Comparison of prior authorization across insurers: cross sectional evidence from Medicare Advantage
BMJ 2024; 384 doi: https://doi.org/10.1136/bmj-2023-077797 (Published 07 March 2024) Cite this as: BMJ 2024;384:e077797
- Ravi Gupta, assistant professor and staff physician12,
- Jay Fein, data scientist3,
- Joseph P Newhouse, professor4567,
- Aaron L Schwartz, assistant professor and staff physician8910
- 1Division of General Internal Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, USA
- 2Department of Health Policy and Management, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA
- 3Medidata Solutions, New York, NY, USA
- 4Department of Health Care Policy, Harvard Medical School, Boston, MA, USA
- 5Department of Health Policy and Management, Harvard TH Chan School of Public Health, Boston, MA, USA
- 6Harvard Kennedy School, Cambridge, MA, USA
- 7National Bureau of Economic Research, Cambridge, MA, USA
- 8Department of Medical Ethics and Health Policy, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
- 9Division of General Internal Medicine, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
- 10Michael J Crescenz VA Medical Center, Philadelphia, PA, USA
- Correspondence to: R Gupta ravigupta{at}jhmi.edu
- Accepted 13 February 2024
Abstract
Objective To measure and compare the scope of US insurers’ policies for prior authorization (PA), a process by which insurers assess the necessity of planned medical care, and to quantify differences in PA across insurers, physician specialties, and clinical service categories.
Design Cross sectional analysis.
Setting PA policies for five insurers serving most of the beneficiaries covered by privately administered Medicare Advantage in the US, 2021, as applied to utilization patterns observed in Medicare Part B.
Participants 30 540 086 beneficiaries in traditional Medicare Part B.
Main outcome measures Proportions of government administered traditional Medicare Part B spending and utilization that would have required PA according to Medicare Advantage insurer rules.
Results The insurers required PA for 944 to 2971 of the 14 130 clinical services (median 1899; weighted mean 1429) constituting 17% to 33% of Part B spending (median 28%; weighted mean 23%) and 9% to 41% of Part B utilization (median 22%; weighted mean 18%). 40% of spending ($57bn; £45bn; €53bn) and 48% of service utilization would have required PA by at least one insurer; 12% of spending and 6% of utilization would have required PA by all insurers. 93% of Part B medication spending, or 74% of medication use, would have required PA by at least one Medicare Advantage insurer. For all Medicare Advantage insurers, hematology and oncology drugs represented the largest proportion of PA spending (range 27-34%; median 33%; weighted mean 30%). PA rates varied widely across specialties.
Conclusion PA policies varied substantially across private insurers in the US. Despite limited consensus, all insurers required PA extensively, particularly for physician administered medications. These findings indicate substantial differences in coverage policies between government administered and privately administered Medicare. The results may inform ongoing efforts to focus PA more effectively on low value services and reduce administrative burdens for clinicians and patients.
Introduction
Governments must decide whether and how private firms should provide health insurance in their countries. Private provision of health insurance occurs not only in the US but also in several European countries, including Switzerland, Germany, and the Netherlands. The US, however, is unique in offering a mix of public and private health insurance programmes.1 The largest segment of the US population (49%) in 2022 enrolled in private health insurance plans sponsored by their employers. Public health insurance programmes include, among others, Medicare for elderly and disabled individuals (15%) and Medicaid for people with a low income (21%). Medicare offers beneficiaries the choice of enrolling in fee-for-service, government administered health insurance or plans that are funded by the government but administered by private health insurance firms (see box for further details).
Glossary of terms
Durable medical equipment—Medical devices and equipment that must be prescribed by a healthcare professional and are designed for reuse, usually in the home, and are appropriate for patients with certain medical conditions. Examples of durable medical equipment include wheelchairs and scooters, walkers and canes, hospital beds, oxygen equipment, and blood glucose monitors. Within traditional Medicare, Part B provides coverage for some durable medical equipment.
Fee-for-service—A payment model in which healthcare providers receive payment for each service, test, or procedure rendered to a patient. These fees may be administratively set or determined through negotiations between healthcare providers and payers (ie, insurance companies or government administered health insurance programmes). Traditional Medicare, which applies to government administered health insurance programmes including Medicare Part A and Part B, operates largely through a fee-for-service system.
Medicare Advantage (Medicare Part C)—A health insurance programme in which the US government pays fixed amounts for each enrolled person per period of time to private health insurance firms to pay for and administer medical and drug benefits to Medicare beneficiaries. Medicare Advantage plans include coverage of Medicare Part A and Part B services, and many plans also include prescription drug coverage (otherwise available through Medicare Part D). Beneficiaries have the option of enrolling in Medicare Advantage or traditional Medicare.
Medicare Part A—One of the two main components of the traditional Medicare health insurance programme in the US. Part A provides coverage for inpatient care in a hospital, skilled nursing facility care, nursing home care, hospice care, and some home health care.
Medicare Part B—One of the two main components of the traditional Medicare health insurance programme in the US. Part B provides coverage for outpatient medical services, preventive services, and some other medical services. Examples of covered services include doctor visits, laboratory tests, diagnostic screenings, durable medical equipment, physician administered medications, and ambulance services.
Medicare Part C—Medicare Advantage.
Medicare Part D—A voluntary outpatient prescription drug coverage programme (non-physician administered medications) in the US for people with Medicare and offered through private health insurance plans that are approved by and contracted with the federal government. Beneficiaries choose between standalone prescription drug plans to supplement traditional Medicare or a Medicare Advantage plan that includes prescription drug coverage.
Physician administered medications—Medications administered by a healthcare provider in an outpatient setting. These medications are typically injected or infused. Within traditional Medicare, Part B provides coverage for physician administered medications.
Physician services—A term that encompasses medical services provided by healthcare professionals. Within traditional Medicare, Part B provides coverage for many physician services, including, for example, doctor visits, surgical procedures performed in an outpatient setting, diagnostic tests, preventive services, and durable medical equipment.
Prior authorization—A process by which providers must obtain coverage determination from insurers for specific medical services before they are provided to patients. Through this process, insurers assess the necessity of medical services before they are provided.
Traditional Medicare—The original government administered health insurance programme in the US, which primarily serves elderly people. It consists of two parts: Medicare Part A (hospital insurance) and Part B (medical insurance). Traditional Medicare primarily operates as a fee-for-service system.
To address the challenge of constraining healthcare spending, private insurers may employ different tools from those used by governments. One controversial tool that health insurers use for managing care in the US is prior authorization (PA), a process by which insurers assess the necessity of medical services before they are provided. As with other utilization management techniques used by insurers, PA could serve to curb wasteful spending if it discourages care that is of low value; such low value care seems to be common.2 Two features distinguish PA from other tools that insurers use to reduce spending, such as patient cost sharing or denials of submitted claims. First, PA is prospective, requiring the determination of coverage before a patient receives costly services. Second, PA targets particular treatments or diagnostic services, especially high priced medications, or procedures. Although PA has been used since the 1980s, the practice has recently prompted controversy. Provider groups, scholars, and policy makers have expressed concerns that PA is administratively burdensome345 and can discourage appropriate care.6789 In the past year, several proposals to regulate PA have been advanced in Congress, state governments, and federal rulemaking.1011121314
Despite widespread policy interest in PA, research on the extent and effects of PA is lacking. PA requests and services deterred by PA are rarely identifiable in standard claims datasets for tracking healthcare use. Existing studies provide an incomplete understanding of the scope of PA policies. A recent analysis of PA requests to private health insurers in the US, for example, did not account for the services that were deterred by PA.15 Accounting for deterrence is important, as PA requirements may result in fewer requests and therefore lower denial rates over time.16 Thus, it remains unclear how often physician services require PA and to what degree PA policies vary across insurers. Answers to these questions could inform ongoing policy debates. Substantial variation in PA policies would imply disagreement among insurers about which services are susceptible to overuse. Services uniformly subject to PA, however, would reflect a consensus among insurers that these services may be susceptible to overuse, thereby implying that these services could be potential targets for interventions that aim to reduce low value care. Furthermore, if PA requirements were found to be extensive, this finding could inform ongoing policy reform to focus PA on low value services and to avoid administrative effort where the costs exceed benefits.
Understanding PA is particularly important for Medicare policy because the use of PA differs between government and privately administered insurance in the US. Several countries offer public health insurance with options to purchase substitutive or supplemental private coverage. The US is unique in offering publicly subsidized options to enroll in traditional Medicare, the government administered insurance that covers hospital care (Part A) and physician services (Part B), and the privately administered alternative to traditional Medicare, called Medicare Part C or Medicare Advantage. PA is minimal in traditional Medicare,17 which imposes coverage restrictions primarily through retrospective denials of claims. In contrast, private insurers can require PA in both Medicare Advantage and Medicare Part D prescription drug plans.18 Unlike PA policies for prescription drugs, however, the PA policies for physician services have not been well characterized across insurers.1920 In this study, we measured and compared the scope of PA policies for physician services, including physician administered medications, among five health insurers that collectively serve most of the Medicare Advantage market, quantifying differences in PA across insurers, physician specialties, and clinical service categories.
Methods
Study design and overview
We measured the scope of PA policies of Medicare Advantage insurers by determining the number of services delivered in traditional Medicare Part B that would have required PA by each insurer, and then calculated the associated spending. This study design allowed us to detect services that would have been subject to PA regardless of whether PA would have been approved, denied, or never sought owing to the sentinel effect of PA policy (ie, deterrence of services without denial of PA).21
Data and study sample
Our initial sample included PA policies among health insurers with Medicare Advantage market shares of 2% or greater, with two exceptions. We excluded Kaiser Permanente because of its integrated insurer-provider organizational model, and we excluded BlueCross and BlueShield affiliates because PA policies were available on independent websites, and these were too numerous to feasibly obtain. The included firms and their 2021 market shares were UnitedHealthcare (27%), Humana (18%), CVS Health (11%), Centene/Wellcare (4%), and Cigna (2%).22
We identified all unique Healthcare Common Procedure Coding System (HCPCS) codes in traditional Medicare claims using the 2021 Centers for Medicare and Medicaid Services National Physician Fee Schedule Relative Value Files.23 For each code, we assessed whether the service required PA by each Medicare Advantage insurer in 2021. We extracted data on PA policies from public insurer documents or websites published between July and September 2021.2425262728 Each insurer publishes regularly updated PA policies differently for its Medicare Advantage and commercial insurance plans. We manually extracted HCPCS codes subject to PA from available policy documents containing coverage rules for Medicare Advantage plans. If instead the policies were available on insurer websites through search applets, we extracted HCPCS codes subject to PA using an algorithm written in the Selenium package of Python (see supplement file for more details). One insurer, Wellcare, made its PA policies available on a consolidated website across multiple lines of business.
We obtained Part B HCPCS level service counts, or utilization, and allowed charges (ie, spending) from the Medicare Part B National Summary Data File.29 Our primary analyses employed 2021 data, the most recent available at the time of our study; in a sensitivity analysis, we used 2019 spending and utilization data to confirm that our findings were not driven by changes in utilization associated with the covid-19 pandemic. We used Medicare administrative enrollment files to calculate spending and utilization for each enrolled beneficiary.30
Analysis of insurer prior authorization policies
To compare the scope of PA policies, for each insurer we totaled Part B spending and utilization counts of HCPCS codes that would have required PA. We refer to these quantities as PA spending and PA utilization, respectively. We also calculated PA spending and PA utilization for three hypothetical policies based on the degree of insurer consensus for each service: the narrowest hypothetical policy required PA for services at all insurers, the broadest hypothetical policy required PA for services by at least one insurer, and an intermediate policy required PA for services requiring PA by any of the insurers with the three largest market shares.
To examine the types of services subject to PA, we assessed the extent of PA across and within clinical service categories. We assigned each HCPCS code to one of five clinical groups: durable medical equipment and transport; radiology, pathology, and laboratory; other medical services; surgery and anesthesia; and medications. We further subcategorized medications into nine therapeutic areas using the Cerner Multum Lexicon database31: autoimmune and musculoskeletal; cardiovascular, diabetes, and hyperlipidemia; genitourinary and sex hormones; hematology and oncology; infectious diseases; neurology; psychiatry; pulmonology; and other.
For each of the five actual and the three hypothetical insurer PA policies and each service category we calculated several quantities. To assess the scope of PA, we calculated the ratio of PA utilization in the service category to all service utilization in the category, and the ratio of PA spending in the category to all spending in the category (eg, the proportion of traditional Medicare Part B medication spending that would require PA). To assess the distribution of PA across service categories, we calculated the ratio of PA utilization in the category to all PA utilization, and the ratio of PA spending in the category to all PA spending (eg, the proportion of all PA spending that was for Part B medications). We summarized these outcomes across insurers by calculating medians and means weighted by insurers’ market share.
To quantify differences in PA by physician specialties, we used the publicly available 2020 Medicare Physician and Other Practitioners-by Provider and Service dataset. We extracted information on services and procedures provided to traditional Medicare Part B beneficiaries by physician and other healthcare professionals, aggregated by provider (based on national provider identifier) and service (based on HCPCS codes).32 We combined specialties as appropriate (see supplement file for more details), excluding services attributed to institutions, non-clinicians, or pediatrics (since Medicare covers almost no children). For each physician specialty, we calculated the proportion of total specialty spending for each HCPCS code. We then calculated the total proportion of Part B spending for which each insurer would have required PA for each physician specialty. Finally, we determined the most expensive and most common individual services subject to PA according to annual spending and utilization.
Data analysis employed Python software, version 3.8 (Python Software Foundation) and Stata version 16 (StataCorp).
Patient and public involvement
No members of the public were formally involved in the design or implementation of this study because no funding was set aside for public or patient involvement. However, clinician authors’ experiences caring for patients whose healthcare requires navigating prior authorization policies informed the motivation for and development of the research question and interpretation of the results.
Results
Among the 30 540 086 beneficiaries in traditional Medicare Part B in 2021, mean annual spending was $4649 (£3698; €4338) per beneficiary, or $142bn in total (see supplement table 2). Of the 14 130 Part B clinical services, Medicare Advantage insurers required PA for 944 to 2971 services (median 1899; weighted mean 1429); 4044 services (29%) required PA by at least one Medicare Advantage insurer and 239 (2%) required PA by all Medicare Advantage insurers.
Across the five Medicare Advantage insurers, the proportion of Part B spending requiring PA ranged from 17% to 33% (median 28%; weighted mean 23%) (fig 1); the proportion of Part B utilization requiring PA ranged from 9% to 41% (median 22%; weighted mean 18%; see supplement figure 1). Forty per cent of spending ($57bn) and 48% of service utilization would have required PA by at least one Medicare Advantage insurer. In a sensitivity analysis employing 2019 spending and utilization data, the proportion of spending on PA services did not substantially change (see supplement figure 2). Twelve per cent of spending and 6% of utilization would have required PA by every Medicare Advantage insurer.
Share of spending requiring prior authorization, by clinical service and insurer. Each column presents the 2021 prior authorization spending in each clinical service category at each Medicare Advantage insurer, all insurers, and at least one insurer. Each value is a share of total 2021 Medicare Part B spending. A (29%), B (11%), and D (18%) are large insurers (Medicare Advantage market share), whereas C (2%) and E (4%) are small insurers by Medicare Advantage market share. See supplement table 3 for numerical values represented in the figure
For every Medicare Advantage insurer, physician administered medications represented the plurality of PA spending (range 46-62%; median 57%; weighted mean 57%; table 1). Medicare Advantage insurers required PA for 56% to 89% of medication spending (median 86%; weighted mean 72%; see supplement table 5), accounting for 14% to 67% of medication utilization (median 35%; weighted mean 28%; see supplement table 5). Ninety three per cent of Part B medication spending, or 74% of medication use, would require PA by at least one Medicare Advantage insurer. Fifty one per cent of Part B medication spending, or 10% of medication use, would require PA by every Medicare Advantage insurer. Supplement figure 3 and supplement tables 9 and 10 present corresponding calculations for intermediate degrees of consensus about PA policies among insurers. Surgery and anesthesia services accounted for the second greatest proportion of PA spending (range 20-27%; median 21%; weighted average 24%; table 1).
Distribution of prior authorization spending and utilization, by insurer, clinical service, and medication therapeutic area
Figure 2 shows the distribution of PA spending across medication therapeutic areas. For all Medicare Advantage insurers, hematology and oncology drugs represented the largest proportion of PA spending (range 27-34%; median 33%; weighted mean 30%). The proportion of spending subject to PA ranged widely by physician specialty (fig 3).
Share of medication spending requiring prior authorization, by therapeutic area and insurer. Each column presents the 2021 prior authorization spending in each medication therapeutic area category at each Medicare Advantage insurer, all insurers, and at least one insurer. Each value is a share of total 2021 Medicare Part B medication spending. A (29%), B (11%), and D (18%) are large insurers (Medicare Advantage market share), whereas C (2%) and E (4%) are small insurers by Medicare Advantage market share. Per cent spending on psychiatry and infectious diseases drugs is too small to be detected visually. See supplement table 3 for numeric values represented in the figure
Share of spending for clinical services requiring prior authorization, by physician specialty and insurer. Each datapoint represents the proportion of a specialty’s Medicare Part B spending that would require PA according to each Medicare Advantage insurer. A (29%), B (11%), and D (18%) are large insurers (Medicare Advantage market share), whereas C (2%) and E (4%) are small insurers by Medicare Advantage market share. See supplement file for details on specialty groupings and spending calculations
Table 2 presents individual PA services with the largest associated spending, mostly injectable medications. Among services requiring PA by at least one Medicare Advantage insurer, the 10 costliest accounted for 10% of all Part B spending. Among services uniformly requiring PA by all Medicare Advantage insurers, the 10 costliest accounted for 7% of all Part B spending.
Services requiring prior authorization with the greatest spending*
Discussion
Debate in the US about the proper role of PA by private health insurers has occurred amid considerable uncertainty around its extent and variation across insurers, physician specialties, and clinical service categories. In this cross sectional analysis of PA policies across five Medicare Advantage insurers, we found that nearly half of traditional Medicare Part B spending and utilization would have been subject to PA by at least one Medicare Advantage insurer. The Medicare Advantage insurer with the broadest PA policy would encompass one in three dollars spent in Medicare Part B; even the Medicare Advantage insurer with the narrowest policy would encompass one in six dollars. However, the medical services uniformly subject to PA by all five Medicare Advantage insurers accounted for a substantially smaller proportion of spending.
Our study advances efforts to understand the role of private insurers in setting coverage policy and insurers’ use of PA in several ways. We extended methods previously applied to a single Medicare Advantage insurer,21 studying several insurers that collectively served more than 60% of the Medicare Advantage market, allowing a more generalizable and nuanced description of PA and its variation. Instead of using data from clinician surveys33 or PA applications15 to quantify the frequency of encountering PA, we applied coverage policies to the universe of medical services delivered in traditional Medicare Part B. This allowed for consistent and objective quantification of the scope of PA in a national and policy relevant sample. Finally, we studied PA in the setting of clinician services rather than Part D medications, which have been more thoroughly studied.1920 Our approach yielded three key findings relevant to the debate over the proper role of PA.
First, we found widespread use of PA for medical services in Medicare Advantage, which is in marked contrast to traditional Medicare. These findings add to previous literature suggesting that Medicare Advantage reduces healthcare use relative to traditional Medicare.343536 The results are also consistent with previous studies showing a high prevalence of medical services suspected of being low value in fee-for-service Medicare.23738 Furthermore, this finding suggests that insurers’ financial savings from PA programmes exceed insurers’ costs to administer them. However, whether PA of medical services results in substantial aggregate savings remains unclear because our analysis did not account for several costs: the administrative and compliance costs of PA for clinicians, insurers, patients, and regulators; the reduction in spending on restricted services; the increase in spending on services that substitute for restricted services; and costs associated with any adverse events as a result of, or prevented by, PA.
Nonetheless, the large contrast in PA between Medicare Advantage and traditional Medicare suggests widely differing approaches to coverage policy. These divergent approaches to constrain health spending are unlikely to both be optimal, highlighting challenges faced by health systems such as those in the US, where national governments aim to incorporate private firms in the provision of health insurance as well as fully public systems elsewhere. Further research should quantify the relative benefits and harms of the differences in use of PA between Medicare Advantage and traditional Medicare. One recent analysis found that extensive PA policies for prescription drugs produced benefits exceeding their cost.20 It is, however, unclear whether extensive PA is similarly valuable for the (non-prescription) medical services we studied. A recent report from the Office of Inspector General found that in some cases Medicare Advantage insurers denied medical services after a PA request that met Medicare coverage criteria.8 Though Medicare Advantage insurers are required to follow government coverage determinations, many medical services are implicitly covered by Medicare without undergoing a formal coverage determination process.39 Thus, some disagreements about coverage may result from a lack of sufficient criteria to determine when it is appropriate for Medicare Advantage insurers to apply additional coverage restrictions. Beyond Medicare policy, this result highlights the importance of policy efforts to reform PA, such as increasing transparency, streamlining PA processes,40 and “gold-carding” rules that allow certain providers to be exempt from PA requirements based on performance measures.41 The considerable scope of services requiring PA may lead to reductions in the provision of low value services, but it can impose an administrative burden on the clinicians who must navigate these policies, and the insurer staff who adjudicate the claims. Patients may also face barriers or delays in obtaining appropriate care,424344 particularly if therapeutic substitutes are lacking. Because most PA requests are eventually approved, reforms may be particularly promising if they reduce administrative burdens for services most likely to be approved45; to infer whether a PA request is likely to be approved, insurers might employ readily available data about a patient or clinician (eg, the patient’s previous diagnoses or the physicians’ rate of PA approvals).
Second, we found noticeable variation in the use of PA within Medicare Advantage. Only 30% of the Medicare spending subject to PA at any Medicare Advantage insurer was subject to PA at every Medicare Advantage insurer; most of this spending was devoted to a small number of costly services. These findings highlight the challenge of uniformly identifying such low value services, which lack a consistent definition or legal standard. Other explanations for variation in PA across insurers may be variation in practice patterns across the different markets that the insurers serve; differences in bargaining power between insurers and providers across these markets, which could affect negotiations over the use of PA; differences in the use of alternative utilization management or payment policies employed by insurers, such as risk based contracts in which providers are responsible for cost and quality of care provided to patients, which may also impact an insurer’s PA policies; and differences in the degree to which insurers enforce their PA rules. In the present study, the services that consistently required PA across insurers could guide PA policy reforms as a focus for coverage policy: PA in private insurance plans could be narrowed to focus on these services, whereas PA in government administered Medicare could be expanded to include these services. Future research should examine whether the services uniformly subject to PA across most insurers are indeed more likely to represent low value care.
In addition, differing PA policies across insurers might disrupt care for patients switching between health plans with different coverage policies for ongoing diagnostic or treatment plans. Previous studies examining coverage decisions also found substantial variation among private insurers,46 but not necessarily variation by clinical factors.47 Other studies have found variation in the evidence used by commercial health plans in their coverage policies for specialty drugs, which occasionally diverge from evidence included in FDA required labeling.4849 Though these latter studies did not examine PA specifically, variation in the evidence reviewed to determine PA requirements may explain some of the differences we observed among Medicare Advantage insurers. It is possible that dissimilarity of PA policies has also been driven by differences in the use of particular medical services by each insurer’s mix of patients and in-network clinicians.
Third, we found that most spending on medical services requiring PA was concentrated in physician administered drugs, which are typically injected or infused. Nine in 10 dollars spent on medications would be subject to PA by at least one Medicare Advantage insurer. These findings are consistent with a recent study documenting the prevalence of PA policies among large Medicare Advantage insurers for the top 20 drugs by Part B spending in 2020.50 Use of PA for such specialty medications may result from their high cost and unit prices, delayed competition and market entry of biosimilars, and opaque interactions between pharmacy benefit managers, insurers, and drug manufacturers incentivizing preferred formulary status and rebates for expensive drugs.51 Moreover, assigning appropriateness criteria to non-drug services may be more challenging,52 which may explain why the costliest services uniformly subject to PA were mostly medications.
Strengths and limitations of this study
Our study has several limitations. First, we examined the scope of PA only for physician services and did not address other services such as inpatient hospital admission, post-acute care, or Part D prescription drugs. Our research design would not have been feasible for studying inpatient hospital admissions because PA policies typically only apply to elective hospital admissions, which are difficult to identify in administrative claims. Similarly, it was not feasible to examine PA for prescription drugs in Part D because PA is widespread in Part D53 and our research design relied on healthcare utilization data in a setting that lacked PA.54 Second, our estimates were based on utilization patterns of traditional Medicare beneficiaries, which may differ from those of Medicare Advantage beneficiaries; again, this limitation resulted from our need to employ utilization data from a setting lacking PA. Third, our analysis uses traditional Medicare prices, which can differ slightly from Medicare Advantage. Fourth, we compared the PA policies of five Medicare Advantage insurers, which may differ from those of other insurers. Fifth, we were unable to include data from BlueCross and BlueShield affiliates in our analysis, as many of the affiliates’ PA policies were not available for collection on a centralized website. Nonetheless, the included insurers represent most Medicare Advantage enrollees.22 Sixth, we utilized data on insurers’ stated policies on services requiring PA. An evaluation of implementation of such policies, including how frequently insurers initially denied and subsequently approved specific services, was not feasible.
Conclusion
In this cross sectional study, the coverage policies of private insurers in the US would have required PA for a large portion of fee-for-service Medicare Part B spending, particularly spending on medications. There was little consensus among insurers, however, about which medical services required PA. This lack of uniformity suggests it can be challenging to identify and discourage some low value services. Our study informs ongoing efforts to reform PA and reduce its administrative burdens. Given the continued growth of privately administered Medicare insurance plans, PA is likely to remain an important feature of the Medicare programme. This aspect of managed care may be improved if PA policies can be narrowly targeted at the specific services most likely to represent low value care.
What is already known on this topic
In the US, health insurers use a variety of tools to limit the quantity of services provided, including prior authorization (PA), a process by which insurers assess the necessity of planned medical care; however, associated administrative burdens have prompted calls for policy reform
Although minimal PA exists in government administered traditional Medicare, the insurance programme that covers hospital admissions (Part A) and physician services (Part B), PA is more extensive in the privately administered alternative, Medicare Advantage
Evidence is lacking on the extent of PA and its variation across insurers, physician specialties, and clinical services
What this study adds
The coverage policies of private insurers would have required PA for a large portion of fee-for-service Medicare Part B spending, particularly spending on physician administered medications
The lack of consensus among insurers suggests it can be challenging to discourage some low value services; reforms could focus PA on services for which there is the greatest consensus about the risk of overuse
A substantial difference in coverage policies exists between government administered and privately administered Medicare, highlighting the importance of policies governing the role of private health insurers with respect to PA in the US
Ethics statements
Ethical approval
This study used public, non-identifiable data and was exempt from institutional review board review.
Data availability statement
Additional data are available at https://github.com/ravigupta-github.
Acknowledgments
We thank Katherine Tanis and Sara Riaz for their research assistance.
Footnotes
Contributors: RG and AS contributed to the design of the study. RG, JF, and AS directed the analyses, which were conducted by RG and JF. All authors contributed to interpreting the results. RG and AS drafted the manuscript. All authors critically revised the manuscript for important intellectual content. RG is the guarantor. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.
Funding: This work was supported by grant IK2HX003634 from the US Department of Veterans Affairs (Office of Research and Development Career Development Award). The funders had no role in considering the study design or in the collection, analysis, interpretation of data, writing of the report, or decision to submit the article for publication.
Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare: support from the US Department of Veterans Affairs for the submitted work; RG reports receiving consulting fees from Alosa Health and Sunday Health for unrelated work; JF is employed by Medidata Solutions, which played no role in the development or publication of this paper; ALS reports receiving speaking fees from VBID Health and is an employee of the Veterans Health Administration; no other financial relationships with any organizations that might have an interest in the submitted work in the previous three years. The statements, findings, conclusions, and views expressed in this article are those of the authors and do not necessarily reflect those of the US Department of Veteran Affairs or the US government.
The lead author affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.
Dissemination to participants and related patient communities: We plan to disseminate our work publicly through a variety of venues, including lay press coverage, social media, press releases and blog posts by the authors’ affiliated organizations, and presentations at virtual and in-person conferences.
Provenance and peer review: Not commissioned; externally peer reviewed.
This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.