Potential impacts on costs and premiums related to the elimination of prior authorization requirements in Massachusetts

17335 Golf Parkway

Suite 100

Brookfield, WI 53045

USA

Tel +1 262 784 2250

milliman.com

MILLIMAN REPORT

Potential Impacts on Costs and Premiums Related to

the Elimination of Prior Authorization Requirements in

Massachusetts

Commissioned by Massachusetts Association of Health Plans

November 29, 2023

Fritz S. Busch, FSA, MAAA Peter Fielek, FSA, MAAA

Principal and Consulting Actuary Consulting Actuary

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Table of Contents

I. EXECUTIVE SUMMARY ........................................................................................................................................ 1

II. BACKGROUND ...................................................................................................................................................... 4

III. MEDICAL AND DRUG ALLOWED COST RESULTS ............................................................................................ 6

IV. RANGE OF IMPACTS TO PREMIUMS AND MEMBER COST SHARING ............................................................ 9

V. CONSIDERATIONS ............................................................................................................................................. 16

VI. METHODOLOGY ................................................................................................................................................. 18

VII. CAVEATS ............................................................................................................................................................. 20

APPENDIX A

APPENDIX B

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 1

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

I. EXECUTIVE SUMMARY

Purchasers of health insurance, including individuals and families, employers, and government agencies, expect

payers

to manage the cost and quality of prescription drug and medical services delivered. To meet this expectation

payers, monitor and manage utilization of services, treatments, and medications to ensure that health care resources

are used appropriately, and evidence-based care is delivered at the right place, at the right time, and in the most

appropriate setting.

Prior-authorization (PA) is one tool employed by commercial payers and managed care organizations or MCOs

(hereafter collectively referred to as “plans”) in the managed care industry to optimize the utilization of various

high-dollar or potentially low-value services that are covered under typical comprehensive medical coverage. Services

that could be over-utilized or misused, might not meet medical necessity criteria, are considered unsafe, or have

alternative treatments available, are likely candidates for PA. Other services that are candidates for PA are those that

are rare, but extremely expensive or where clinical safety is less certain.

Prior authorization can be an effective cost control tool depending on the scope of services subject to PA, as well as

the stringency of the application of the authorization rules.

It can also help members avoid unnecessary risks

and / or paying additional out-of-pocket cost sharing and improve health outcomes by ensuring that members are

treated in accordance with the latest clinical guidelines and evidence-based medicine. Likewise, PA can improve

community practice by ensuring that providers are meeting nationally and locally accepted evidence-based standards

of medical practice.

However, PA can add to the administrative burden for healthcare providers and obtaining PA

approval is often cited by providers as the primary reason for delays in care or the patient not seeking the treatment at

all,

which can lead to increased long-term costs.

Given that PA programs impact clinical care and can affect member and provider satisfaction, premiums, and member

cost sharing, it is important that stakeholders understand the tradeoffs involved with limiting or eliminating PA, including

the potential increase in healthcare costs associated with placing limitations on or even eliminating the use of PA. In

the commercial market, increased costs impact insurance premiums and member cost sharing, while increased costs

associated with limiting or eliminating the use of PA in Medicaid can impact a state’s fiscal budget.

The Massachusetts Association of Health Plans (MAHP) engaged Milliman to model the potential cost impacts that can

result from generally

limiting or eliminating PA on the commercial and Medicaid markets in Massachusetts. The

commercial market includes both self-insured and fully insured employer groups, the individual and small business

ACA market, and the state’s Group Insurance Commission which administers health insurance to the Commonwealth’s

employees, retirees, and dependents. For Medicaid, we considered enrollees under managed care plans (enrolled with

an MCO) only and not fee-for-service enrollees.

Employers, who typically pay a large share of premiums under employer-sponsored health insurance programs, and

some consumers in the individual market could see premium increases if the use of PA is either limited or eliminated.

Likewise, commercially insured patients, who typically pay cost sharing (deductibles, copays, etc.), under their health

plan, could see increases in their out-of-pocket costs as they would be paying cost sharing on, for example, potentially

medically unnecessary services or on services with lower cost alternatives or more appropriate site of care. While cost

sharing is limited in Medicaid, state Medicaid programs could see higher expenditures or capitation rates if PA criteria

is limited or eliminated.

For this paper, we directly model the elimination of PA. Limitations of PA short of complete elimination could have

impacts ranging from minimal to estimates closer to the costs of full elimination provided in this report. To understand

the potential range of cost impacts resulting from eliminating PA in the commercial market, we analyze 2022 claims

experience (trended to 2023 cost levels) including the member cost sharing portion, for a Massachusetts sample of

commercial enrollees that purchased or were provided a typical major medical policy in 2022.

For Medicaid, we relied

on 2020 Massachusetts Medicaid data from the Transformed Medicaid Statistical Information System (T-MSIS) monthly

Third-party claims payers, such as commercial plans, Medicaid managed care organizations, Medicare advantage plans, federal and state

governments. This paper focuses on commercial payers and MCOs.

https://www.nihcr.org/wp-content/uploads/Altarum-Prior-Authorization-Review-November-2019.pdf

https://www.gao.gov/assets/700/692138.pdf

Massachusetts law requires that medical necessity guidelines be evidence-based. See General Law - Part I, Title XXII, Chapter 176O, Section 16

(malegislature.gov)

https://www.hsgac.senate.gov/subcommittees/investigations/hearings/examining-health-care-denials-and-delays-in-medicare-advantage/

https://www.ama-assn.org/system/files/2021-04/prior-authorization-survey.pdf

In Massachusetts, Medicaid is ~30% of the state budget. https://budget.digital.mass.gov/govbudget/fy24/budbriefpdf/fy24_bb13_masshealth.pdf

While there are ongoing discussions and proposals regarding prior authorization requirements, we do not model any specific proposal or legislation.

Typical is defined as coverage that includes medical and pharmacy services and generally includes the essential health benefits (EHBs) as defined by

the ACA. More information on EHBs can be found here: https://www.cms.gov/CCIIO/Resources/Data-Resources/ehb

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 2

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

claims files, a 100% dataset from CMS. It should be noted, that for the claims period analyzed, health plans in

Massachusetts relaxed PA in some circumstances due to the COVID-19 pandemic.

Using PA code lists from plans in Massachusetts, we construct code lists with scopes characterized as “broad”

(relatively more codes subject to PA review) and “narrow” (relatively fewer codes subject to PA review) to identify

medical procedures and drugs that are subject to PA restrictions. We then apply a measure of program effectiveness,

which is defined as the net percentage reduction in medical and prescription drug costs attributable to the PA program,

to each of the scopes of services. Estimates of net effectiveness were based on data and information provided by

Massachusetts plans. To calculate program effectiveness, we use a range of service denial rates provided by plans

that provide commercial insurance, as well as managed Medicaid in Massachusetts.

Finally, using a literature review, we estimate the “sentinel effect”

of eliminating PA. The “sentinel effect” is the

increase in medical and prescription drug expense after PA is removed that is related to requests for authorizations

that were previously unsubmitted when PA was in place because providers did not expect an approval. It is expected

that these authorizations will be submitted once PA requirements are removed.

A summary of results can be found in Figures 1 and 2 below.

Figure 1: Premium Impacts (PMPM) of PA Removal

Broad and Narrow

Commercial Market

% of Premium Impact

Broad PA List

Narrow PA List

Low

Sentinel

Med

Sentinel

High

Sentinel

Low

Sentinel

Med

Sentinel

High

Sentinel

Effectiveness

6.6%

5.1%

Sentinel

5.6%

11.2%

16.7%

4.1%

8.1%

12.2%

Total Impact

12.1%

17.7%

23.3%

9.1%

13.2%

17.3%

$ PMPM Impact

Effectiveness

$36.66

$28.48

Sentinel

$31.21

$62.42

$93.63

$22.71

$45.42

$68.13

Total Impact

$67.86

$99.07

$130.28

$51.19

$73.90

$96.61

Figure 2: Capitation Rate Impacts (PMPM) of PA Removal

Broad and Narrow

Medicaid Managed Care

% of Capitation Impact

Broad PA List

Narrow PA List

Low

Sentinel

Med

Sentinel

High

Sentinel

Low

Sentinel

Med

Sentinel

High

Sentinel

Effectiveness

4.1%

2.1%

Sentinel

3.8%

7.6%

11.3%

1.7%

3.5%

5.2%

Total Impact

7.9%

11.7%

15.5%

3.9%

5.6%

7.3%

$ PMPM Impact

Effectiveness

$24.14

$12.27

Sentinel

$22.06

$44.12

$66.18

$10.20

$20.39

$30.59

Total Impact

$46.20

$68.27

$90.33

$22.46

$32.66

$42.86

The sentinel effect is additive because the impact of removing the effect of PA only accounts for this effect at a lower

level of claims submission when PA was in effect. The sentinel effect accounts for the additional claims not contained

in the calculation of effectiveness.

Denial rates are a significant, but not the sole, driver of PA effectiveness. See considerations section in the main body of this report for more discussion

related to factors influencing program effectiveness.

Sentinel affect in general is the modification in behavior that occurs when subjects are aware they are being observed and measured.

Please see the full report and Methodology section for sources, full development, and calculation methodology of these results.

All results are applicable to the population modeled, which is broadly representative of a standard population with average age of 45 and covered by a

typical large group major medical plan with an average AV of 85%. Results may differ for specific employers, individual purchasers, geographies or for

younger or older populations.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 3

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Across the Massachusetts commercial market, we estimate a total annual impact to premiums of between $2.2-$5.6B

and across the Medicaid market $0.4-$1.6B

In addition, readers should also consider the following:

▪ Administrative Costs – While premium costs increase if limitations are placed on PAs, our research indicates

that a plan’s administrative costs related to PA would likely be lower, depending on the nature of the limitation.

In the special cases where a PA is eliminated, it may allow a reduction in the staffing hours previously used

to support that authorization. Since eliminating PA does not eliminate health plan requirements to manage

cost and quality, administrative resources most likely would be redirected to retrospective audits. There may

also be cost increases in other areas that offset these reductions. For example, it is likely plans will increase

fraud, waste and abuse controls, monitoring of provider billing patterns, and / or payment integrity efforts.

Plans may also have increased claims volume. Our analysis did not factor in these potential administrative

cost changes. Therefore, the increases to premiums and capitation rates could be higher or lower than shown.

▪ Net Savings – Often PA results in a diversion of care from one course of treatment to another rather than a

wholesale denial of a course of treatment. While the PA-approved treatment, if different from the original

course of treatment, may still incur costs, there is often a net savings relative to the original course of care.

When such PA programs are removed, the additional costs include the loss of these net savings rather than

the full cost of treatment denied under the PA program. It is outside of the scope of this analysis to quantity in

a detailed manner the impact of diversions of care as opposed to full denials for every service. Instead, we

attempt to account for the average net savings by service category. We believe the general conservatism in

our assumptions and methodology elsewhere is sufficient to arrive at a reasonable estimate of net savings.

▪ Value-Based Care (VBC) – Providers have different incentives under value-based payment (VBP) models

than they would under a predominantly fee-for-service environment. Thus, any limitations placed on PA could

have smaller effects on premiums, capitation rates, and cost sharing than shown in this analysis, particularly

in delivery systems that are already efficient. That being said, the authors’ experience is that the realignment

of incentives in commercial and Medicaid VBP models has yet to result in a transformational change in

behavior for providers. To date, a large percentage of commercial and Medicaid managed care members will

not be covered in VBP models,

nor are the incentives sufficient to transform their delivery model.

The ultimate effect of eliminating PA on medical services and prescription drug utilization, which drives commercial

premiums and Medicaid capitation rates, can be influenced by these considerations and others discussed in the body

of this analysis.

The analysis in this report represents the opinion of the authors only.

This report is intended to inform stakeholders, which could include but is not limited to, employers, employees and

other individual consumers, commercial plans, MassHealth, providers, lawmakers, and other policy makers about the

range of potential financial impacts (i.e., changes in premiums, capitation rates under a managed Medicaid program

and consumer cost sharing) of various changes to State laws intended to limit the plan practice of PA as a method of

controlling costs. This report is technical in nature and readers with limited background in healthcare should consult a

qualified professional when interpreting these results.

Enrollment assumptions for total impact are from Enrollment in Health Insurance (chiamass.gov)

https://hcp-lan.org/apm-measurement-effort/2020-2021-apm/2021-infographic/

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 4

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

II. BACKGROUND

Managed care plans use various processes known collectively as utilization management (UM) to improve patient

outcomes, ensure efficacy and safety of medications, treatments, services, and procedures, and optimize the utilization

of various high cost or potentially low-value services that are covered under typical major medical coverage. Likewise,

prior authorization (PA) can increase the amount of care that the medical community adopts that is consistent with

nationally and local-recognized evidence-based standards. UM usually encompasses prospective service review (i.e.,

prior to services being rendered), concurrent review (generally applicable to just inpatient hospital stays and occurring

during the actual service period) and retrospective review (after claim is paid or service is rendered). In comparison to

concurrent and retrospective review, however, prospective review can prevent unnecessary or harmful care from being

delivered at the outset. Our analysis focuses on this front-end, prospective component of UM, which we refer to as prior

authorization (PA). Services that are or could be over-utilized and misused, or services that do not meet medical

necessity criteria, may raise safety concerns, or have alternative treatments available, are likely candidates for PA.

Typically, PA is characterized by requiring specific approval before a service is rendered or a drug is prescribed and

there is a more involved administrative process of notification, documentation, and justification on the part of the

ordering provider before a service is approved to be paid by the plan. Notification, documentation and approval or

denial can be done telephonically or electronically through secure channels. Approval for a particular service can also

be granted retrospectively but typically on exception basis. It is important to note, that the approval being requested is

for plan payment to the provider. The approval or disapproval for payment of a particular service does not mean that

the service cannot be provided.

Prior authorization can be an effective cost and quality control tool for plans depending on the services subject to PA,

as well as the strictness of the application of the authorization rules. Controlling medical and prescription drug costs is

an efficient way for plans to keep health insurance premium rates affordable for individuals, employers, and state

Medicaid programs, as medical and prescription drug costs are the largest portion of costs underlying a typical major

medical insurance policy. By identifying inappropriate, unnecessary, or ineffective services, PA can reduce financial

and medical risk for plans and employers and help commercially insured individuals avoid additional out-of-pocket cost

sharing on those services.

In addition to reducing costs, PA can be valuable in reducing variation in health care delivery through the use of

evidence-based standards of determination. Stakeholders including provider groups, hospitals, state and federal

agencies, professional societies and patient advocacy groups all have an interest in reducing non-evidence-based

medicine as it can affect member outcomes adversely.

On the other hand, PA can add to the administrative burden for healthcare providers and is cited by providers as the

primary reason for delays in care or for the patient not seeking the treatment.

In severe cases, (despite expedited

review and approval processes being available), delayed or foregone treatment could ultimately increase costs if a

member does not receive care until the condition worsens. Provider groups, such as the American Medical Association

have criticized PA and multiple states, including Massachusetts, have proposed legislation seeking to limit PA practices

of plans.

In the 2023-2024 legislative session, 37 bills have been filed by specialty societies, such as the

Massachusetts Medical Society, Massachusetts Radiological Society, and the Massachusetts Psychiatric Association

seeking to eliminate PA for specific services and treatments, as well as an omnibus bill, Senate Bill 1249 / House Bill

1143, which would eliminate PA requirements for all generic medications and any admission, service, treatment,

procedure or medication with low denial rates or an evidence-base for treatment. In addition, the bill would eliminate

the ability of plans to issue a PA denial based on an administrative or technical defect in the claim, such as an

incomplete authorization request and would require a PA approval for a service or medication to remain approved for

the duration of treatment, among other provisions. At the federal level, recent legislation seeking to place new

requirements on Medicare Advantage plans using PA was estimated by the CBO to add $16 billion in costs over a

10-year period as a result of greater use of services.

Prior authorization programs affect both member and provider satisfaction, as well as having clinical considerations. It

is also important that stakeholders understand the potential cost impacts to eliminating the use of PA on commercial

insurance premiums and member cost sharing, as well as state Medicaid program expenditures. Generally, PA reduces

unnecessary utilization and shifts the mix of services utilized, both of which have the effect of reducing costs, all else

equal. If PA is removed, it is expected that more services would be utilized and the mix of services would shift towards

higher cost treatments, both of which would increase overall costs.

https://www.ama-assn.org/system/files/2021-04/prior-authorization-survey.pdf

https://www.ama-assn.org/practice-management/prior-authorization/what-prior-authorization

https://www.cbo.gov/system/files/2022-09/hr3173_0.pdf

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 5

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

The Massachusetts Association of Health Plans (MAHP) engaged Milliman to model the cost impacts of limiting or

eliminating PA on:

▪ Massachusetts employers - who generally pay a large portion of premium under employer sponsored plans

▪ Massachusetts consumers - either individual purchasers or employees, who pay a portion of the premium,

as well as deductibles, coinsurance, and copays

▪ The Commonwealth of Massachusetts, which includes budget impacts to MassHealth (Massachusetts’s

Medicaid and Children’s Health Insurance Program) and the State’s obligations to comply with the

Massachusetts’ health care cost growth benchmark

This report is intended to inform stakeholders, which could include but is not limited to, employers, employees and

other individual consumers, commercial plans, MassHealth, providers, lawmakers, and other policy makers about the

range of potential financial impacts (i.e., changes in premiums, capitation rates under a managed Medicaid program

and consumer cost sharing) of various changes to State laws intended to limit the plan practice of PA as a method of

controlling costs. This report is technical in nature and readers with limited background in healthcare should consult a

qualified professional when interpreting these results.

ANALYSIS PROCESS OVERVIEW

We used the following analysis process to arrive at the results in this paper:

1. Construct hypothetical but representative PA code lists from data contributed by Massachusetts health plans

for both Medicaid and commercial business.

2. Apply the codes to Milliman data and flag claims data that contains services subject to PA.

3. Summarize this claims experience to understand the scope of costs subject to PA.

4. Using data contributed by health plans in Massachusetts, calculate the average effect (i.e., net reduction in

costs) of PA programs on medical and prescription drug costs subject to PA.

5. Apply these effectiveness measures to claims subject to PA and recalculate implied premiums or capitation

rates assuming effective PA is removed.

6. Using data summarized from publicly available studies, calculate an average “sentinel effect.”

7. Apply the sentinel effect to costs, similar to Step 5.

8. Add together the effects of Steps 5 and 7 to arrive at a total potential impact (cost increase) of removing PA.

The remainder of this report follows the above outline.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 6

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

III. MEDICAL AND DRUG ALLOWED COST RESULTS

We requested information from MAHP member plans operating in the Massachusetts commercial and managed

Medicaid markets relating to medical and prescription drug claim codes, including Healthcare Common Procedure

Coding System (HCPCS), Current Procedural Terminology (CPT), and National Drug Codes (NDCs). We use this

information to identify services and drugs that are currently subject to PA.

The scope of services and associated procedure codes that are subject to PA vary by plan. To accommodate this

variation, we construct two scenarios representing PA programs in Massachusetts characterized as either a broad or

narrow scope of services. The broad list reflects codes commonly reviewed by plans with more comprehensive PA

programs. The narrow PA list is a subset of broad and reflects only the codes more commonly reviewed by plans with

less comprehensive PA lists. We also note considerable differences in the services reviewed between commercial and

the managed Medicaid markets,

so we created separate broad and narrow lists for each market. Please note, that

the broad and narrow lists do not represent any single plan operating in Massachusetts, but rather are composites that

we believe are generally representative of practices in the commercial and Medicaid markets. Additionally, the

hypothetical broad and narrow lists overlap considerably in terms of services that are included.

We then use these lists to identify the portion of allowed

claims cost in each service or benefit category that is subject

to PA. We use calendar year 2022 claims experience from Milliman’s proprietary database for commercial claims and

use calendar year 2020 claims data from the Transformed Medicaid Statistical System (T-MSIS) for managed Medicaid,

trended to 2023 claims levels. These costs represent composite, Massachusetts commercial and managed Medicaid

benefit plans with average actuarial values of ~85% and 98%, respectively.

We map the broad and narrow scope

code lists to the medical and prescription drug claims data and summarize the claims subject to PA below in Figures 3

and 4.

Figure 3: Summary of Benefit / Services Subject to PA

Commercial

(a)

(b)

(c)

(d) =

(b) / (a)

(e) =

Benefit / Service Categories

Total

Allowed

Claims

PMPM

Claims

Subject to

Broad PA

List PMPM

Claims

Subject to

Narrow PA

List PMPM

Broad % of

Total

Narrow % of Total

Inpatient

$90.75

$71.09

$62.52

78.3%

68.9%

Outpatient Surgery

$72.62

$42.22

$19.24

58.1%

26.5%

Drugs in the Medical Benefit

$43.63

$31.80

$29.61

72.9%

67.9%

Other Outpatient

$41.36

$23.91

$5.45

57.8%

13.2%

Radiological Services

$28.18

$18.04

$15.71

64.0%

55.7%

Other Professional

$55.12

$17.37

$2.17

31.5%

3.9%

Physical / Occupational / Speech

Therapies

$6.43

$6.04

$5.27

93.8%

81.9%

DME / Prosthetics / Medical Supplies

$10.14

$5.79

$3.83

57.1%

37.8%

Cardiovascular Services

$8.14

$5.09

$3.17

62.5%

38.9%

Ancillary Services

$27.77

$2.85

$1.78

10.3%

6.4%

Laboratory / Pathology

$22.55

$2.01

$1.17

8.9%

5.2%

Specialty Drugs

$85.67

$73.20

$67.66

85.4%

79.0%

Brand Drugs

$52.23

$8.03

$3.48

15.4%

6.7%

Generic Drugs

$22.73

$1.87

$0.12

8.2%

0.5%

Total Medical

$406.69

$226.20

$149.93

55.6%

36.9%

Total Drug

$160.63

$83.10

$71.25

51.7%

44.4%

Total

$567.32

$309.30

$221.18

54.5%

39.0%

Employers purchasing coverage, state Medicaid agencies, and Medicare establish PA requirements based on the population served, clinical practice

guidelines, peer-reviewed medical literature, and pharmacoeconomic studies which will result in variation in services subject to PA.

Allowed amount is the total amount received by the provider from the payer and the member.

Actuarial value is the average percentage of patient costs paid by the plan or (plan liability) / (total allowed cost)

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 7

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Figures 3 and 4 illustrate the portion of medical and prescription drug allowed costs subject to a broad or narrow PA

program, by benefit category, expressed as both PMPM amounts and percentages of total allowed costs for both

commercial and Medicaid.

From Figure 3 (Commercial), we note the following:

▪ Under the broad scope PA scenario, $309 PMPM, or 54.5%, of authorized medical costs are subject to PAs

of various types, as compared to $221 PMPM, or 39.0%, for the narrow scope PA scenario.

In previous

work,

we have observed similar differences in broad versus narrow costs.

▪ For prescription drugs, about 52% of authorized allowed costs are subject to PA, driven heavily by specialty

drugs, where 85% of costs were for drugs subject to PA under the broad scope PA scenario.

▪ A significant portion of costs subject to the broad and narrow scope PA programs (about 71% and 81% of

allowed costs, respectively) are found in just the top three medical categories plus specialty drugs. This implies

that any limitations to PA applied to these categories could have a larger impact on premiums than the

remaining categories.

Figure 4: Summary of Benefit / Services Subject to PA

Medicaid Managed Care

(a)

(b)

(c)

(d) =

(b) / (a)

(e) =

Benefit / Service Categories

Total

Allowed

Claims

PMPM

Claims

Subject to

Broad PA

List PMPM

Claims

Subject to

Narrow PA

List PMPM

Broad % of

Total

Narrow % of

Total

Inpatient

$64.37

$9.81

$2.81

15.2%

4.4%

Outpatient

$136.62

$39.33

$17.57

28.8%

12.9%

Professional

$115.76

$56.41

$6.08

48.7%

5.3%

Ancillary

$2.25

$0.50

$0.31

22.1%

13.8%

Specialty Drugs

$77.90

$56.61

$45.76

72.7%

58.7%

Brand Drugs

$85.89

$25.86

$16.46

30.1%

19.2%

Generic Drugs

$42.18

$10.03

$2.78

23.8%

6.6%

Total Medical

$319.00

$106.05

$26.77

33.2%

8.4%

Total Drug

$205.97

$92.50

$65.00

44.9%

31.6%

Total

$524.97

$198.55

$91.78

37.8%

17.5%

From Figure 4 (Medicaid), we note the following:

▪ Under the broad scope PA scenario, $199 PMPM, or 37.8%, of authorized medical costs are subject to PAs

of various types, as compared to $92 PMPM, or 17.5%, for the narrow scope PA scenario.

▪ For prescription drugs, about 45% of authorized allowed costs are subject to PA, driven heavily by specialty

drugs, where 73% of costs were for drugs subject to PA under the broad scope PA scenario.

▪ Compared to commercial, a lower percent of medical services are flagged as subject to PA, with more of the

claims subject to PA coming from prescription drugs.

▪ The Commonwealth of Massachusetts Medicaid program has various requirements related to PA that MCOs

impact the list of service codes that are flagged.

These percentages, while reasonable to convey the general magnitude of medical cost typically subject to PA, are understated as they represent the

cost after the effect of the PA program.

Potential impacts on commercial costs and premiums related to the elimination of prior authorization requirements (milliman.com)

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 8

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

The distributions for both commercial and Medicaid above suggest that significant portions of health care costs can be

subject to PA. Moreover, limitations to, or eliminations of, PA programs will have varying effects on health care costs

depending on a plan’s current scope of services subject to PA and whether limitations on the use of PA are applied to

all services or targeted to specific service categories. For example, the removal of PA requirements for medical or

prescription drug costs that are currently heavily managed using PA (such as inpatient admissions, outpatient surgeries

or specialty drugs) will have a greater impact on overall costs than eliminating PA programs on lower cost categories

of care (such as therapies or generic drugs).

It should be noted that the above distribution of allowed claims costs that are currently subject to PA review in

Massachusetts does not include any new, innovative, and likely expensive therapies that may otherwise be candidates

for effective PA. As these innovations proliferate, a larger percentage of costs would likely be subject to PA than what

is shown and any removal or limitation of PA will also have correspondingly larger impacts than shown.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 9

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

IV. RANGE OF IMPACTS TO PREMIUMS AND MEMBER COST

SHARING

We noted earlier that commercial premiums and capitation rates are likely to increase as a result of increasing utilization

once PA is removed. We estimate the impact of removing PA on commercial premiums by calculating the increase in

medical and prescription drug paid benefit expense (i.e., the portion of allowed claims for which the plan is responsible)

by a measure of program effectiveness (shown in Table 5) and for a “sentinel effect,” discussed further below. We

increase benefit expense for this effect and then recalculate premiums.

For Medicaid, due to data limitations, we assume that the allowed claims also represent the MCO’s liability (i.e., allowed

claims are reasonably close to paid benefit expense) to calculate the impact of PA on Medicaid managed care capitation

rates. Due to very low member cost sharing under Medicaid, this simplifying assumption will not change overall

conclusions of our analysis.

The impact to commercial premiums or Medicaid managed care capitation rates when PA is removed will be comprised

of two additive effects:

▪ Authorizations submitted – Requests for approval of services or drugs that are denied (with or without an

alternative treatment path) when PA is active will no longer be denied or redirected when PA is removed.

▪ Authorizations previously unsubmitted – Under PA, requests for approval of various services may not be

getting submitted as the provider is reasonably sure the request will be denied or redirected to another

treatment path. Once PA is removed, these previously unsubmitted requests to have services covered will

now will likely be performed. This phenomenon is referred to as a “sentinel effect,” which we discuss in more

detail below.

These effects are additive because the authorizations in each category are mutually exclusive. In the case of the

authorizations currently (with PA in place) being submitted, the effect of these is contained within the claims data used

for this analysis, since all plans in our database use PA. However, for the authorizations not currently being submitted,

they are, by definition, excluded from our data and must be added to the impact of the first category.

We analyze each of these effects and discuss them separately. We also provide a combined impact at the end of this

section.

IMPACT ON SUBMITTED AUTHORIZATIONS AND RELATED CLAIMS EXPENSE

We define “effectiveness” of PA as the net impact on submitted claims costs following the approval, denial, or redirection

of a service or prescription drug undergoing PA. For example, a procedure requested in the inpatient setting could be

redirected to the outpatient setting as a result of PA, which may result in lower costs, but not a full reduction in costs

from the original requested service. Again, a particular specialty drug may be denied, but an alternative course of

treatment may be approved, such that there is some net effect of the denial and the approval. An effectiveness rate in

this analysis of 8% means that if PA had not been in place, medical or prescription drug utilization (and thus costs and,

ultimately, premiums or capitation rates) would have been 8% higher.

We derive the best estimates for medical and pharmacy PA effectiveness from information and data supplied by plans

that are members of MAHP. Overall effectiveness can vary from plan to plan, even when the scope of services is

similar, due to varying clinical considerations, such as medical policy and clinical judgement, as well as the plan’s

resources and rigor applied to the PA process internally. We adjust for outlier values and calculate unweighted average

effects across the plans that submitted data. The effectiveness of PA by benefit category from this plan survey data is

shown in column (b) in Figures 5 and 6 below. Column (a) shows the paid benefit expense that is subject to PA under

a broad scope definition for commercial and allowed benefit expense for Medicaid managed care. Similar tables

(Figures 7 and 8) for narrow PA scope are shown in Appendix A.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 10

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Figure 5: Paid Benefit Expense Impacts of PA

Broad PA Scope

Commercial

(a)

(b)

Benefit / Service Category

Costs Subject

to Prior

Authorization

(PMPM)

Effectiveness*

$ Increase

(PMPM)

Inpatient

$62.79

$2.28

Outpatient Surgery

$34.02

$2.56

Drugs in the Medical Benefit

$29.29

$1.87

Radiological Services

$13.13

$0.84

Phys. / Occ / Speech Therapies

$3.82

$0.20

Cardiovascular Services

$3.93

15%

$0.69

DME / Prosthetics / Medical Supplies

$4.96

10%

$0.55

Pathology / Lab

$1.69

25%

$0.56

Other Professional

$13.52

$0.71

Other Outpatient

$20.41

$1.54

Ancillary Services

$2.31

$0.17

Specialty Drugs

$70.11

20%

$17.53

Brand Drugs

$7.28

20%

$1.82

Generic Drugs

$1.13

15%

$0.20

Total

$268.39

10.5%

$31.52

*Source: MAHP member plans.

Figure 6: Benefit Expense Impacts of PA

Broad PA Scope

Medicaid Managed Care

(a)

(b)

Benefit / Service Category

Costs Subject

to Prior

Authorization

Effectiveness*

$ Increase

Inpatient

$9.81

$0.46

Outpatient

$39.33

10%

$4.37

Professional

$56.41

$4.25

Ancillary

$0.50

$0.03

Specialty Drugs

$56.61

10%

$6.29

Brand Drugs

$25.86

15%

$4.56

Generic Drugs

$10.03

15%

$1.77

Total

$198.55

9.9%

$21.73

*Source: MAHP member plans.

Since PA is a common industry practice, the claims data utilized for this analysis inherently contains the effects of these

programs.

Therefore, in calculating the cost impacts of eliminating PA in column (c), we assume that PA is currently

reducing costs covered by the percentage in column (b). Therefore, in Figures 5 through 8, we calculate values in

column (c) by first grossing up the portion of benefit expense subject to PA (in column (a)) by a factor of 1 less the

effectiveness rate and then subtracting the original portion subject to PA in column (a).

Using estimates from Figures 5 through 8, we calculate the potential increases to premium for commercial plans and

potential increases to managed care capitation rates.

The data also inherently contains the additional but unknown impact of suppression of submitted services due to the sentinel effect mentioned above.

For additional discussion of sentinel effects and other key considerations and limitations of the data, see the “Sentinel Effect” section below.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 11

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Our analysis and results are based on a composite or typical plan with a composited broad (or narrow) scope of

services. The actual impact of removing or limiting a PA program will vary considerably between plans. Each plan’s

definition of services covered by their PA program will be unique, as will be the effectiveness of their specific program.

Even within a payer, it is possible that their definition of PA services and the stringency of their protocols may vary for

each service category or line of business. While we present a range of possible impacts, actual outcomes could be less

than or greater than these results.

Figure 9: Premium Impacts (PMPM) of PA Removal

Broad and Narrow Scope

Commercial

Scenario 1: Broad PA

Scope

Scenario 2: Narrow PA

Scope

Total Paid Ben. Expense Modeled

$481.17

Total Premium Modeled (86% Loss Ratio)

$559.50

Benefit Expense Subject to PA (Figures 5 and 7)

$268.39

$195.30

Effectiveness of PA

10.5%

11.1%

Benefit Expense Increase

$31.52

$24.49

Premium Increase (86% Loss Ratio)

$36.66

$28.48

Recalculated Premium

$596.15

$587.97

Premium Increase %

6.6%

5.1%

Figure 10: Capitation Rate Impacts (PMPM) of PA Removal

Broad and Narrow Scope

Medicaid Managed Care

Scenario 1: Broad PA

Scope

Scenario 2: Narrow PA Scope

Total Allowed Expense Modeled

$525.01

Total Capitation Modeled (90% Loss Ratio)

$583.34

Allowed Expense Subject to PA (Figures 6 and 8)

$198.55

$91.78

Effectiveness of PA

9.9%

10.7%

Benefit Expense Increase

$21.73

$11.04

Capitation Increase (90% Loss Ratio)

$24.14

$12.27

Recalculated Capitation

$607.49

$595.61

Capitation Rate Increase %

4.1%

2.1%

As shown in Figure 9, we estimate PA removal could increase commercial premiums approximately $28 and $37

PMPM, or 5.1% and 6.6%, without considering the sentinel effect. Similarly, we estimate removal of PA could increase

Medicaid capitation rates by between approximately $12 and $24 PMPM, or 2.1% to 4.1%, as shown in Figure 10.

We recognize that other aspects of PA, and any limitations placed on the practice, will influence underlying benefit

expense costs and result in premium or capitation rate increases once PA is removed. These aspects include the

potential for decreased administrative costs related to PA for both provider and plan, alternate paths of care after denial

of an authorization, and others. We discuss these in more detail in the “Considerations” section below. Rather than

attempting to consider each of these impacts explicitly in this analysis, we broadly consider these effects through

general conservatism in estimating effectiveness rates and by the use of ranges.

Moreover, the increases in utilization of services and the change in mix of services towards more costly treatments may

not immediately manifest and when these do manifest, plans may or may not reflect these changes immediately or

uniformly in commercial premiums. Likewise, impacts to Medicaid costs may not be immediately reflected in the setting

of capitation rates. However, we believe ultimately as costs rise, this will need to be reflected in commercial premiums

and Medicaid capitation rates or plans will incur ongoing financial losses.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 12

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Figure 11 shows the development of the estimated impacts to member cost sharing for commercial.

Note, that

member cost sharing percentage increases will not match premium percentage increases as the distribution of member

cost sharing on services subject to PA across benefit categories is not the same as the paid claims distribution used to

develop premiums.

Figure 11: Impacts to Member Cost Sharing (PMPM) of PA Removal

Broad and Narrow Scope

Commercial

Scenario 1:

Broad PA Scope

Scenario 2:

Narrow PA Scope

(a)

Allowed Costs Before PA

$567.32

(b)

Benefit Expense Before PA

$481.17

Cost Sharing

$86.15

(d)

Allowed After PA

$602.35

$594.04

(e)

Paid After PA

$512.69

$505.66

(f) = (d) – (e)

Cost Sharing

$89.66

$88.38

(g) =(f) – (c)

Change in Cost Sharing $

$3.51

$2.23

(h) = (f) / (c)

Change in Cost Sharing %

4.1%

2.6%

Member cost sharing impacts reflect the lower proportion of costs paid by plan members, typically due to out-of-pocket

maximums in the benefit plans. Note, we do not model the impact of cost sharing in Medicaid, as cost sharing is typically

limited.

IMPACT OF UNSUBMITTED AUTHORIZATIONS AND RELATED CLAIMS EXPENSE (SENTINEL EFFECT)

The analysis in the immediately preceding section utilizes claims data reflecting provider behavior that has been

influenced by the presence of PA. It has been demonstrated in other studies that the addition or removal of PA affects

utilization patterns.

It is possible that a portion of claims for services that require PA were not submitted, leading

to an overall lower level of costs in existing plan claims data. One possible explanation for this phenomenon is that

providers may anticipate the requested service will be denied or directed to an alternative treatment path; therefore,

they forgo requesting an authorization entirely and the resulting costs never materialize.

The effect of costs being suppressed to a lower level in the presence of PA is a specific instance of what is generally

referred to as the “sentinel effect.” This effect is the tendency for people to act differently when they know they are

being observed and measured. In the case of PA, a sentinel effect occurs when the requirement to request a PA deters

the provider from submitting the request in the first place. For example, the sentinel effect may happen when the

provider anticipates the request will not meet the plan’s medical necessity criteria, and thus the provider perceives the

submission as wasted effort. The sentinel effect is also seen when a provider notifies a member that a course of

treatment is not viable because of the existence of PA and redirects the member to alternative treatments. In some

cases, this may result in the provider selecting an alternative service or alternative lower level of care, which will be

reflected in the claims submission. Studies have shown a sentinel effect on various medical services and prescription

drugs to be material and notably greater than the effect of the denials or redirection of care on already submitted claims,

as previously quantified above.

When PA is removed, there may be a period when utilization of a particular service might stay at previous levels, as

providers have become accustomed to the PA criteria for approval resulting in a continued lower level of utilization.

After some time, utilization may begin to increase as providers begin to perform procedures they previously would not

have, knowing that they no longer will be denied. Utilization may also increase as providers new to the market do not

get evidence-based feedback. While our estimates use assumptions from the literature to account for the ultimate

sentinel effects, we did not account for this durational aspect in our analysis.

Cost sharing impacts are not estimated for Medicaid due to data limitations. However, given Medicaid cost sharing is typically very small, we expect

small, immaterial impacts.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8822442/

https://www.nejm.org/doi/10.1056/NEJM199506153322406?url_ver=Z39.88-

2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200www.ncbi.nlm.nih.gov

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1885015/pdf/bcp0061-0341.pdf

https://www.jmcp.org/doi/pdf/10.18553/jmcp.2014.20.5.447

See Appendix B for references

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 13

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

To estimate the sentinel effect, we performed a literature review and analyzed studies related to the removal or addition

of PA requirements on various services. We also considered a limited number of studies that did not relate to PA per se,

but rather to other utilization management measures that would have a similar effect or provided insights into physician

behavior in the presence or absence of various incentives. The list of studies we reviewed can be found Appendix B

(Figure 12), along with the numerical results from each.

As seen in literature, summarized in Appendix B (Figure 12), the effect of the removal of PA or the general response

of providers to changes in financial incentives is significant, ranging from 104% to 793% increase in the utilization of a

particular service. However, for purposes of this analysis, we use a range based on our judgement, and we choose to

apply a substantial discount to these estimates that is much more conservative than the average effect observed across

all of these studies.

The reasons for this adjustment and the use of a range are:

▪ Many of the benefit (service) categories used in this analysis are not reflected directly in any of the available

studies. For example, we found no studies that applied directly to inpatient admits and outpatient surgeries,

two of the largest categories in this analysis.

▪ The studies in the literature review are performed on very specific services. The impact of PA on a specific

service does not adequately represent the impact of the removal of PA on an entire service category (see any

of Figures 3 through 6 for the service categories used).

As an example, our literature review cites a study done by the Government Accountability Office (GAO) on

power mobility equipment and non-emergency ambulance use, among others. The impact of PA

implementation was to reduce by over half the number of services being utilized. This suggests that by

removing PA, utilization would greater than double. Likewise, our literature review cites a study

noting

decreases to imaging services, such as CT scans and MRI’s of between 50% and 94% that correspondingly

imply very large increases in utilization if PA were removed.

In these cases, there were specific targeted approvals required under the PA programs that focused on these

specific areas of overuse. It would not be reasonable assumption to take these magnitudes of decrease in

utilization (or increases when PA is removed) and apply them generally across broad categories that likely do

not contain the same degree of overuse or misuse across the entire broad category.

▪ Some of the studies included were performed outside the U.S. and application may be less appropriate for

the U.S. market.

▪ The sentinel effect deals with human behavior and decision making, which can be hard to predict in specific

situations and even harder to generalize across a broad range of situations.

For this reason, we use conservative estimates of 10%, 20% and 30% for the sentinel effect applied to the entire

amount of benefit expense subject to PA (line (c) in the tables below).

Building on our analysis from Figures 9 and 10, we calculate an additive premium or capitation rate impact of the

sentinel effect. Figures 13 and 14 below assume the broad PA scope. For brevity we include equivalent tables for the

narrow scope in Appendix A (Figure 15 and 16).

Figure 13: Premium Impacts (PMPM) due to Sentinel Effect

Broad Scope

Commercial

(a)

Total Paid Ben. Expense Modeled

$481.17

(b)

Total Premium Modeled (86% Loss Ratio)

$559.50

(c)

Paid Benefit Expense subject to PA

$268.39

(d)

% Increase in Submitted Claims

10%

20%

30%

(e) = (c) * (d)

Paid Benefit Expense Increase

$26.84

$53.68

$80.52

(f) = (e) / .86

Premium Increase (86% Loss Ratio)

$31.21

$62.42

$93.63

(g) = (f) + (b)

Recalculated Premium

$590.71

$621.91

$653.12

(h) = (g) / (b) – 1

Premium Increase %

5.6%

11.2%

16.7%

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8822442/pdf/10.1177_2333392817732018.pdf

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 14

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Figure 14: Capitation Rate Impacts (PMPM) due to Sentinel Effect

Broad Scope

Medicaid Managed Care

(a)

Total Allowed Ben. Expense Modeled

$525.01

(b)

Total Capitation Modeled (90% Loss Ratio)

$583.34

(c)

Benefit Expense subject to PA

$198.55

(d)

% Increase in Submitted Claims

10%

20%

30%

(e) = (c) * (d)

Benefit Expense Increase

$19.86

$39.71

$59.57

(f) = (e) / .90

Premium Increase (90% Loss Ratio)

$22.06

$44.12

$66.18

(g) = (f) + (b)

Recalculated Capitation

$605.41

$627.47

$649.53

(h) = (g) / (b) – 1

Capitation Increase %

3.8%

7.6%

11.3%

COMBINED IMPACTS

Using our best estimates of PA effectiveness and the corresponding impacts of removing PA on services currently

being captured in our data from Figures 5 and 6, combined with the ranges of impacts for sentinel effects on potentially

unsubmitted authorizations not contained in our data, Figures 17 and 18 show the total range of impact across two PA

scopes (broad and narrow) and three sentinel effect estimates for commercial and Medicaid.

Figure 17: Overall Premium Impacts (PMPM) of PA Removal

Broad and Narrow Scope

Commercial

Total Paid Ben. Expense

Modeled

$481.17

(a)

Total Premium Modeled

(86% Loss Ratio)

$559.50

Broad Scope PA

Narrow Scope PA

(b)

Benefits subject to PA

$268.39

$195.30

(c)

PA Effectiveness on Submitted

Claims (Figure 5 and 7)

$31.52

$24.49

(d)

Sentinel Effect

10%

20%

30%

10%

20%

30%

(e) =

(d) x (b)

Sentinel Effect on Unsubmitted

Claims (Figure 13 and 15)

$26.84

$53.68

$80.52

$19.53

$39.06

$58.59

(f) =

(e) + (c)

Total Benefit Expense

Increase

$58.36

$85.20

$112.04

$44.02

$63.55

$83.08

(g) =

(f) / .86

Premium Increase

(86% Loss Ratio)

$67.86

$99.07

$130.28

$51.19

$73.90

$96.61

(h) =

(g) + (b)

Recalculated Premium

$627.36

$658.57

$689.78

$610.68

$633.39

$656.10

(j) =

(h) / (a) -1

Premium Increase %

12.1%

17.7%

23.3%

9.1%

13.2%

17.3%

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 15

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Figure 18: Overall Capitation Rate Impacts (PMPM) of PA Removal

Broad and Narrow Scope

Medicaid Managed Care

Total Allowed Expense Modeled

$525.01

(a)

Total Capitation Modeled (90% Loss

Ratio)

$583.34

Broad Scope PA

Narrow Scope PA

(b)

Benefits subject to PA

$198.55

$91.78

(c)

PA Effectiveness on Submitted Claims

(Figure 6 and 8)

$21.73

$11.04

(d)

Sentinel Effect

10%

20%

30%

10%

20%

30%

(e) =

(d) x (b)

Sentinel Effect on Unsubmitted Claims

(Figure 14 and 16)

$19.86

$39.71

$59.57

$9.18

$18.36

$27.53

(f) =

(e) + (c)

Total Allowed Expense Increase

$41.58

$61.44

$81.29

$20.22

$29.39

$38.57

(g) =

(f) / 0.9

Capitation Increase (90% Loss Ratio)

$46.20

$68.26

$90.32

$22.47

$32.66

$42.86

h =

(g) + (b)

Recalculated Capitation

$629.55

$651.61

$673.67

$605.81

$616.00

$626.20

(j) =

(f) / (a) -1

Capitation Rate Increase %

7.9%

11.7%

15.5%

3.9%

5.6%

7.3%

Overall, the elimination of PA in the commercial market under the broad scope PA scenario could result in commercial

premium increases ranging from 12.1% to 23.3%. Under the narrow PA scenario, commercial premium increases range

from 9.1% to 17.3%.

Similarly, the elimination of PA in Medicaid under the broad scope PA scenario could result in capitation rate increases

ranging from 7.9% to 15.5%. Under the narrow PA scenario, capitation rate increases range from 3.9% to 7.3%.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 16

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

V. CONSIDERATIONS

The information presented above estimates costs associated with eliminating PA based on variations in the scope of

services subject to PA, averages of PA effectiveness across numerous plans, application of these effectiveness rates

to broad service type categories and estimates of sentinel effects with material actuarial judgement applied. These are

simplifications that, in our estimation, do not affect the magnitude of impacts of removing PA. However, calculating

precise impacts to changes in PA programs can be challenging for various reasons that would otherwise complicate

analysis of impacts. We consider some of these below.

KEY CONSIDERATIONS

Net Savings

Prior authorization often results in substituting one set of services for another. Examples of this include physical therapy

in lieu of back surgery, a lower cost imaging in place of higher cost imaging, redirecting a service to an appropriate

lower cost level of care or the use of a lower cost, therapeutically similar prescription drug first before going to a more

expensive one (otherwise known as step therapy). Thus, a limitation or elimination of a particular PA rule may not result

in an increase in claims for the full cost of that service, to the extent a lower cost service is subsequently utilized. For

example, in an analysis of PA restrictions in Medicare, researchers calculated that PA policies resulted in a $112 per

beneficiary-year reduction in prescription drug spending on drugs with PA, and a $16 per beneficiary-year increase in

spending on unrestricted drugs.

It is outside the scope of this analysis to account directly for all of the various situations

where a PA rule achieves a net savings as opposed to a full savings. Instead, plans submitting data for this study

attempted to account for the net savings by service category in their estimates of effectiveness rates.

Plan and Provider Administrative Costs

PA programs have administrative costs associated with them that could be reduced if a smaller scope of services is

considered under the program or if PA were eliminated altogether. Therefore, any increase in claim costs associated

with eliminating PA programs is likely offset with at least some administrative expense savings. In certain cases,

multiple departments may be involved in doing some form of PA or have some related cost to it (e.g., medical personnel

versus IT staff). Likewise, provider costs for administering their side of the PA process would decrease, either improving

financial performance for providers or serving to reduce physician compensation / fee schedules. If the latter were the

case, there would be an additional net savings.

However, since eliminating PA does not eliminate health plan requirements to manage cost and quality, a portion of

administrative resources would most likely be redirected to other programs to manage quality and cost. As a result,

administrative costs could increase in other areas that offset the reductions noted above related to PA directly. For

example, it is likely plans will increase retrospective reviews including fraud, waste and abuse controls and payment

integrity efforts. These alternate activities in lieu of PA could also have some mitigating effects on costs increases

shown herein.

Our analysis does not account for either plan or provider administrative costs, whether increases or decreases, that

are associated with PA; therefore, impacts on premiums in this analysis could be overstated or understated depending

on the net of effect of these changes.

OTHER CONSIDERATIONS

Purpose of Prior Authorizations

In our experience with plan clients, we note multiple purposes for the use of PA. Most noted the need for medical

necessity review, mitigating the overuse or misuse of services, intervening with alternate treatment paths that are either

more effective or equally effective but lower cost, identifying members that may benefit from care management and

ensuring safety controls. Depending on the reason for the implementation of a PA rule and the nature of the limitation

on the use of PA, the impact to premiums or capitation rates could vary. For example, if a blanket prohibition on PA

was required on drugs on the medical benefit, the impact would be substantial, as these drugs are very high cost on a

per service basis. On the other hand, limitations on PA related to physical therapy may be less impactful, not simply

because it is lower cost per service, but also because the nature of PA, as it relates to physical therapy, is typically only

regarding number of services or length of treatment, not a complete denial.

https://www.nber.org/system/files/working_papers/w30878/w30878.pdf

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 17

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Plans point to continued increases in unit prices for hospital and provider services and excessive increases in

prescription drug costs as the primary challenges in keeping health care affordable for employers and consumers. Prior

authorization mitigates unnecessary care (physicians themselves estimate that nearly a quarter of care is

unnecessary

) and helps protect against potentially harmful care. Inappropriate care can be more than merely

wasteful, it can be harmful, such as exposure to unnecessary radiation, missed diagnoses and false positives, and

ineffective procedures and treatments, and decreases in health care affordability.

Value-Based Care Impacts

Impacts of eliminating PA might be smaller than they might otherwise be prior to the advent of value-based care and

providers being at some degree of risk for utilization under these types of contracts. This is particularly true for systems

or provider groups that are already efficient. Providers have different incentives under value-based care models than

they would under a predominantly fee-for-service environment, and this may induce less over-utilization of services.

However, this is heavily dependent, at a minimum, on the nature of the risk-based contract and the precise incentives

(or disincentives) involved. In some cases, the narrower PA lists we observed for some plans may be linked to situations

where they have risk-based contracts with providers shifting the responsibility for utilization management onto the

provider. That being said, the authors’ experience is that the realignment of incentives in commercial and Medicaid

VBP models has yet to result in a transformational change in behavior for providers. To date, a large percentage of

commercial and Medicaid managed care members will not be covered in VBP models

nor are the incentives sufficient

to transform their delivery model.

Medicaid Drug Rebate Program (MDRP) Implications

Due to the statutory rebate requirements of the MDRP, oftentimes the more expensive branded prescription drugs have

a lower net cost to state Medicaid programs than their generic alternatives. States use a preferred drug list (PDL) to

increase utilization of these lower net cost products, as well as negotiate supplemental rebates through PDL placement.

States typically require a PA for non-preferred products and most states, including Massachusetts, require MCO to

follow the state’s PDL. Limiting or eliminating the use of PA in Medicaid would decrease the State’s ability to maximize

rebates and would shift utilization to products that do not have the lowest net cost to the State.

Impact on the Cost Growth Benchmark

Massachusetts was the first state in the nation to develop a health care expenditure target. With the passage of

Chapter 224 of the Acts of 2012, plans and providers became subject to an annual target for the rate of growth in total

health care expenditures. Set annually by the Health Policy Commission (Massachusetts’ independent state agency

charged with developing policy to reduce health care cost growth and improve the quality of patient care), the

benchmark is set at 3.6% (the growth rate of potential state gross domestic product) for 2023. Total health care

expenditures (THCE) is calculated on a per capita basis to control for increases in spending due to population growth

and includes three components: 1.) all medical expenses paid to providers by private and public plans, including

Medicare and Medicaid; 2.) all patient cost-sharing amounts; and 3.) the net cost of private insurance (e.g., admin

expenses and operating margins for commercial plans). Limiting or eliminating the use of PA could impact plans’ ability

to keep THCE growth under the benchmark and could shift utilization to medications, services, and treatments with

higher costs / or to sites of care with higher costs.

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0181970

https://hcp-lan.org/apm-measurement-effort/2020-2021-apm/2021-infographic/

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 18

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

VI. METHODOLOGY

DATA SOURCES

For commercial claims, we relied on Milliman’s 2022 Consolidated Health Cost Guidelines Sources Database (CHSD),

limited to members from Massachusetts.

For managed Medicaid claims, we relied on 2020 data from T-MSIS Research Identifiable Files (RIF), limited to

members in Massachusetts. T-MSIS contains detailed Medicaid claims and enrollment data for 100% of the Medicaid

population. This dataset contains header and detail records for inpatient hospital services, long-term care services,

other services, and pharmacy claims.

For both commercial and managed Medicaid, prescription drug costs and any effects do not include manufacturer

rebates.

The T-MSIS data has several limitations which could impact our results:

▪ For Medicaid encounter claims, only utilization data is available. We estimated Medicaid managed care costs

based on the average charge by service (DRG, procedure code, or NDC adjusted for days supply) in the

Massachusetts fee-for-service Medicaid data.

▪ Limited information was reliably populated to group medical claims by detailed service category, so we only

group medical claims based on major service category (inpatient, outpatient, professional, and other) based

on the benefit type indicated in T-MSIS.

▪ Several fields of interest to the analysis may be less reliable than the propriety dataset used for modeling the

impact on the commercial market. However, we generally believe the fields used are reliable for the scope of

this analysis.

▪ The Medicaid data is older than our commercial dataset and may not properly align with current clinical

standards for PA and new procedures and drugs.

▪ The COVID-19 pandemic had a material impact on Medicaid enrollment, utilization, and treatment patterns in

2020 and beyond; it is possible the utilization and cost patterns in 2020 may not reflect 2023.

▪ The T-MSIS RIF data is subject to redaction to protect PHI and data exports with fewer than 10 observations.

For our modeled PA population, we only include members subject to managed care (as indicated by T-MSIS enrollment)

but include all claims available for those members.

PRIOR AUTHORIZATION SERVICES

We received a list of HCPCS, CPT, and NDC codes subject to each plan’s commercial and Medicaid PA review

program. We reviewed these lists to identify the commonalities and differences and to develop representative lists of

codes subject to PA. We note considerable variation between plans, with some plans reviewing a significantly larger

number of codes, and others focusing on a smaller set of services, which were often also reviewed by the broader

scope plans. We also notice differences between commercial and managed Medicaid. Plans typically reviewed fewer

medical services (HCPCS / CPTs) for managed Medicaid than commercial. However, Medicaid plans typically review

a larger set of NDCs for pharmacy claims than for commercial, as MassHealth has a single preferred drug list (PDL),

making PA more critical.

Accordingly, we made separate lists of codes subject to PA for commercial and Medicaid, and for each selected a

representative set of services for both a plan with a broad and narrow PA program. The PA lists are not intended to

reflect any specific plan, but rather a representative range of services for a typical plan based on the codes submitted

by all plans.

In our claims datasets, we identify services that are subject to PA, assuming all services on the list are reviewed. For

inpatient facility claims, HCPCS / CPT is typically not populated. We identify inpatient facility claims as subject to PA if

the same member had a service (HCPCS / CPT) subject on or between their admit and discharge dates.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 19

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

EFFECTIVENESS MEASURES

We also received post-appeal effectiveness rates by benefit / service category for MAHP plans’ commercial and

Medicaid PA programs. These reflect the net cost savings of PA denials, including an estimate of any offsetting ultimate

costs from partial denial or care redirection. We reviewed the distribution of effectiveness measures by plan and service

category. There is considerable variance in effectiveness of PA by service category. Prescription drugs and

pathology / lab services show high effectiveness, while inpatient has the lowest effectiveness from the set of service

categories. Effectiveness also varies for a specific service among different plans, although it generally fell into a

relatively small range other than a few outliers. We developed a best estimate of PA effectiveness for each service

category based on the average among plans.

We apply the effectiveness measures and sentinel effect to paid benefit expense for commercial. For Medicaid, due to

data constraints and the relatively small member cost sharing, we applied the effectiveness and sentinel effects to

allowed charges.

For commercial, we convert allowed charges into paid, by calculating an Actuarial Value (AV) ratio of allowed to paid

based on historical data in the CHSD for commercial. A unique AV was calculated for each service category, preserving

the uniqueness of each categories' estimates for paid quantities of reduced effectiveness and sentinel values with the

removal of PA. Cost Sharing metrics were derived as the difference between allowed and paid estimates. To translate

paid claims to premium, we assume commercial plans price to an 86% MLR. Massachusetts has a statutory minimum

MLR of 88% after certain adjustments. We reviewed historical MLR data for Massachusetts plans to determine a pricing

MLR of 86%.

We did not convert allowed claims to paid for Medicaid due to data limitations. We use allowed claims to represent the

approximate liability for MCOs, due to limited cost sharing in Medicaid.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page 20

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

VII. CAVEATS

The report is intended to help quantify the potential range of impacts to commercial health insurance premiums and

managed Medicaid capitation rates due to eliminating prior authorization programs. Other uses may be inappropriate.

The results in this report represent estimates and actual results for any given payer could vary significantly. The

estimates are intended to provide a framework in which to discuss potential impacts of proposed legislation or regulation

of PA practices. It is not intended to provide pricing impacts nor be an estimate for any specific piece of legislation.

This report may not be distributed to any third parties without the prior consent of Milliman. To the extent that the

information contained in this report is provided to third parties, the document, including all appendices, should be

distributed in its entirety. Milliman makes no representations or warranties regarding the contents of this report to third

parties. Similarly, third parties are instructed to place no reliance upon this report prepared by Milliman that would result

in the creation of any duty or liability under any theory of law by Milliman or its employees to third parties.

We relied on claims data from Milliman’s CHSD, as well as Transformed Medicaid Statistical Information System

Research Identifiable Files. We also relied on information submitted by MAHP member plans, specifically the list of

codes (CPT, NDC, and HCPCs) that are subject to PA and effectiveness measure by major benefit / service category.

This information was reviewed for reasonableness but accepted without audit. To the extent the data and information

relied upon is not accurate, or is not complete, the values and conclusions provided in this report may likewise be

inaccurate or incomplete.

Models used in the preparation of our analysis were applied consistently with their intended use. We have reviewed

the models, including their inputs, calculations, and outputs for consistency, reasonableness, and appropriateness to

the intended purpose and in compliance with generally accepted actuarial practice and relevant actuarial standards of

practice (ASOP). The models, including all input, calculations, and output may not be appropriate for any other purpose.

Where we relied on models developed by others, we have made a reasonable effort to understand the intended

purpose, general operation, dependencies, and sensitivities of those models. We relied on input, review, and validation

by other experts in the development of our models.

The results of this report are technical in nature and are dependent upon specific assumptions and methods. No party

should rely on these results without a thorough understanding of those assumptions and methods. Such an

understanding may require consultation with qualified professionals.

Fritz S. Busch and Peter Fielek are members of the American Academy of Actuaries and meet the Qualification

Standards of the American Academy of Actuaries to render the actuarial opinion contained herein.

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

APPENDIX A

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page A - 1

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Figure 7: Paid Benefit Expense Impacts of PA

Narrow PA Scope

Commercial

(a)

(b)

Benefit / Service Category

Costs Subject

to Prior

Authorization

Effectiveness*

$ Increase

Inpatient

$55.22

$2.00

Outpatient Surgery

$15.50

$1.17

Drugs in the Medical Benefit

$27.28

$1.74

Radiological Services

$11.43

$0.73

Phys. / Occ / Speech Therapies

$3.34

$0.18

Cardiovascular Services

$2.45

15%

$0.43

DME / Prosthetics / Medical Supplies

$3.28

10%

$0.36

Pathology / Lab

$0.98

25%

$0.33

Other Professional

$1.69

$0.09

Other Outpatient

$4.65

$0.35

Ancillary Services

$1.44

$0.11

Specialty Drugs

$64.81

20%

$16.20

Brand Drugs

$3.15

20%

$0.79

Generic Drugs

$0.07

15%

$0.01

Total

$195.30

11.1%

$24.49

Figure 8: Allowed Benefit Expense Impacts of PA

Narrow PA Scope

Medicaid Managed Care

(a)

(b)

Benefit / Service Category

Costs Subject

to Prior

Authorization

Effectiveness*

$ Increase

$2.81

$0.13

Outpatient

$17.57

10%

$1.95

Professional

$6.08

$0.46

Ancillary

$0.31

$0.02

Specialty Drugs

$45.76

10%

$5.08

Brand Drugs

$16.46

15%

$2.91

Generic Drugs

$2.78

15%

$0.49

Total

$91.78

10.7%

$11.04

Figure 15: Premium Impacts (PMPM) due to Sentinel Effect

Narrow Scope

Commercial

(a)

Total Paid Ben. Expense Modeled

$481.17

(b)

Total Premium Modeled (86% Loss Ratio)

$559.50

(c)

Paid Benefit Expense subject to PA

$195.30

(d)

% Increase in Submitted Claims

10%

20%

30%

(e) = (c) * (d)

Paid Benefit Expense Increase

$19.53

$39.06

$58.59

(f) = (e) / .86

Premium Increase (86% Loss Ratio)

$22.71

$45.52

$68.13

(g) = (f) + (b)

Recalculated Premium

$582.21

$604.92

$627.63

(h) = (g) / (b) – 1

Premium Increase %

4.1%

8.1%

12.2%

APPENDIX A

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans Page A - 2

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Figure 16: Capitation Rate Impacts (PMPM) due to Sentinel Effect

Narrow Scope

Medicaid Managed Care

(a)

Total Allowed Expense Modeled

$525.01

(b)

Total Capitation Modeled (90% Loss Ratio)

$583.34

(c)

Allowed Expense subject to PA

$91.78

(d)

% claim Increase from Sentinel Effect

10%

20%

30%

(e) = (c) * (d)

Allowed Expense Increase

$9.18

$18.36

$27.53

(f) = (e) / .90

Capitation Increase (90% Loss Ratio)

$10.20

$20.39

$30.59

(g) = (f) + (b)

Recalculated Capitation

$593.54

$603.74

$613.94

(h) = (g) / (b) – 1

Capitation Rate Increase %

1.7%

3.5%

5.2%

MILLIMAN REPORT

_______________________________________________________________________________________________________________________

Massachusetts Association of Health Plans

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

APPENDIX B

Massachusetts Association of Health Plans Page B - 1

Potential Impacts on Costs and Premiums Related to the Elimination of Prior Authorization Requirements in Massachusetts

November 29, 2023

Figure 12: Summary of Literature Quantifying the Sentinel Effect

Therapeutic Area

Implementation or

Removal

Implied

Increase*

Notes

Source

Opioids

PA Implementation

104%

Claims for long-acting opioid analgesics decreased

by 4.1% after implementation of a prior

authorization program in Massachusetts Medicaid.

https://www.jmcp.org/doi/pdf/10.1

8553/jmcp.2014.20.5.447

Cyclooxygenase-

2 (COX-2)

Specific Inhibitors

PA Implementation

293%

The cost per success (defined as no serious

gastrointestinal event) for COX-2 specific inhibitors

with PA was $278 versus $422 without PA.

https://www.jmcp.org/doi/epdf/10.

18553/jmcp.2003.9.4.327?role=ta

Power Mobility

PA Implementation

129%

Compares average monthly expenditures for a PA

demonstration program.

https://www.gao.gov/assets/700/6

92138.pdf

Non-Emergency

Ambulance

PA Implementation

150%

Compares average monthly expenditures for a PA

demonstration program.

https://www.gao.gov/assets/700/6

92138.pdf

Hyperbaric

Oxygen Therapy

PA Implementation

220%

Compares average monthly expenditures for a PA

demonstration program.

https://www.gao.gov/assets/700/6

92138.pdf

Home Health

PA Implementation

750%

Compares average monthly expenditures for a PA

demonstration program.

https://www.gao.gov/assets/700/6

92138.pdf

Cefuroxime

PA Implementation

667%

Cefuroxime experienced an 85% relative reduction

in the use of the drug during the PA period despite

the request rejection rate of only 8%.

https://www.ncbi.nlm.nih.gov/pmc

/articles/PMC1885015/

Direct-Acting Oral

Anticoagulants

Other

714%

DOACs' utilization and costs were much higher in

Canadian provinces with more liberal mechanism

of reimbursement. Utilization was as much as 86%

higher than the national average in one province

with more liberal mechanism of reimbursement.

https://www.ncbi.nlm.nih.gov/pmc

/articles/PMC5749525/

Parenteral

Antimicrobials

PA Implementation

147%

Average monthly expenditures for parenteral

antimicrobials during the 6 months immediately

after full implementation of the PA requirement

decreased by 32%, compared to the 6 months

immediately before implementation.

https://pubmed.ncbi.nlm.nih.gov/9

332517/

Celecoxib and

etoricoxib

Removal of PA

615% and

793%

In Spain, celecoxib use increased 615% in terms of

the daily doses per thousand inhabitants per day

once the PA was removed. Similarly, etoricoxib

experienced a relative increase of 793%.

https://onlinelibrary.wiley.com/doi/

epdf/10.1111/jcpt.12490

Non-generic

NSAIDs

PA Implementation

213%

Following the implementation of a PA program in

Tennessee Medicaid, expenditures decreased by

53% following switching to generic NSAIDs, and

there was a 19% overall NSAID use.

https://www.nejm.org/doi/10.1056

/NEJM199506153322406?url_ver

=Z39.88-

2003&rfr_id=ori:rid:crossref.org&r

fr_dat=cr_pub%20%200www.ncb

i.nlm.nih.gov

CT and MRI

PA Implementation

200%

CT utilization decreased between 76 to 90% one

year following implementation of a nondenial PA

program, and decreased 52 to 88% five years

following implementation. Similarly, MRI utilization

decreased 86 to 94% after one year and between

50 to 75% after five years.

https://www.ncbi.nlm.nih.gov/pmc

/articles/PMC8822442/pdf/10.117

7_2333392817732018.pdf

CT and MRI

Other

N/A

In a review of 459 CT and MRI examinations, 26%

were not considered clinically appropriate.

https://www.jacr.org/article/S1546

-1440(09)00589-4/fulltext

*For PA implementation, the decrease in utilization or cost is converted into an implied increase if PA were removed but using the following formula: 1/ (1-reduction).

For more information about Milliman,

please visit us at:

milliman.com

Milliman is among the world’s largest providers of actuarial and related

products and services. The firm has consulting practices in life

insurance and financial services, property & casualty insurance,

healthcare, and employee benefits. Founded in 1947, Milliman is an

independent firm with offices in major cities around the globe.

milliman.com

CONTACT

Fritz S. Busch

fritz.busch@milliman.com

Peter Fielek

peter.fielek@milliman.com

does not certify the information, nor does it guarantee the accuracy and completeness of such information. Use of such information is voluntary and should not be relied upon

unless an independent review of its accuracy and completeness has been performed. Materials may not be reproduced without the express consent of Milliman.