experiences with PA by their use of ePA technology in a national sample of providers from the United States.
METHODS:
An online, nationwide survey was sent to licensed providers in 2020. The survey obtained information on provider experiences with PA, including PA workload, time to the communication of a PA decision, and other PA-related interactions with plans/PBMs. Provider characteristics and survey responses by provider use of ePA were compared using analysis of variance tests for items based on ordinal scales and a chi-square test statistic for nominal variables. Multivariable regression analyses identified associations between ePA use and PA experience variables, controlling for provider characteristics.
RESULTS:
Of 1,147 providers submitting usable surveys, 58% reported personally using ePA for at least some of their PA submissions, 88% submit at least 1 PA in a typical week, and most (82%) spend up to 5 hours a week on PA submissions. A majority (58.5%) reported that manual PA is often required. Those using ePA submitted a higher volume of PAs (P < 0.001) and spent more time on PA submissions (P = 0.003) than providers not using ePA software, but the duration of time from start to finish for a PA submission did not differ (P = 0.211). Providers who use ePA reported more difficulty identifying step therapy requirements (P = 0.005) and more frequently needing to submit additional documentation (P = 0.022). PA-related communication failures did not differ. Those using ePA reported a shorter time to PA decision (P = 0.004) than those not using ePA. Univariate descriptive findings were supported by multivariable analyses.
CONCLUSIONS:
This large, nationwide survey identified that a provider’s use of ePA was not associated with less provider time or fewer challenges in preparing and submitting PA requests. However, the use of ePA was associated with a reported shorter PA decision time. Efforts to standardize PA requirements, support of real-time benefits check functionality, and better integration of ePA processes into provider workflows may help reduce PA burden and treatment delays.
Plain language summary
Health plans must approve the use of some drugs before they will cover those drugs. This study surveyed doctors and other prescribing health care providers on this approval process. It found that providers who use software to ask for approval generally do not find the process quicker or easier, but health plans may take less time to reach a decision. Improvements are needed in the process and software to reduce the time doctors spend on requests.
Implications for managed care pharmacy
Providers who use electronic prior authorization (ePA) report a quicker time to approval but no improvement in the time spent on prior authorization (PA) or communication with health plans. Partial adoption of ePA plus the lack of a real-time exchange of PA requirements, clinical information, and decisions are possible reasons why provider burden is not lessened. The adoption of a real-time benefits check and better integration of ePA processes into provider workflows should reduce PA burden and reduce PA-related treatment delays.
Managed care organizations use prior authorization (PA) as a utilization management strategy to promote the safe and appropriate use of medications. PA can help health plans manage costs of care by maximizing the therapeutic value of drug treatment.1-3 However, PA is also associated with administrative burden to providers.4-6 Noted challenges providers face include difficulty determining if a medication requires PA and, if so, payer-specific requirements for review and approval. PA is also associated with inefficiencies due to the provider and staff needing to navigate various communications channels (eg, fax, telephone, and web portals). Finally, the time to payers’ PA decisions is unpredictable and often lengthy.7-9
A documented consequence of the PA process is treatment delay.4 PA delays may start from the time of prescribing, with 60% of medical practices reporting that they do not initiate PA requests until made aware of the need for PA by the dispensing pharmacy.10 Once submitted, 16% of providers wait a few hours or less for a health plan’s response, whereas more than half (64%) report waiting 1 business day or longer.10 Unfortunately, delays can lead to nonadherence, with estimates of PA-related prescription abandonment rates ranging from 37% to 78%.10,11 Regardless of cause, nonadherence is a concern because of its adverse impact on health outcomes.
Given these issues, numerous medical and pharmacy organizations endorse changes in the PA approach to ease the burden on providers and patients.7,12 Recommendations include judicious application of PA to medications, improvement in PA processes, and standardization of PA criteria. These same entities also support the adoption of electronic PA (ePA) solutions, which allow for digital preparation and submission of some or all aspects of PA requests. ePA tools may be incorporated into an electronic health record (EHR) system or may function outside of the EHR. ePA can facilitate aspects of the PA process, such as allowing providers to submit PA requests and supporting documentation and receive PA determinations from health plans.13
Research has shown that ePA can decrease processing and turnaround time for PA decisions.13-15 However, studies also highlight barriers to ePA use, including no access and incomplete or inaccurate information on PA requirements at the time of prescribing.13,14 Furthermore, ePA solutions vary in functionality in terms of integration with the provider’s EHR, and access to benefit information varies by ePA tool and health plan/pharmacy benefit manager (PBM). Given the current limitations, providers using ePA may engage multiple PA processes, including the occasional use of manual processes. These factors can increase cognitive load, increase administrative costs, and create delays in PA submissions.6,14
However, the literature evaluating how ePA affects provider burden and PA-related communications between providers and plans/PBMs is limited. This gap in the literature is relevant, as plans/PBMs seek to improve PA processes through ePA. Therefore, this study’s objective is to describe provider experiences with PA by their use of ePA technology in a national sample of providers in the United States.
Methods
This survey study, conducted in late 2020, was conducted as part of an effort to design a digital solution for medication PA funded in part by a National Institute of Mental Health Small Business Technology Transfer Grant. The study sampled providers in specialties most likely to prescribe PA medications.16 Psychiatry providers were oversampled, as they are the current focus of the digital “software-as-a-solution” platform under development by the study investigators.
Investigators developed the survey content based on focus groups and stakeholder interviews. Select survey items were designed to identify provider experiences with PA, including PA volume and time spent on PA processes; time to payer communication of a PA decision; and interactions with plans/PBMs, including PA requirements and communication issues. Survey items included in this article are provided in Supplementary Survey 1 (206.1KB, pdf) (available in online article).
Invitation emails with an embedded hyperlink were sent in October and November of 2020 to 100,000 licensed MD/DO, nurse practitioner (APRN), and physician assistant providers identified from a curated mailing list. Data were collected anonymously, and all participants provided electronic consent prior to taking the survey (University of Nebraska, IRB no. 00000672).
Descriptive statistics were used to report provider characteristics and survey responses by the provider self-reported use of ePA. The number and percentage of responders are provided for each response option. Some item categories were collapsed to facilitate results interpretation. Analysis of variance tests were used to compare results by ePA use for survey items based on ordinal scales, and a chi-square test statistic was used for nominal variables. Multivariable ordered logistic regression analyses were performed to identify associations between ePA use and measures of provider PA experience, controlling for provider age, sex, specialty, patient volume, and the provider’s average PA volume. An a priori power analysis indicated that a sample of 1,150 providers would appropriately power the study analyses.
Results
Respondent individual and practice characteristics are reported in Table 1. Of the 100,000 providers contacted, a total 1,173 responded to the survey (1.2% response rate), of which 1,147 surveys were at least 50% complete and included in these analyses. Of these responses, 58% reported personally using ePA for at least some of their PA submissions. In terms of respondent demographics, 73% were White, 53% were female, and the mean age (SD) was 51 (13) years. A majority (76%) were physicians (MD or DO). Responses by specialty largely reflect the sampled population, with psychiatry providers representing the largest proportion by specialty (45%), followed by internal medicine (18%) and dermatology (13%) (Supplementary Table 1 (206.1KB, pdf) ).
Table 1.
Characteristics of Responding Providers by the Provider’s Use of ePA Software
| Overall | Does not use ePA | Uses ePA | P value b | |
|---|---|---|---|---|
| No. of responses a | 1,147 | 481 | 663 | |
| Mean (SD) age (n = 985), years | 50.6 (13.8) | 51.3 (13.8) | 50.1 (12.0) | 0.156 |
| Age by categories, n (%) | ||||
| < 40 years | 243 (24.7) | 102 (25.5) | 141 (24.1) | 0.064 |
| 40-49 years | 236 (23.9) | 88 (22.0) | 148 (25.3) | |
| 50-59 years | 232 (23.6) | 83 (20.8) | 149 (25.5) | |
| ≥ 60 years | 274 (27.8) | 127 (31.7) | 147 (25.1) | |
| Sex (n = 1,075) | ||||
| Female | 569 (52.9) | 215 (48.3) | 354 (56.2) | 0.011 |
| Male | 506 (47.1) | 230 (51.7) | 276 (43.8) | |
| Race (n = 1,023), n (%) | ||||
| White | 748 (73.1) | 310 (73.3) | 438 (73) | 0.368 |
| Black | 22 (2.2) | 6 (1.4) | 16 (2.7) | |
| Asian | 197 (19.3) | 88 (20.8) | 109 (18.2) | |
| Other | 33 (3.2) | 10 (2.4) | 23 (3.8) | |
| Mixed | 23 (2.3) | 9 (2.1) | 14 (2.3) | |
| Hispanic/Latino ethnicity (n = 994), n (%) | 46 (4.6) | 10 (2.4) | 36 (6.2) | 0.006 |
| Practitioner type (n = 1,147), n (%) | ||||
| Physician (MD/DO) | 872 (76) | 395 (82.1) | 474 (71.5) | < 0.001 |
| Nurse practitioner | 158 (13.8) | 56 (11.6) | 102 (15.4) | |
| Physician assistant | 117 (10.2) | 30 (6.2) | 87 (13.1) | |
| Practice size (n = 1,142), n (%) | ||||
| < 5 prescribers | 503 (44.2) | 170 (35.4) | 333 (50.5) | < 0.001 |
| 5-15 prescribers | 365 (32.1) | 161 (33.5) | 204 (31) | |
| > 15 prescribers | 271 (23.8) | 149 (31) | 122 (18.5) | |
| Patient panel size (n = 1,142), n (%) | ||||
| ≤ 100 patients | 428 (37.5) | 223 (46.7) | 205 (31) | < 0.001 |
| 101-200 patients | 162 (14.2) | 63 (13.2) | 99 (15) | |
| > 200 patients | 550 (48.3) | 192 (40.2) | 358 (54.1) | |
| Specialty (n = 1,147), n (%) | ||||
| Psychiatry | 513 (44.8) | 211 (43.9) | 302 (45.6) | 0.575 |
| Internal medicine | 206 (18) | 101 (21) | 105 (15.8) | 0.025 |
| Dermatology | 149 (13) | 40 (8.3) | 109 (16.4) | < 0.001 |
| Gastroenterology | 92 (8) | 46 (9.6) | 46 (6.9) | 0.107 |
| Neurology | 68 (5.9) | 28 (5.8) | 40 (6) | 0.881 |
| Oncology | 66 (5.8) | 37 (7.7) | 29 (4.4) | 0.017 |
| Rheumatology | 50 (4.4) | 18 (3.7) | 32 (4.8) | 0.376 |
| No. of PAs completed per week (n = 1,143), n (%) | ||||
| None | 251 (22) | 155 (32.2) | 96 (14.6) | < 0.001 |
| 1-5 | 523 (45.8) | 215 (44.7) | 308 (46.7) | |
| 6-20 | 306 (26.8) | 92 (19.1) | 214 (32.4) | |
| > 20 | 61 (5.4) | 19 (4) | 42 (6.4) | |
| Hours on PA in a week (n = 1,147), n (%) | ||||
| < 5 | 936 (81.8) | 411 (85.5) | 525 (79.2) | |
| 5-10 | 162 (14.2) | 58 (12.1) | 104 (15.7) | 0.003 |
| > 10 | 46 (4) | 12 (2.5) | 34 (5.1) | |
aNot all questions were answered by each respondent; 1,144 responded to ePA use.
bAnalysis of variance for ordinal scale variables, chi-square test for nominal variables.
ePA = electronic prior authorization; PA = prior authorization.
In evaluating respondents by ePA use, greater proportions of women (P = 0.011) and providers of Hispanic/Latino ethnicity (P = 0.006) reported using ePA. Providers who report using ePA platforms come from practices with fewer prescribers but with higher patient loads (P < 0.001 for both). Dermatology providers were more likely to use ePA (P < 0.001), whereas internal medicine and oncology providers reported lower ePA use (P < 0.05 for both). Note, the survey questions queried the provider’s personal experience with PA volume and workload and their use of ePA and not that of the overall practice.
PROVIDER EXPERIENCE WITH PA
Overall, 88% of providers reported personally submitting at least 1 PA in a typical week, with 5.4% submitting more than 20 PAs in a week (Table 2). Most providers (82%) reported that they spend up to 5 hours a week on PA submissions. Those using ePA for at least some PA requests submitted a higher volume of PAs (P < 0.001) and spent more time on PA submissions (P = 0.003) than providers who reported not using ePA. Sixty percent reported that from start to finish, it takes 1 day or less to complete a PA submission, including gathering clinical documentation; 29% reported that the start-to-finish duration is 1 to 2 weeks. However, the duration of time from start to finish for a PA submission did not differ by ePA use (P = 0.211).
Table 2.
Provider Experience With Prior Authorizations by Provider’s Use of ePA Software
| Overall n (%) | Does not use ePA n (%) | Uses ePA n (%) | P value b | |
|---|---|---|---|---|
| No. of PAs completed per week (n = 1,143) | ||||
| None | 251 (22) | 155 (32.2) | 96 (14.6) | < 0.001 |
| 1-5 | 523 (45.8) | 215 (44.7) | 308 (46.7) | |
| 6-20 | 306 (26.8) | 92 (19.1) | 214 (32.4) | |
| > 20 | 61 (5.4) | 19 (4) | 42 (6.4) | |
| Hours on PA in a week (n = 1,147) | ||||
| < 5 | 936 (81.8) | 411 (85.5) | 525 (79.2) | 0.003 |
| 5-10 | 162 (14.2) | 58 (12.1) | 104 (15.7) | |
| > 10 | 46 (4) | 12 (2.5) | 34 (5.1) | |
| Time to complete a PA submission: from start to finish (n= 1,138) | ||||
| ≤ 1 day | 686 (60.3) | 278 (58) | 408 (61.9) | 0.211 |
| 1-7 days | 334 (29.4) | 148 (30.9) | 186 (28.2) | |
| > 1 week | 118 (10.4) | 53 (11.1) | 65 (9.9) | |
| Average time for a PA response from health plan/PBM (n = 1,003) | ||||
| ≤ 1 business day | 365 (36.3) | 133 (33.9) | 232 (38) | 0.004 |
| 2 business days | 270 (26.9) | 93 (23.7) | 176 (28.8) | |
| 3-5 business days | 289 (28.8) | 123 (31.4) | 166 (27.2) | |
| > 5 business days | 81 (8.1) | 43 (11) | 37 (6.1) | |
| How challenging is it to find step therapy requirements (n = 1,136) | ||||
| Not at all | 103 (9.1) | 52 (10.9) | 51 (7.8) | 0.005 |
| Somewhat | 354 (31.2) | 158 (33.1) | 196 (29.8) | |
| Moderately | 331 (29.1) | 143 (29.9) | 188 (28.6) | |
| Highly | 218 (19.2) | 89 (18.6) | 129 (19.6) | |
| Extremely | 130 (11.4) | 36 (7.5) | 94 (14.3) | |
| How often do you need to provide additional documentation (n = 1,119) | ||||
| Very rarely or never | 56 (5) | 31 (6.5) | 25 (3.9) | 0.022 |
| Rarely | 197 (17.6) | 79 (16.7) | 118 (18.3) | |
| Often | 413 (36.9) | 190 (40.1) | 223 (34.6) | |
| Very often | 307 (27.4) | 124 (26.2) | 183 (28.4) | |
| Extremely often | 146 (13.1) | 50 (10.6) | 96 (14.9) | |
| How often is manual PA processing required (n = 1,112) | ||||
| Very rarely or never | 122 (11) | 53 (11.4) | 69 (10.7) | 0.012 |
| Rarely | 339 (30.5) | 118 (25.3) | 221 (34.3) | |
| Often | 409 (36.8) | 178 (38.1) | 231 (35.8) | |
| Very often | 165 (14.8) | 79 (16.9) | 86 (13.3) | |
| Extremely often | 77 (6.9) | 39 (8.4) | 38 (5.9) | |
| How often are you not notified of PA approval (n = 1,119) | ||||
| Very rarely or never | 114 (10.2) | 53 (11.2) | 61 (9.4) | 0.238 |
| Rarely | 328 (29.3) | 143 (30.3) | 185 (28.6) | |
| Often | 356 (31.8) | 146 (30.9) | 210 (32.5) | |
| Very often | 202 (18.1) | 83 (17.6) | 119 (18.4) | |
| Extremely often | 118 (10.6) | 47 (10) | 71 (11) | |
| How often are you not notified of PA denial (n = 1,114) | ||||
| Very rarely or never | 137 (12.3) | 69 (14.7) | 68 (10.6) | 0.185 |
| Rarely | 384 (34.5) | 158 (33.6) | 226 (35.2) | |
| Often | 333 (29.9) | 138 (29.3) | 195 (30.3) | |
| Very often | 166 (14.9) | 69 (14.7) | 97 (15.1) | |
| Extremely often | ||||
| How often software indicates PA was denied when it was actually approved (n = 1,112) | ||||
| Very rarely or never | 246 (22.2) | 131 (28) | 115 (17.9) | < 0.001 |
| Rarely | 408 (36.8) | 183 (39.1) | 225 (35.1) | |
| Often | 308 (27.8) | 111 (23.7) | 197 (30.7) | |
| Very often | 96 (8.7) | 28 (6) | 68 (10.6) | |
| Extremely often | 51 (4.6) | 15 (3.2) | 36 (5.6) | |
a Not all questions were answered by each respondent; 1,144 responded to ePA use.
b Analysis of variance for ordinal scale variables, chi-square test for nominal variables.
ePA = electronic prior authorization; PA = prior authorization; PBM = pharmacy benefit manager.
When providers were asked how difficult they find it to identify step therapy requirements for PA, 60% reported that it is moderately to extremely challenging to identify step therapy requirements. In addition, a majority (77%) reported that they often, very often, or extremely often need to provide additional documentation for PA approval. These challenges were more frequently noted by ePA users compared with nonusers (P < 0.05 for both). A majority (58.5%) reported that manual PA is often to extremely often required by health plans/PBMs, although those using ePA reported that manual PA is required less frequently than those who do not use ePA (P = 0.012).
When asked how long they wait on average from the health plan for a PA response, 36% report receiving a PA response in 1 business day or less. When evaluated by ePA use, 38.0% of providers who use ePA report that they receive a PA response in 1 business day or less vs 33.9% of those who report not using ePA (P = 0.194). The overall time to a PA decision differed by ePA use; those using ePA responded that the typical response time from the health plan is shorter (P = 0.004).
Finally, providers report that health plan PA decisions were not well communicated to them, with 60% and 53% reporting that they were often to extremely often not notified of a PA approval or denial, respectively. These communication challenges did not differ by ePA use. However, almost 47% of ePA users reported that they were often to extremely often notified by the ePA software that the PA request was denied but later discovered that the PA request was, in fact, approved.
The findings of the descriptive analyses were upheld in multivariable logistic regression analyses examining the association between PA experiences and ePA use, controlling for provider characteristics, including specialty, patient load, and PA volume (Table 3; Supplementary Tables 2-7 (206.1KB, pdf) ).
Table 3.
Multivariable Analyses of Association of Provider’s Use of ePA Software With PA-Related Experiences (n = 974)
| Survey question | Odds ratio a | SE | P value | 95% CI |
|---|---|---|---|---|
| Time to complete a PA submission from start to finishb | 0.870 | 0.122 | 0.323 | (0.661-1.146) |
| Average time waiting for a PA response from health plansb | 0.573 | 0.079 | < 0.001 | (0.438-0.751) |
| How challenging is it to find step therapy requirementsc | 1.223 | 0.151 | 0.103 | (0.96-1.557) |
| How often do you need to provide additional documentationd | 1.059 | 0.132 | 0.648 | (0.829-1.352) |
| How often are you not notified of PA approvald | 1.029 | 0.128 | 0.819 | (0.807-1.312) |
| How often are you not notified of PA deniald | 1.041 | 0.131 | 0.748 | (0.814-1.332) |
aAnalyses: ordered logistic regression with all models controlling for age, sex, race, specialty, number of patients in practice, and number of PAs submitted by the provider in an average week.
bOdds of being in a higher time category with ePA use vs no ePA use
cOdds of finding more challenges with ePA use vs no ePA use.
dOdds of event occurring more often with ePA use vs no ePA use.
ePA = electronic prior authorization; PA = prior authorization.
Discussion
This article presents results related to provider perceptions of their PA experiences, with more than half of the survey respondents reporting that they use ePA in practice. Overall, this survey identified that responding providers who use ePA do not spend less time preparing and submitting PA requests or face fewer challenges with PA submissions than those who do not use ePA. However, provider use of ePA was associated with reduced time for a PA decision once submitted.
This survey contributes to the existing research.17 Although much of the prior PA research has assessed provider burden and outcomes related to PA at the practice level, our survey asked providers about their personal experience, which makes a direct comparison between this study and others more tenuous. For instance, the American Medical Association (AMA) reported that medical practices complete an average of 41 PA requests per week per physician,4 whereas a majority of providers responding to the current study reported that they personally submit 5 or fewer PAs in a typical week. The difference in PA burden between studies may reflect that a portion of the practice PA burden is handled by administrative and other clinical staff. Considering this context, our survey explores the prescriber’s unique personal experience and helps to highlight if and how incorporation of ePA into their workflows influences that experience.
Several expected benefits of ePA use were not noted by survey respondents, such as our expectation that providers using ePA would spend less of their time on PA submissions. However, prior research has highlighted barriers to the implementation of ePA in the United States by providers that may help explain this finding. One recent survey reported that despite the substantial integration of ePA software into EHR, pharmacy, and payer system platforms, providers report maintaining multiple processes, including submitting some PA requests manually.6 Further, without a real-time clinical benefit check at the time of prescribing, a PA requirement may not be realized until a prescription claim is denied at the pharmacy. In such cases, even with the use of ePA, this process falls outside of the clinical workflow, which may be offset by other efficiency gains.
The current study identified other issues with PA that appear to be either more evident or more problematic for providers who use ePA software. This includes challenges in identifying step therapy requirements and how often providers find it necessary to send additional clinical documentation for coverage authorization. This finding highlights that although ePA can facilitate the completion and submission of PA request forms, real-time access to current PA clinical criteria and documentation requirements is lacking. The standardization of PA requirements and adoption of data standards for real-time communication between providers and plans/PBMs, including benefit checks, will be essential to minimize the administrative burden of PA.
Another issue that could be reduced by reciprocal communication with ePA use is effective and accurate communication of PA decisions between health plans/PBMs and providers. In the current study, providers reported that they are often not made aware of PA decisions or were informed by their ePA platform that a medication request was denied when, in fact, it was approved by the health plan. This survey does not examine causes of communication delays and failures. However, a recent initiative to facilitate ePA adoption by plans/PBM and providers found that the level of ePA adoption may impact providers’ ability to navigate ePA for information. Notably, providers who used ePA for most patients reported that they found it easier to view PA decisions than providers who used ePA for some or a few of their patients.13 Others have noted issues leading to breakdowns in communication, such as providers not knowing to or forgetting to check an inbox for PA denials.14 Thus, communication issues may not be driven by failure to communicate per se, rather the communication of decisions may not be optimally integrated into clinical workflows. Thus, plans/PBMs and ePA developers should work with providers to integrate PA decisions into existing provider workflows.
Overall, the current survey identifies opportunities for ePA process improvements that are aligned with a consensus statement on improving the PA process sponsored by the AMA, BlueCross BlueShield Association, American’s Health Insurance Plans, and others regarding greater transparency and communication and the automation of PA processes with ePA solutions.7,12 We contend that these steps are interrelated, as transparency and communication must be improved overall and supported by ePA solutions for optimal provider adoption. This is not to minimize the importance of other recommendations regarding PA improvement, including the judicious use of PA and the regular review of PA medications with volume adjustment.7,12 For instance, we identified differences in PA volume and provider burden by provider specialty (data not shown). These findings indicate the need to examine PA burden at the specialty and/or therapeutic level to avoid over burdening select providers and, by extension, their patients.
We believe that plans/PBMs can play a key role in reducing PA burden for providers as well as their own organizations by standardization and support for real-time benefit check technology. However, if the ePA solutions fail to consider existing clinical workflows, providers may find it impractical to integrate ePA into their practice. Ideally, the development of ePA solutions should take a user-centered approach. Design thinking is one such mechanism, defined as “a human-centered approach to innovation that draws from the designer’s toolkit to integrate the needs of people, the possibilities of technology, and the requirements for business success.”18 Designed using a human-centered approach, the optimal ePA solution would consider the unique needs of providers, other health professionals involved in the PA process, and patients, as well as the requirements of plans/BPMs.
LIMITATIONS
This survey-based study identified differences in provider experience with PA processes by ePA use. Its strengths are a large, nationwide sample with more than 1,000 usable survey responses, making it one of the largest provider surveys examining the effects of PA. However, the survey response rate was low. Although low survey participation by physicians is well documented,19 it may diminish the representativeness of the survey responses. The demographic profile of those who responded matched well to national physician characteristics, but there is a risk of nonresponse bias based on unmeasured factors. Second, this survey was cross-sectional and can be used to examine associations, but it does not establish temporal relations or causation. Further, we asked if providers used ePA software but did not specifically ask providers for the proportion of PA submissions that they submit electronically or to differentiate their PA experiences by ePA vs manual submissions. Thus, the findings reflect their overall experience, and findings may be confounded by unmeasured factors that led providers to adopt an ePA tool. Finally, this survey represents provider perceptions of their experience. Future research that also includes clinic staff, that uses a prestudy-poststudy design, and that has more objective measures of PA volume and burden would be insightful for exploring causal relationships between ePA use, PA burden, and communication challenges.
Conclusions
This large, nationwide survey of providers identified that providers who use ePA do not report spending less time preparing and submitting PA requests and do not face fewer challenges with PA submissions than those who do not use ePA. However, ePA use was associated with a shorter average time for a PA decision by the plan/PBM once submitted. Efforts to standardize PA and support real-time clinical benefits checks in a manner that is optimally integrated into provider workflows should help to reduce the burden of PA on providers and ultimately help reduce PA-related delays in treatment.
References
- 1.Agha L, Kim S, Li D. Insurance design and pharmaceutical innovation. Working Paper 27563. National Bureau of Economic Research. 2020. Accessed February 7, 2021. https://www.nber.org/papers/w27563 [Google Scholar]
- 2.2018-2019 Academy of Managed Care Pharmacy Professional Practice Committee. Prior authorization and utilization management concepts in managed care pharmacy. J Manag Care Spec Pharm. 2019;25(6):641-4. doi: 10.18553/jmcp.2019.19069 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Dieguez G, Saehney TG, Mirchandani H. Evolution of the use of restrictions in commercial formularies. Milliman, Inc. 2016. Accessed January 14, 2022. https://www.milliman.com/-/media/milliman/importedfiles/uploadedfiles/insight/2016/evolution-restrictions-commercial-formularies.ashx [Google Scholar]
- 4.American Medical Association. 2021. AMA prior authorization (PA) physician survey. American Medical Association. Accessed January 17, 2022. https://www. ama-assn.org/system/files/2021-04/prior-authorization-survey.pdf [Google Scholar]
- 5.American Medical Association. Measuring progress in improving prior authorization. American Medical Association. Accessed January 17, 2022. https://www.ama-assn.org/system/files/2021-05/prior-authorization-reform-progress-update.pdf [Google Scholar]
- 6.CoverMyMeds. Electronic prior authorization. CoverMyMeds. Accessed January 16, 2021. https://www.covermymeds.com/main/insights/scorecard/
- 7.AMCP Partnership Forum: Optimizing prior authorization for appropriate medication selection. J Manag Care Spec Pharm. 2020;26(1):55-62. doi: 10.18553/jmcp.2020.26.1.55 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Jones LK, Ladd IG, Gionfriddo MR, Gregor C, Evans MA, Graham J. Medications requiring prior authorization across health insurance plans. Am J Health Syst Pharm. 2020;77(8):644-8. doi: 10.1093/ajhp/zxaa031 [DOI] [PubMed] [Google Scholar]
- 9.Office of the National Coordinator for Health IT, Health and Human Services. Request for information: Electronic prior authorization standards, implementation specifications, and certification criteria. Federal Register. 2022. Accessed June 3, 2022. https://www.federalregister.gov/documents/2022/01/24/2022-01309/request-for-information-electronic-prior-authorization-standards-implementation-specifications-and [Google Scholar]
- 10.CoveryMyMeds. Electronic prior authorization. Accessed July 15, 2022. https://www.covermymeds.com/main/medication-access-report/electronic-prior-authorization/
- 11.CoverMyMeds. 2020 medication access report. Accessed June 6, 2021. https://assets.ctfassets.net/2in405srp47m/4SyH0ZdIFQwuAuCzhuAmTo/973d7b3266a843c94c074fce698de9ea/CMM36517MARExecutiveSummary Digital.pdf
- 12.Consensus statement on improving the prior authorization process. American Medical Association. Accessed January 15, 2021. https://www.ama-assn.org/sites/ama-assn.org/files/corp/media-browser/public/arc-public/prior-authorization-consensus-statement.pdf
- 13.RTI International. Evaluation of the fast prior authorization technology highway demonstration: Final report. AHIP. 2021. Accessed January 26, 2022. https://www.ahip.org/wp-content/uploads/Fast-PATH-Evaluation.pdf [Google Scholar]
- 14.Lauffenburger JC, Stults CD, Mudiganti S, et al. Impact of implementing electronic prior authorization on medication filling in an electronic health record system in a large healthcare system. J Am Med Inform Assoc. 2021;28(10):2233-40. doi: 10.1093/jamia/ocab119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Birdsall AD, Kappenman AM, Covey BT, Rim MH. Implementation and impact assessment of integrated electronic prior authorization in an academic health system. J Am Pharm Assoc (2003). 2020;60(4):e93-9. doi: 10.1016/j.japh.2020.01.012 [DOI] [PubMed] [Google Scholar]
- 16.American Association of Medical Colleges. Physician specialty data report 2019: Number of people per active physician by specialty. AAMC. 2020. Accessed January 14, 2022. https://www.aamc.org/what-we-do/mission-areas/health-care/workforce-studies/interactive-data/number-people-active-physician-specialty-2019 [Google Scholar]
- 17.Petitt CE, Kiracofe E, Adamson A, Barbieri JS. Prior authorizations in dermatology and impact on patient care: An updated survey of US dermatology providers and staff by the American Academy of Dermatology. Dermatol Online J. 2021;27(1):13030/qt990631n9. [PubMed] [Google Scholar]
- 18.Brown T. Change by Design, Revised and Updated. Harper Collins; 2019. [Google Scholar]
- 19.Dykema J, Jones NR, Piché T, Stevenson J. Surveying clinicians by web: Current issues in design and administration. Eval Health Prof. 2013;36(3):352-81. doi: 10.1177/0163278713496630 [DOI] [PubMed] [Google Scholar]