health section news · as a result of the soa's governance audit we've been hearing about...

Issue No. 49 | March 2005

“For Pro fess iona l Recogn i t ion o f the Hea l th Actuary”

Many prescription drugs are cost-effec-tive treatment options.i With so manyprescription drugs available, which

ones should be covered and encouraged byhealth plans and which should not? It is theFood and Drug Administration’s (FDA) job toensure that only safe and effective pharmaceuti-cals are available in the United States. Given thatthis agency does its job well, why give the issueany further thought?

In the past, this line of thinking may have beenacceptable. Health plans could allow theirmembers access to whatever drugs wereprescribed by their physicians. The difficulty hascome with the explosion in cost in this area ofhealth care. While total health care expendituretrends have ranged from 9 percent to 16 percentover the past five years, pharmaceutical benefittrends have increased at rates of 17 percent to 18percent.ii These pharmaceutical cost trends havebeen attributed to increases in drug utilization(39 percent), increases in drug prices (37 percent),and shifts to higher priced drugs (24 percent).iii

The more recent innovations in biotechnologyhave helped to fuel this trend and seem poisedto continue to do so. The term “biotechnologydrug” refers to a pharmaceutical treatment withthree characteristics. First, the drug is derivedfrom particularly sophisticated technology.Second, these drugs require more complicatedadministration. They are injectable drugs, somerequiring physician administration. Third,because of the expensive development costs andthe additional administration costs, these drugsare very expensive, typically more than $1000per month per patient. For example, Xolair is abiotechnology drug for the treatment of asthmathat costs about $1000 per month. Another

example is Fabrazyme for Fabry’s Disease.While this condition is rare, the cost of the drugis about $250,000 per patient per year.

There are over 100 biotechnology drugscurrently available, and the drug pipeline prom-ises many more after 2006. These drugscurrently account for roughly 10 percent of thepharmacy budget.

Pharmacy Benefit Managementand the Use of DataTo manage this expense, health plans in recentyears have had to consider carefully which

The Formulary Decision Process:What Are They Doing in There andCan Actuaries Help?by Jill Van Den Bos, M.A., John Watkins, R.Ph., MPH, Kristin Reed, MPH, and Jonathan Shreve, FSA

Health Section News

(continued on page 20)

2 | MARCH 2005 | HEALTH SECTION NEWS

CONTENTSCOVER

1 The Formulary Decision Process: What AreThey Doing in There and Can Actuaries Help?

Jill Van Den Bos, John Watkins, Kristin Reed and Jonathan Shreve

Published by the Health Section Council of the Society of Actuaries

475 N. Martingale Road, Suite 600Schaumburg, IL 60173-2226

Phone: (847) 706-3500 • Fax: (847) 706-3599 •Web: www.soa.org

This newsletter is free to section members. A subscrip-tion is $15.00 for nonmembers. Current-year issues areavailable from the Communications Department. Backissues of section newsletters have been placed in theSOA library and on the SOA Web site: (www.soa.org).Photocopies of back issues may be requested for anominal fee.

Facts and opinions contained herein are the sole respon-sibility of the persons expressing them and should not beattributed to the Society of Actuaries, its committees, theHealth Section or the employers of the authors. We willpromptly correct errors brought to our attention.

2005 SECTION LEADERSHIPKarl G. Volkmar, ChairpersonLori A. Weyuker, Vice-ChairpersonWilliam R. Lane, Secretary/TreasurerMark E. Bill ingsley, Council MemberDamian A. Birnstihl, Council MemberCraig S. Kalman, Council MemberJohn C. Lloyd, Council MemberBryan F. Miller, Council MemberLisa F. Tourville, Council Member

Gail Lawrence, Newsletter EditorPHONE: (515) 224-4380E-MAIL: [email protected]

Susan Martz, Project Support SpecialistPHONE: (847) 706-3558FAX: (847) 273-8558E-MAIL: [email protected]

Clay Baznik, Publications DirectorPHONE: (847) 706-3548FAX: (847) 273-8548E-MAIL: [email protected]

Susie Ayala, Graphic DesignerPHONE: (847) 706-3548FAX: (847) 706-3599

Copyright © 2005, Society of Actuaries • All rights reserved. • Printed in the United States of America.

OPINION3 Chairperson’s Corner

Karl G. Volkmar

3 Book Review

William R. Lane

FEATURES

ANNOUNCEMENTS

4 What Will Be the Impact of the New Medicare Prescription Drug Benefit?

Corey N. Berger

8 Impact of the Medicare Prescription Drug,

Improvement and Modernization Act on

Medicare Supplement Insurance Plans

Dennis Hare

10 The Art and Science of Pricing Small Group

Medical Coverage

William R. Lane

14 The Bottom Line on Behavioral Health Care

Costs

Steve Melek

16 The Cost of Mental Health Parity

Steve Melek

13 The SOA Announces Co-Sponsorship of

the 2005 GUAA Annual Meeting

26 IAA Health Section Update

27 New Opportunities for Health Section

Participation

28 Overview of Sessions at the 2005 Health /

Pension Spring Meeting

28 Availability of the Disability Chart Book

The "section-year" runs from late October tolate October, beginning and ending duringthe annual meeting each fall. This is when

the Health Section Council (HSC) turns over—three of the nine HSC members officially rotate offthe HSC (but, thankfully, they usually still helpout!) and three newly elected members join thecrew. Please note, however, that this year is differ-ent than prior years—everyone is effectively newto the HSC this year. This is not due to a strike or amass exodus—six of the HSC are, in fact, return-ing—but due to significant changes beingimplemented in conjunction with this "changing ofthe guard." As you may know, the elected HSC hasserved section members for nearly 25 years. Theyhave served their members primarily by:

• Providing continuing education opportunities primarily through the Spring and Annual SOAmeetings;

• Providing communication and networking opportunities (e.g., Health Section News, meet-ing events, etc.); and,

• Soliciting, oversight and sponsorship of practicalshort-term research.

The Health Benefit Systems Practice AdvancementCommittee (HBSPAC) is an appointed committeeof the SOA that has served all practicing healthactuaries by (for example):

• Providing thought leadership to the SOA.

• Developing and maintaining external relation-ships with other professions, associations, etc.

• Soliciting, oversight and sponsorship of longer–term research.

As a result of the SOA's governance audit we'vebeen hearing about for some time, the decision wasmade to merge these two groups into one. Thesurviving group is called the Health SectionCouncil, but it really could be renamed. It will onlybe through the effective integration of the energy,

Chairperson’s Corner

by Karl G. Volkmar

Actuaries who work with medical insurance,either as plan actuaries who must price thecost of coverage, or as employee benefits

actuaries who must advise employers as to howwell disease management programs are working,will find the “Disease Management ProgramEvaluation Guide” a good working reference.

The Disease Management Association of America(DMAA) has compiled a lot of basic information onthe evaluation of disease management programs intoone relatively short book (less than 80 pages).

The book talks about the issues facing anyone whois attempting to evaluate disease management(DM) programs for cost effectiveness. It discussesthe most accurate ways to make such evaluationsand why these approaches tend to be very difficultto implement in the real world. 3 has a chartcovering a dozen ways to study cost effectivenessand gives some comparative thoughts on each,such as relative accuracy, relative ability to imple-ment in the real world, relative time frame toimplement and whether a control group is neededin order to use this approach.

The book touches briefly on a number of importantconsiderations such as timing, trend, appropriatemeasures, how to establish a “population,” causa-tion, regression to the mean and the general validityof the results. It also provides a checklist for evaluat-ing your own DM evaluation process. Since it is ashort book, it cannot cover these topics in depth andoften simply highlights the problem without lengthydiscussion as to how to solve the problem.

One aspect of the book that I found difficult wasthe heavy use of “insider” language. “Pre-post”designs may be familiar to DM specialists, but I hadsome difficulty getting up to speed with the lingo.To alleviate this problem, readers should consideralso acquiring the companion volume, “Dictionaryof Disease Management Terminology.”

I found the book well worth reading and wouldstrongly recommend it to anyone who needs tostudy the cost benefit of medical interventions. h

Book Review

Disease Management Program Evaluation Guide

by William R. Lane

HEALTH SECTION NEWS | MARCH 2005 | 3

Karl G. Volkmar, FSA,

MAAA, is a principal of

United Health Actuarial

Services, Inc. in

Carmel, Ind. He can

be reached at

kvolkmar@uhasinc.

com.


William R. Lane is

principal at Heartland

Actuarial Consulting,

LLC in Omaha, Neb.

He can reached at

(402) 778-0297 or

[email protected].

Introduction

On Dec. 8, 2003, President Bush signed intolaw the “Medicare Prescription Drug,Improvement and Modernization Act of

2003.” Two of the primary goals of this legislationwere to offer a prescription drug discount card toMedicare recipients starting in 2004 and, moreimportantly, an insured prescription drug benefit toMedicare recipients starting in 2006. These new bene-fits, along with the restrictions imposed on the abilityof Medigap policies to offer prescription drug cover-age to new enrollees starting in 2006, will have asignificant impact on prescription drug coverage forthe Medicare population and could have an impacton coverage for the commercial population. The totalannual cost to the federal government of this newprescription drug coverage was originally estimatedat $400 billion, but those estimates have increasedsince the legislation was passed. For purposes of thisarticle, our estimates are based on this $400 billionestimate. If the actual costs are higher, the distribu-tion among recipients would likely be the samewhile the dollar amounts would increase.

The impact of this legislation on the entire phar-maceutical distribution chain, includingpharmaceutical manufacturers, pharmacy benefitmanagers (PBMs), pharmacies, insurers, employersand insureds, is still unclear. For all of these enti-ties, a number of questions will need to beanswered in order to determine the impact:

• Who in the distribution chain will see the $400 billion estimated cost, and how much will eachentity see? How much of the $400 billion is tocover additional utilization created by the drugcard in 2004 and the insured benefit in 2006versus just shifting costs to the federal govern-ment (i.e., does the pie grow or just getreallocated)?

• How will the Medicare prescription drug plan affect the commercial market? Will pharmaciesreduce the discounts they offer to PBMs, insurersand employers in order to compensate for theloss of cash-paying customers, and will pharma-

ceutical manufacturers reduce their averagerebates per prescription?

• What will potential prescription drug plans (PDPs) offer as benefits? Among those MedicareAdvantage plans offering prescription drugplans (MA-PDs), what benefits will be offered?And will any offer supplemental benefits?

Who will see the $400 billionestimated cost?Many published news stories indicate that theHMO/insurance industry and the pharmaceuticalmanufacturers are the big winners from theMedicare prescription drug benefit. Many of thesenews stories are based on quotes from detractors ofthe bills. In reality, neither of these industries maycome out as huge winners from this legislation.

In fact, public sector and private sector employerswho currently provide prescription drug benefitsto their Medicare recipients appear to have theclearest-cut benefit. These employers will eitherreceive a direct tax-free subsidy from the federalgovernment for 28 percent of the gross cost ofprescription drugs between $250 and $5,000 start-ing in 2006, or they could eliminate theirprescription drug benefit and migrate theirMedicare recipients into a Medicare prescriptiondrug (Part D) plan. Original published estimatesfrom several sources including the government’sown estimates of the cost of this subsidy rangefrom $71 to $86 billion out of the $400 billion, orabout 20 percent of the total cost of the bill.

The impact on HMOs/insurers or any other entitythat wants to be a PDP is less obvious. Clearly theywill receive some of the $400 billion for providingadministrative services for the prescription drugbenefit. A rough estimate of the value of the admin-istrative services is about $50 billion. Unlike thesubsidies to the employers, however, this moneywill not go straight to the bottom line of PDPs.While some of this money could result in bottom-line profits, we would expect that a large percent ofthis revenue would cover real additional expenses.

What Will Be the Impact of theNew Medicare PrescriptionDrug Benefit?by Corey N. Berger, FSA, Senior Consultant, Reden & Anders, Ltd.


Corey N. Berger is a

consultant at Reden &

Anders Ltd in Duluth,

Ga. He can be reached

at (678) 417-4905 or

Corey.Berger@

reden-anders.com


WHAT IS THE IMPACT OF THE NEW MEDICARE PRESCRIPTION DRUG PLAN?

Who will see the rest of the money? Pharmaceuticalmanufacturers will likely see some, although theincrease in prescriptions filled (utilization) may beminimal. Studies have indicated that approxi-mately 75 percent of Medicare beneficiariescurrently have some level of coverage, either fromemployers, Medicare+Choice HMOs, Medigapplans, existing prescription drug plans, Medicaidor programs offered by the manufacturers. Thelevel of coverage for these individuals is not clearand varies significantly depending on the cover-age, but those with employer coverage, Medicaid,or some of the richer Medigap plans are likely toalready utilize prescription drugs at the same level(or potentially an even higher level) than theywould under the Medicare prescription drug plan,so they would be unlikely to increase their utiliza-tion. (In fact, if some of these people lose theircurrent prescription drug coverage and move intothe standard prescription drug plan, their utiliza-tion may actually decrease.)

In addition, analysis of Reden & Anders’ internalprescription drug databases indicates that onlyabout 10 percent of Medicare beneficiaries wouldhave costs in excess of $5,100 and only 33 percentwould have costs in excess of $2,250 in 2006. Theseare the two breakpoints in the formula forMedicare prescription benefits. Assuming a 10percent increase in utilization from the introduc-tion of the new coverage would mean additionalrevenue to the pharmaceutical manufacturers ofabout $50 billion in total from 2006 to 2013.Considering the industry had over $150 billion inrevenues from the United States in 2004, a $50billion increase in total from 2006 to 2013 wouldincrease revenue by only 2-4 percent over that timeperiod.

The remaining cost of the $400 billion, or about$200 billion (the largest part), would likely go toreimburse directly Medicare beneficiaries whocurrently pay for their drugs themselves by reduc-ing the out-of-pocket costs for those beneficiaries.

How will the Medicareprescription drug plan affect thecommercial market?As mentioned previously, statistics show that about75 percent of Medicare beneficiaries currently havesome level of prescription drug coverage. For amajority of Medicare beneficiaries, however, thediscounts available through that coverage are less

than the discounts currently available for commer-cial insureds. For example, for brand prescriptionsmost employers receive a discount of between 12percent and 15 percent, before factoring in rebates,based on a survey performed by Reden & Anders.Most Medicare recipients not covered under anemployer plan or a Medicare+Choice plan likelyreceive a discount of less than 10 percent on brandprescriptions, and those without any coveragemost likely receive a discount of 5 percent or less,based on Reden & Anders’ knowledge of thosemarkets. If all Medicare recipients moved to adiscount of 12 percent from their current estimateddiscounts, the reduction in revenue to the pharma-cies could be as much as $2 billion. The net incomeof Walgreen’s, CVS and Rite Aid combined for thetrailing 12 months (as of 12/10/2003) was $1.8billion based on filed financial statements.However, most Medicare members without insur-ance do not have the opportunity to receiveprescription drugs from other sources, such as mailorder. If these members were offered this optionand chose to get mail order prescriptions, thiswould impact the revenue generated from otheritems sold by the retail prescription drug stores. Theretail pharmacies will need to take this combinationof factors into consideration when negotiating theircontracts with the Part D carriers for Medicareinsureds, which may impact their commercialcontracts or result in entirely separate contracts forMedicare beneficiaries even though that would addadditional administrative complexity.


The same issue obviously applies to the pharma-ceutical manufacturers. They currently providerebates to the PBMs or employers based on marketshare and volume. These rebates are only paid,however, to entities that have a contract with themanufacturer. For Medicare beneficiaries that havelimited or no insurance for their prescription drugcoverage, the manufacturers do not pay a rebate.Once these beneficiaries start to enroll in aninsured plan, however, the PBMs or insurers arelikely to include these new prescriptions in thetotal volume they report to the manufacturers andrequest a rebate for these prescriptions.

The United States market for prescription drugsconsisted of approximately 3.2 billion prescriptionsin 2003 according to the NACDS Web site (accessedon 7/23/2004). Assuming that one third of thattotal is for Medicare, about 1.1 billion prescriptionsare filled for Medicare recipients. If 40 percent ofthose prescriptions are filled by people that do nothave any plan that would receive rebates, and 40percent of those individuals’ prescriptions wouldbe eligible for an average rebate of $2, about 18million prescriptions would receive rebatescurrently not paid. The additional rebates themanufacturers would have to pay would be $360million in current dollars, (which may be applica-ble under the prescription drug card) and likelyeven more in 2006. As with the pharmacies, themanufacturers may face the same issue regardingthe level of rebates they pay for Medicare benefici-aries compared to their current payments forcommercial members, especially if the inclusion ofMedicare members under insured coverage doesnot increase utilization.

If the pharmacies and manufacturers refuse toextend their existing levels of discounts and rebatesto new Medicare beneficiaries, the result will beeither a two-tier level of discounts and rebates (onefor commercial and one for Medicare) or new levelsfor discounts and rebates. Since most PBMs typi-cally have a global contract with each pharmacychain that covers all prescriptions filled by thatchain for all members covered by the PBM, devel-oping a split reimbursement schedule may be

difficult, or, at a minimum, undesirable.Alternatively, developing new contracts with lowerdiscounts and rebates that would maintain thecurrent profit levels for the pharmacies and totalpayout for the manufacturers would result in lowerdiscounts and rebates than what the commercialmarket currently receives. The result could be aspike in prescription drug costs for all entitiescurrently providing prescription drug coveragethat receive the higher discount over and above thetrends we already see for pharmacy.

What will PDPs offer for benefits,and will any offer supplementalbenefits?The legislation defines a standard prescriptiondrug benefit starting in 2006 as the following:

• From $0 to $250 in total costs (not including administration), the member pays 100 percent.

• From $250.01 to $2,250 in total costs (not includ-ing administration), the plan pays 75 percent of the cost and the member pays 25 percent.

• From $2,250.01 to $5,100 in total costs (not including administration), the member pays 100 percent.

• Above $5,100 in total costs (not including admin-istration), the member pays the greater of 5percent or $2 for generic or multi-sourceprescriptions and $5 for all other prescriptionsand the plan pays the balance. The $5,100 intotal costs is equal to an out-of-pocket expendi-ture for the member of $3,600.

The PDP must submit a bid for covering the cost ofthe standard prescription drug coverage (or anactuarially equivalent plan) that includes the costof administration. This bid is called the “directsubsidy” and covers the 75 percent of costsbetween $250 and $2,250 and approximately 15percent of the costs after the member ’s out-of-pocket expenses are greater than $3,600. Theremaining 80 percent of costs above the out-of-pocket maximum are reimbursed directly by thegovernment as part of a “reinsurance subsidy.”

The legislation then requires that an “adjustednational average monthly bid amount” be calcu-lated using all of the accepted bids. The “nationalaverage monthly premium” is then calculated as


WHAT IS THE IMPACT OF THE NEW MEDICARE PRESCRIPTION DRUG PLAN? | FROM PAGE 5

The United States market for prescriptiondrugs consisted of approximately 3.2 billionprescriptions in 2003


approximately 25.5 percent of the total of this“adjusted national average monthly bid amount”plus an average of the expected reinsurancesubsidy that all bidders must also include as part oftheir overall bid. Finally, the premium an individ-ual member must pay will be the “national averagemonthly premium” plus or minus the differencebetween the PDP bid and the “adjusted nationalaverage monthly bid amount” (i.e. if the PDP bid is$10 above the “adjusted national average monthlybid amount,” the entire additional $10 would becharged to the member.)

Since the “adjusted national average monthly bidamount” is not known when the bids are submit-ted, especially in 2006, there is clearly some risk insubmitting a bid. One likely result from thismandatory adjustment in member premium inorder to reflect the difference between the bid andthe “adjusted national average monthly bidamount” is that in the first couple of years in whichthe Medicare prescription drug plans are offered,bids will likely be made conservatively, unless aPDP was looking to enroll a vast majority of thebeneficiaries in their region or service area.

In addition to the basic prescription drug benefit,PDPs can offer supplemental coverage.Supplemental coverage is an enhancement to thebasic prescription drug benefit, and has some addi-tional aspects:

• The entire cost of the supplemental coverage must be paid for by the Medicare beneficiary.

• The supplemental coverage can take the form of a reduction in the deductible or coinsurance or an increase in the initial coverage threshold. There is no mention of a change in the annual out-of-pocket threshold and whether this can be reduced.

• The individual reinsurance provision that reim-burses the PDP for 80 percent of their costs oncea member hits the out-of-pocket maximumwould be worth LESS since it would requiremore total claims for an individual to hit the out-of-pocket maximum of $3,600 if the member ispaying less per prescription, on average, thanthey would under the basic benefit.

• The aggregate reinsurance provisions that apply when costs are in excess of 102.5 percent ofexpected costs do not apply to the portion ofcoverage that is for supplemental benefits.

• The cost for supplemental coverage can reflect an assumption for additional utilization due to selection.

These elements of the supplemental coverage maydiscourage plans from offering supplementalprescription drug coverage, even with the ability toadjust the bid to reflect additional utilization. Sincemembers will have to pay the full cost of the addi-tional benefit, and since prescription drug costs areamong the most predictable of all medical costs, theindividuals that will pay for the additional benefitsare extremely likely to use it, and highly likely touse more services than the excess premium wouldcover. The lack of any additional reinsuranceprotection for this adverse selection means that ifplans do offer supplemental coverage, they willlikely price it very conservatively since they are atrisk for all of the supplemental costs with minimalreinsurance from the federal government.

ConclusionsClearly, this new Medicare prescription drug bene-fit will impact the pharmaceutical industry, manyHMOs and insurers, employers sponsoring retireeprescription benefits and Medicare beneficiariesthemselves. What remains to be seen is who theultimate winners and losers will be, but identifyingsome likely repercussions and planning for themnow can provide you with a competitive advantageunder the new paradigm. 2006 will be here soon. Sothose who manage retiree prescription programsshould start the review and planning soon. h

WHAT IS THE IMPACT OF THE NEW MEDICARE PRESCRIPTION DRUG PLAN?


President Bush signed the MedicarePrescription Drug, Improvement andModernization Act (MMA) on Dec. 8, 2003.

Among the many requirements of MMA was theaddition of prescription drug benefits in the newMedicare Part D. Provision of drug benefits viaMedicare required the NAIC to make severalchanges to the Model Regulation to Implement theNAIC Medicare Supplement Insurance MinimumStandards Model Act, i.e. Model Regulation, inorder to conform to the federal law. These changeshad a direct impact on certain Medicare supple-ment, or Medigap, products. On Sept. 8, 2004 theNAIC adopted amendments to the MedicareSupplement Model Regulation in response to therequirements of MMA.

The amendments to the Medicare SupplementModel Regulation that implement the followingfour requirements of MMA are described below inmore detail:

1. Add two new plans (called K and L) to the stan-dard Medigap plans A through J;

2. Prohibit the sale of outpatient prescription drug coverage in Medigap plans after Dec. 31, 2005 (i.e., when Part D comes into effect);

3. Revise Medigap plans to eliminate outpatient prescription drug coverage for those who enroll in Medicare Part D;

4. Make any other changes to the model regulation that might be required as a result of the legislation.

New Plans K and LMedigap plans K and L have been added to theModel Regulation in response to the MMA require-ment that two new standardized Medigap benefitpackages be developed. These federally prescribedbenefit packages generally increase the cost sharingby the insured over that of plans A through J.

Coverage under Plan K includes 50 percent of thecost sharing otherwise applicable under MedicareParts A and B, except there is no coverage for thePart B deductible and any cost sharing otherwiseapplicable for preventive benefits is covered at 100percent. All hospital inpatient coinsurance and 365

extra lifetime days of coverage of inpatient hospitalservices are covered as in the current core benefitpackage. After the individual has reached theannual out-of-pocket limit, all cost sharing underMedicare Parts A and B is covered at 100 percentfor the balance of the calendar year. The annualout-of-pocket limit is $4,000 for 2006, and isindexed for inflation in future years.

Benefits under Plan L are identical to Plan K exceptthat 75 percent, rather than 50 percent, of cost shar-ing for Parts A and B is covered until the annualout-of-pocket limit is reached, and the initial out-of-pocket maximum in 2006 is $2,000 rather than$4,000.

The outlines of coverage illustrated in the ModelRegulation have been revised to reflect these twonew plans.

Ban on Future Sale of OutpatientPrescription Drug CoverageAfter Dec. 31, 2005, companies are no longerallowed to issue Medigap plans that provideoutpatient prescription drug coverage. BeginningJan. 1, 2006, companies will only be allowed toissue plans H, I and J or related plans in waiverstates if they have been modified to eliminateprescription drug benefits and the premiums havebeen adjusted accordingly.

Plan Revisions for Policies WithDrug CoverageOn Jan. 1, 2006 the rules for renewal of Medigappolicies providing outpatient prescription drugcoverage change. Renewal options for standard-ized and pre-standardized plans with drug benefitsthat were sold prior to Jan. 1, 2006 depend uponwhether the insured has elected to enroll inMedicare Part D.

If the insured does not enroll in Part D, the policycan be renewed at the option of the insured with-out any modification. However, there are twooptions available to insureds if they elect to enrollin Part D during the initial enrollment period thatends May 15, 2006. The insured may choose tocontinue their current policy, but the insurer must

Impact of the MedicarePrescription Drug, Improvementand Modernization Act on MedicareSupplement Insurance Plansby Dennis Hare

Dennis K. Hare is an

actuary with GE

Insurance Solutions

in Overland Park, Kan.

He can be reached at

(913) 676-5281 or

[email protected].


eliminate the drug benefits and appropriatelyadjust the premium, or the insured may canceltheir current policy and purchase under a guaran-teed issue provision a new standardized plan A, B,C, F, K or L offered for sale by their current insurer.This guaranteed issue offer begins on the date theindividual receives notice from the carrier duringthe 60-day period immediately preceding the initialPart D enrollment period and ends 63 days afterthe effective date of the insured’s coverage underPart D. If the insured elects to enroll in Part D afterthe initial enrollment period, they may continuetheir current policy without drug benefits, but theydo not have the guaranteed issue option to switchto another plan offered by the insurer.

To assist insurers with the premium modificationfor these plans, the NAIC Accident and HealthWorking Group of the Life and Health ActuarialTask Force is currently working on a set of princi-ples to guide carriers when Medigap plans aremodified to remove drug benefits. Proposedlanguage for these principles is expected to befinished in the first quarter of 2005, and amend-ments to the Model Regulation or the NAICMedicare Supplement Insurance Model RegulationCompliance Manual will follow.

The Working Group previously consideredwhether changes to the premium refund provisionof the Model Regulation are necessitated by MMA.A single change was made to the benchmark lossratio formula used in the premium refund calcula-tion. The relatively minor change expands theformula to include use of earned premiums for thelifetime of the policy rather than stopping after thefirst 15 years as required in the previous formula.During discussions of other possible changes theWorking Group agreed that for purposes of thepremium refund calculation, plans where drugcoverage has been eliminated should be combinedwith like plans that continue to provide outpatientprescription drug coverage.

Another amendment requires an issuer to file anyriders or amendments to the policy or certificateforms used to delete outpatient prescription drugbenefits only with the commissioner in the state inwhich the policy or certificate was issued.

The discontinuance of outpatient prescription drugbenefits is reflected in the revised outlines of cover-age illustrated in the Model Regulation.

Other Changes to the ModelRegulationAll Medigap carriers are required to provide notifi-cation regarding the insured’s options and rights toeach individual insured with a Medigap plan thatcovers outpatient prescription drugs during the 60-day period immediately preceding the initial PartD enrollment period. Generally, the initial enroll-ment period for Part D is the same as the initialenrollment period established for Part B. However,for those individuals already eligible to enroll inPart D as of November 15, 2005, the initial enroll-ment period for Part D begins on November 15,2005. The Secretary of Health and Human Servicesis working in consultation with the NAIC todevelop the required notification.

Certain provisions of the Model Regulation requirereinstitution of previous Medigap coverage thatwas suspended at the option of the policyholderfor specified reasons. Clarification was added tothe Model Regulation stating that if the suspendedMedigap policy provided coverage for outpatientprescription drugs and the insured enrolled inMedicare Part D while their policy was suspended,the reinstituted policy will not have outpatientprescription drug coverage, but will otherwise besubstantially equivalent to the coverage before thedate of the suspension.

The required eligibility questions included on theMedigap application and the text of the notice tothe applicant regarding replacement of Medicaresupplement insurance have been revised toimprove clarity and understanding.

The changes to the Medicare Supplement ModelRegulation that were adopted by the NAIC reflectonly the unambiguous provisions of MMA thatdirectly relate to Medigap plans, with some minorexceptions for clarification purposes. This articlehas highlighted the majority of the amendments tothe Model Regulation, but is not an exhaustivedescription. To review all changes to the MedicareSupplement Model Regulation visit www.NAIC.orgto obtain a copy of the adopted version of themodel. In addition to the changes specified inMMA, the NAIC was directed to evaluate the bene-fits provided by the standardized plans andrecommend whether the number of plans availableshould be reduced. The NAIC Senior Issues TaskForce has begun discussing this charge, but at thistime it is difficult to predict what recommendationmight be made. h

MEDICARE SUPPLEMENT INSURANCE PLANS


On the surface, it might appear that there areas many ways to price medical insuranceas there are carriers that underwrite it. In

reality, most carriers use one of two mainapproaches in pricing group medical insurance.These two approaches can go by many names (orno name at all). I refer to them as the “Forecast”approach and the “Rebuilding” approach.

At its core, the Forecast approach works by takinghistorical earned premium and historical incurredclaims and projecting them into the rating period.The projected loss ratio is compared to the desiredloss ratio and the rate action is the increase incurrent premium needed to make the loss ratiosidentical. I refer to this approach as Forecastingbecause it typically explicitly develops an expectedloss ratio for the future rating period in the absenceof a rate action.

In its most simple form, the approach simplycompares the current loss ratio to the desired lossratio and combines the needed corrective actionwith trend.

In other cases, some carriers will develop permember per month (PMPM) incurred claims andPMPM earned premiums by dividing incurredclaims and earned premium by a total membermonth count. They then compare the PMPM claimsto the PMPM premium. Mathematically, when youdivide the PMPM claims by the PMPM premium,you cancel out the member month count. Hence,this approach is still what I refer to as the Forecastapproach.

At its core, the Rebuilding approach works bysplitting the historical incurred claims into variouscomponents usually based on the type of service,supply or additional benefit, and then dividingthese amounts by the number of member monthswhich were included in the historical period. ThesePMPM amounts are adjusted for trend and some-times other factors. The trended amounts thenbecome the basis for the next time period’s pricing.I refer to this approach as Rebuilding because ittypically produces the new rates by “rebuilding”the base rates by type or benefit.

Typically in a rebuilding approach, the cost foreach service, supply or benefit is at 100 percentcoverage without reduction for copays or othercost sharing. This is different from the typicalForecast approach that usually looks at actualincurred claims after cost sharing and compares itto earned premium that has been adjusted for theplan of benefits.

The key difference between the two approaches isthat the Forecast approach incorporates the variousrisk factors by using actual premium that shouldalready have these factors built into it. TheRebuilding approach must acquire the ratingfactors separately because they are not inherent in anormal member month count.

Neither approach is universally better than theother.

The Forecast ApproachThe Forecast approach often uses only financialstatement data or the equivalent. Since this infor-mation is required for other purposes and isheavily audited, the data for pricing is relativelyeasy to acquire and accurate. The calculations arerelatively easy as well since the approach oftenuses a large block of business as a whole. It is alsoeasy to explain. For example, a typical explanationmight be that the loss ratio for the last year came intwo points higher than expected, so the rate actionis trend plus two points.

One common reason for using the Forecastapproach is that the data to perform a Rebuildingapproach is simply not available or is not deemedto be sufficiently accurate.

The Forecast approach has a number of drawbacksas well. It assumes that various rating factors suchas age gender slopes are correct. It also assumesthat if the age gender of the underlying block ischanging, then these factors, as used in settingactual premiums, will compensate in an appropri-ate manner. Since not all rating factors can be usedto the extent of their actual values, this is not acorrect assumption. It is often, but not always,“close enough.”

The Art and Science Of Pricing SmallGroup Medical CoverageTwo General Approaches To Pricingby William R. Lane

William R. Lane is

principal at Heartland

Actuarial Consulting,

LLC in Omaha, Neb.

He can reached at

(402) 778-0297 or

[email protected].


To illustrate what can occur, consider that smallergroups, particularly one employee groups and twoemployee groups, tend to have significantly highermorbidity than larger groups even after healthstatus is considered. If the carrier has no size factoror a size factor that is restricted by law to less thanthe real change in risk, then the carrier is at risk fora change in the average size of its groups. If theaverage group gets smaller in the future, then rateswill be inadequate and vice versa.

Another key difficulty in the Forecast approach istaking prior rate actions and benefit changes intoconsideration. Ideally, you would want the earnedpremium to be based on the same rate basisthroughout the experience period. In reality, this isoften not the case. One relatively simple approachis to split the experience into each renewal month.Generally speaking, this keeps the same base ratesfor all cases and allows for easier consideration ofprior rate actions.

If plan factors are considered to be reasonable anddo not need significant change, there might not beany reason to adjust for plan changes. If they areknown to be inconsistent, then adjusting for thembecomes a challenge. This is quite importantbecause, in the real world, underpriced benefitswill grow as a percentage of the total block andvice versa.

Other factors that can change for the whole blockinclude anti-selection (either a stagnant block thatis increasingly anti-select or an increasingly selectblock with growing new sales), the discount fornegotiated networks, plan factors (which changewith inflation) and shifts within area if the areafactors do not compensate. Since most states do notallow the full range of risk factors or the full annualchange in risk factors, the average risk factorallowed by law may not match the average riskfactor of the block. This also needs an adjustment.

Yet another drawback to the Forecast method isthat trend will need to be leveraged by thedeductible levels of the various benefit designs.This is particularly important in pricing the drugcard benefit where fixed dollar copays can be asizeable percentage of the average cost perprescription.

The Forecast approach generally requires thatspecial studies need to be made to determine if theunderlying rating factors are appropriate or needto be changed. A review of general factor appropri-

ateness can be built into the rating approach, but itisn’t easy to do. One “special study” that is almostalways required is a continuation table whereclaimants are sorted by the size of their annual cost(before deductibles and coinsurance). Othersinclude splitting the benefits by type of service.Special studies also have the drawback that theyfrequently take place during times of lower activityand the data being used is not the same data thatwas used in re-rating the entire block. It is morerecent if nothing else. This can lead to mismatches.

The Rebuilding ApproachThe Rebuilding approach requires significantlymore data. Not only do you need claims andmember month counts, but you also need ratingfactors at the group, subscriber or member level.Member months are generally not audited likefinancial data, particularly outside of the HMOenvironment. In addition, rating factors can bedifficult to obtain. For example, a computer systemmight contain the zip code of the member, but ifthe group is priced according to the zip code of thegroup itself, then the appropriate area factor foreach member must be found in a group level datarecord. Values that are paid by the carrier orcollected by the carrier tend to be heavily audited.Factors that are entered into a system, but not usedin any payment, tend to have more uncorrecteddata entry errors.

Another issue with rating factors is that they needto be as of the date the rates were developed. Thus,unless all cases are re-rated as of issue and changedto the demographics and other rating factors as ofthat date, the computer system ideally wouldcapture the rating factors (and the data item thatcorrelated to them) when the rates are run for thefinal time. Some factors such as group size canchange rapidly, and the changes are not alwaysrandom. For example, it is not unusual for groupsto be larger when rates are initially requested andthen they “shrink” at issue or soon thereafter. Thus,


PRICING SMALL GROUP MEDICAL COVERAGE

To illustrate what can occur, consider thatsmaller groups, particularly one employeegroups and two employee groups, tend to have significantly higher morbidity than largergroups even after health status is considered.

the group size factor used for rating the actual casemight be set at a level appropriate only for a largergroup. Hence, you can’t accurately recreate actualpremium levels using the current size of thesegroups.

One method for testing the validity of membermonth counts and rating factors is the process ofrecalculating what the actual premium should befor a block of business. Typically, the estimatedtotal premium and the actual earned premium willvary by more than an immaterial amount. Thereare many reasons why this is so, but essentiallythey all go back to missing data or inaccurate data.In some cases, however, the difference remains areasonably constant percentage difference. If this isthe case, you might be willing to simply assumethis factor will remain constant in the future ratherthan go to the effort of correcting all rating factors.

There are several advantages to the Rebuildingapproach. It tends to better fit the HMO benefitstructure where benefits are paid at 100 percent lessa copay. Splitting the claims into these benefitamounts and knowing the number of claimantsallows for an easy calculation of plan factors.Negotiated network reimbursements tend to beeasier to apply with the Rebuilding approach.Having the data by benefit generally allows for aneasier time in modeling the impact of changes in anegotiated arrangement.

Historically, HMOs have viewed both theirpremium and their claims on a PMPM basis. Often,senior management does not adjust these amountsfor significant rating factor changes. Thus, seniormanagement might believe that premium is satis-factory simply because the PMPM premium hasreached a specified target. If you are using theRebuilding approach and have the rating factorsavailable, you may be able to spot that the agegender factor has increased due to a lower percent-age of children. In reality, the HMO might not beachieving the premium results that it originallybudgeted. Simply because it does tend to “fit” anHMO mentality, most HMOs use some form of theRebuilding approach.

Generally, there is no need to be concerned withprior rate actions when using the Rebuildingapproach except when attempting to recreatehistorical premium.

If you have captured all of the rating factors bymember, then performing special studies to reviewthe rating factors is much easier. It is also more

likely to be more accurate since the data for thefactor study is identical to the data used in re-rating the entire block.

The Rebuilding approach also has significant draw-backs. The most serious is usually the accuracy ofthe rating data. It is often inaccurate to a degreeand the inaccuracy is not necessarily self-correcting.For example, suppose the computer system withmember data shows whether the “subscriber” hasemployee only coverage, employee plus spousecoverage, employee plus children coverage or fullfamily coverage. Now, also suppose that the carrierallows dependents to have coverage without amember under specific situations. Unless themember record clearly shows that no employee iscovered under these “subscribers,” the assumedexposures will include employees that do not exist.This will lead to a lower than actual historicalPMPM claim value. The following year, the calcula-tion will continue to be lower than actual.

Such problems with data occur in almost allsystems and are intensified by the number of sepa-rate computer systems that carriers use. Data maybe accurate for one system, but not another.

Another drawback to the Rebuilding approachmay be that it is more difficult to explain. Trendedclaims divided by adjusted exposures are not whata typical marketing officer thinks about.

Yet another drawback to the rebuilding approach isthat by building up a number of pricing piecesfrom scratch, it becomes more difficult to applyreasonableness tests. The hospital cost may haverisen significantly, but the physician costs appear tobe reduced. Should the physician costs be used asis or should you assume they actually increased aswell? If the local hospitals have started includingradiology, pathology, anesthesiology and emer-gency physician costs in their charges, thecombination might be correct. When everything is“thrown together” under a typical Forecastapproach, the increase in the total is more likely tobe what you expect. The finer you split the benefitsunder a Rebuilding approach, the more difficult itis to apply reasonableness tests.

The most serious drawback to the Rebuildingapproach is the data accuracy issue. The Forecastmethod uses premiums and claims that can usuallybe compared to readily available financial state-ment values. There are no comparable figuresagainst which the Rebuilding approach values canbe compared. Unless the rating values can be used


PRICING SMALL GROUP MEDICAL COVERAGE | FROM PAGE 11


to recalculate historical premium, there is nosimple method of assuring their overall accuracy.

The Best ApproachWhat then is the “best practice”? Assumingresources can be made available on a cost-effectivebasis, the best approach is to do both. If the sameassumptions are applied to both methods, theyshould result in the same answer. If they do not,something is wrong and it is worthwhile to findout the differences. If nothing else, using thedetailed data from the Rebuilding process to calcu-late historical earned premium and then comparingthis figure to actual earned premium provides anextremely valuable cross check.

In the real world of scarce resources, a number ofconsiderations need to be made in selecting arating approach. Computer resources are a criticalconsideration. The sheer volume of data requiredby a Rebuilding approach is a significant drawbackunless the systems already exist.

The size of the block is also a consideration.Basically speaking, the smaller the block, the lesscredible the experience. Hence, for a smaller block,a Forecast method (which lumps all experiencetogether) tends to be better simply because thevalues produced by a Rebuilding approach are notcredible.

If you are starting a new block of business, youhave no historical experience and will need to baseyour rates off of whatever information seems mostappropriate (probably information purchased froma consultant).

In either approach there are issues that should beconsidered. For example, extremely large claimswill distort the results. If you know the numberand size of the large claims in a block, the Forecastapproach tends to be easier to adjust for an abnor-mally high or low number of large claims. If youhave frequent changes in negotiated networkarrangements, the Rebuilding approach tends tomake it easier to implement these changes in pric-ing. Bonus payments to provider groups tend to behandled easier with the Forecasting approach sincethey are thrown in with all other claims.Distributing bonus payments, after the fact, tendsto complicate the Rebuilding approach.

Whichever approach is used, a valuable cross checkis to use the re-rating information to forecast theexpected experience of the new and renewal blockin the following time period. This should then becompared with the actual experience on an ongo-ing basis. Any significant differences are a justcause for further research and possibly futurerefinements in the rating process. h

PRICING SMALL GROUP MEDICAL COVERAGE

The Society of ActuariesAnnounces Co-Sponsorship ofthe 2005 GUAA Annual MeetingMay 22-25, 2005

Grand Hyatt San Francisco on Union Square

Specific sessions will address actuarial and underwriting issues, such as pricing, product design andindustry experience. A major focus will be on underwriting issues and the integration with the actu-ary’s work. Participants will have the opportunity to see and discuss first hand what happens toactuarial theory when it is practically applied by underwriters in an environment influenced bymarket pressures. Separate tracks will address group life, long- and short-term disability, medical,dental and reinsurance.

Don't miss this great opportunity to learn from and network with our underwriting colleagues.

Please refer to the SOA (www.soa.org) or GUAA (www.guaa.com) Web sites for fee information andsession descriptions. h

Steve P. Melek is a

consulting actuary at

Milliman Inc. in Denver,

Colo. He can be

reached at (303) 299-

9400 or steve.melek

@milliman.com

What are the real costs of behavioral healthcare today? To answer this question mostpayers look at their behavioral health

carve-out spending and claim that costs haveremained level over the past decade or perhapseven decreased. This is a result of the excellent jobthat managed behavioral health care companieshave done in managing specialty behavioral healthcare over this period. However, by only examiningcarve-out spending, payers are overlooking signifi-cant mental health expenditures from treatment byPrimary Care Physicians (PCPs). Under our currenthealth care system, PCP involvement in behavioralhealth care delivery may lead to less efficient andeffective treatment and is an increasing cost thatmany plans haven’t yet begun to address.

If one only considers the carve-out piece whenanalyzing mental health care spending, costs haveindeed remained fairly level. However, thisapproach disregards the portion of mental healthcare that comes from primary care physicians.PCPs are the sole deliverers of about 50 percent ofmental health care, and also prescribe two-thirds ormore of all psychopharmacological drugs. Thesecosts are not included in mental health carve-outexpenditures. Considering these facts, it becomesobvious that mental healthcosts are increasing. The distri-bution of private insurancemental health expendituresover the last decade shows asteady decrease in facility andprofessional costs and anincrease in prescription drugcosts. Together these result in anet increase of approximately 7percent annually. These costsexclude the additional costs oftreatment by PCPs for physicalsymptoms associated withmental health conditions.

The high contribution of PCPcosts to total mental healthexpenditures reflects a lack ofpublic understanding aboutbehavioral health care and is a

serious obstacle to delivering it effectively. In manycases, failure to recognize a mental condition leadsindividuals to seek treatment from their PCP. Thisis in large part due to most people’s inability todifferentiate between the symptoms of a mentalhealth condition and a physical illness. Mentalillness often manifests itself through physicalconditions such as headache, backache, nausea,fatigue, or even chest pain. In fact, there is a posi-tive correlation between the number of physicalsymptoms and the prevalence of mood and anxietydisorders. As the number of physical symptomsincreases, the more likely a mood or anxiety disor-der exists.

In fact, only a small percentage of the populationwith mental disorders or emotional distress willever see a mental health professional and manywill remain untreated. Because a patient generallyspends only 12 to 16 minutes with their PCP, littletime is available to adequately assess multipleand/or vague symptoms, make a diagnosis, anddevelop a treatment plan. Physicians may be ableto treat symptoms in such an environment.However, the underlying mental condition oftenremains undiagnosed. In these cases the individualmay experience temporary symptomatic relief, but


The Bottom Line on BehavioralHealth Care Costsby Steve Melek

Distribution of Private Insurance Mental Health

Expenditures - 1991 through 2001

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

Calendar Year

Facility Professional Rx

Source: National Estimates for Mental Health and Substance Abuse Treatment, 1991-

2001, USDHHS, SAMHSA, March 2004


will often return with different or more persistentproblems. When referrals to behavioral health carespecialists are made, 50-90 percent of patients arenoncompliant and never see the specialist.Ultimately, PCP utilization becomes higher thannecessary and the patient is inadequately treated.

When the PCP does diagnose a mental condition,they may prescribe a psychopharmacologic agent.Sometimes patients will even request such medica-tion on their own. However, these medicationsrequire close monitoring and supervision to deter-mine proper dosage and reduce side effects. PCPssee over 30 patients a day and lack the necessarytime to educate and counsel patients about thesedrugs. Thus, the agents are often used improperlyand are less effective in addressing the patient’scondition.

I believe that a substantial opportunity for savingsin health care exists from more effective and timelytreatment of mental health disorders. After audit-ing total mental health care spending to identifyareas of wasted spending, there are generally threeareas of focus to improve quality and value frommental health care: benefit design, PCP supportand patient education.

For example, changes in bene-fit design to improve qualitymay include removal of barri-ers, such as higher copays tovisit behavioral health special-ists, which discourage patientsfrom seeking proper care fortheir conditions. The use ofpharmacy benefit managers tohelp monitor drug utilizationand patient compliance is anexample of PCP support thatmay help improve mentalhealth care provided in theprimary care office. Finally,disease management programsthat monitor patients andprovide 24-hour call-in linesmay aid in providing patienteducation about their mentaldisorder and treatment plan.

This is particularly important for mental illnesseswhere medications take much longer to becomenoticeably active and have the challenge of moreimmediate side effects.

The bottom line is that, like other areas of healthcare, mental health care costs, when considered intheir entirety, are also increasing rapidly. We needto be continually exploring new methods to reduceand manage them. By focusing on systems that getpatients the right mental health care at the righttime, improving patient adherence to drug treat-ment regimens, and aligning incentives in physicaland behavioral health care benefit plans and deliv-ery systems, not only can health care be greatlyimproved, but the bottom line can be as well. h

2001

the Prevalence of Mood and Anxiety Disorders

Relationship Between Physical Symptoms and

0%

10%

20%

30%

40%

50%

60%

70%

0-1 2-3 4-5 6-8 9 or more

Number of Physical Symptoms

Prevalence of Mood Disorder Prevalence of Anxiety Disorder

Source: Kronke et.al. Physical Symptoms in Primary Care; Predictors of Psychiatric Disorders and

Functional Impairment, Archives of Family Medicine. 1994, 774-779.

BEHAVIORAL HEALTH CARE COSTS

The Mental Health Parity Act of 1996 requiredthat the annual and lifetime dollar limits ofmental health benefits and medical benefits

be equal for employers with at least 50 employeesoffering mental health coverage. Since its imple-mentation, new federal proposals have beenpresented that would extend the 1996 Act, somerequiring full parity for all categories of mentalhealth conditions as listed in the DSM-IV (theDiagnostic and Statistical Manual of MentalDisorders). Opponents of such legislation arguethat the combined pressures of general costincreases and a need to pay fully for mental healthcare will make it impossible for employers tocontinue offering affordable coverage, often citinginitial estimates that placed resulting premiumincreases from full parity between 3.2 percent and8.7 percent. However, as actual experience hasemerged, it has become clear that these estimateswere conservatively high. In fact, with implementa-tion of mental health parity at the same time asmanaged behavioral health care, many states havediscovered that overall health care costs increasedminimally and in some cases were even reduced.

The three primary drivers of cost increases frommental health parity legislation have been identi-fied by both sides of the issue, and include:

• The levels of mental health benefits already existing, including calendar year benefit limits

and levels of insured coinsurance, copayments and deductibles.

• The degree of utilization management that existed or that would be implemented with parity.

• The degree of shift in services from the public sector to the private sector after parity.

While parity does not require mental health cover-age to be offered by employers, when coverage isprovided it may not be limited more than medicalcoverage. Thus, the impact of this legislation willbe minimal if benefits offered under the currentplan are similar in richness to mandated benefits.In the case that mandated benefits are significantlyricher, utilization will likely increase. However,implementing managed care for behavioral healthcare may limit the effects.

As debate over the federal legislation continues, 35states have enacted their own versions of mentalhealth parity laws. The emerging results of theirprograms dispel the cost arguments of parity crit-ics. These states are finding cost increases of lessthan 2 percent and in some cases cost decreases ofup to 50 percent, depending on whether mentalhealth care management was already in place. Thefollowing table summarizes the results from vari-ous state parity programs.


The Costs of Mental HealthParityby Steve Melek

IMPACT OF STATE PARITY PROGRAMS

State Parity Type Managed Care Change Cost impact

North Carolina, 1991 Full Parity for Implemented at Mental Health Costs State Employees time of parity changed from 6.4%

of total health costs to 3.1% in 6 years

Texas, 1991 SMI type for State Implemented at 48% decrease in the Employees time of parity cost of behavioral health

care in managed care plans

Minnesota, 1995 Full Parity No change $0.26 pmpm increase for 1 large plan; 1-2% increase

for state employees

Maryland, 1994 Full Parity No change 0.6% increase in health care costs

Rhode Island, 1994 SMI Parity No change 0.33% increase in health care costs


SMI parity designates parity for severe mentalillnesses only (such as schizophrenia, schizoaffec-tive disorder, bipolar affective disorder, majordepressive disorder, specific obsessive-compulsivedisorder and panic disorder) as defined within thelegislation.

There is also evidence, besides the emerging expe-rience of these states, that the initial costprojections for mental health parity programs weretoo high. Industry experts have made more recentprojections based on current data, the most tellingof which are highlighted below.

• In March of 1998, RAND published a case study of the Ohio State Employee Program’sexperience for mental health and substanceabuse parity. The main result of the study wasthat costs for behavioral health care remainedlow and even declined under managed care.According to the authors, “the implementationof managed care by far overwhelmed the effectof benefit expansion.”

• In October 1999, RAND provided testimony to the U.S. House of RepresentativesSubcommitee on Criminal Justice, Drug Policyand Human Resources that the additional costof adding full parity for substance abuse bene-fits to a plan that previously had provided nosubstance abuse benefits is in the order of 0.3percent for HMOs.

• In June 2000, the National Advisory Mental Health Council (NAMHC) updated their 1998estimate (ranged from 1 percent to 4 percent byplan type) for the cost of mental health parityto an aggregate increase of 1.4 percent, basedon an evolution of assumptions in their modeland new data. In a report to Congress entitled“Parity in Coverage of Mental Health Servicesin an Era of Managed Care,” the NAMHCfound that “based on empirical studies andeconomic stimulations across diverse popula-tions, managed care approaches and paritystructures suggest that the introduction ofparity in combination with managed careresults in lowered costs and lowered premiums(or, at most, very modest cost increases) withinthe first year of parity.” They also included“these findings do not support earlier concernabout potentially high financial costs causedby parity.”

In 2000, PricewaterhouseCoopers produced amental health parity report for the AmericanPsychological Association. They reported that “todate, there are no examples where mental healthparity has been enacted in a state and costs havedramatically increased,” and that there “are noexamples where mental health parity has beenenacted in a state and a measurable increase inuninsured has been detected.”

• The Office of Personnel Management (OPM) is responsible for implementation of the 2001

Steve P. Melek is a

consulting actuary at

Milliman Inc in Denver,

Colo. He can be


9400 or steve.melek

@milliman.com

New Hampshire, 1994 SMI Parity No change 1.5% increase projected; actual increases in health

care costs less than that or even flat

Maine, 1995 SMI Parity No change Behavioral health care costs as a % of all health care costs

changed from 4.66% to just 4.67% of total

Colorado, 1997 SMI Parity No change Increase in total health care costs of 0.2%

Vermont, 1997 Full Parity No change BCBS Plan found that behavioral health care costs rose from 2.30% to 2.47%

of all health care costs

IMPACT OF STATE PARITY PROGRAMS (Cont’d)

State Parity Type Managed Care Change Cost impact

MENTAL HEALTH PARITY

parity coverage for the Federal EmployeeHealth Benefit Plan (FEHBP). They expectedincreases in total health care benefit costs of 0.2percent to 0.5 percent due to parity.

• In July 2001, RAND provided additional testi-mony that “parity in employer-sponsoredhealth plans is not very costly under compre-hensively managed care, which is the standardarrangement in today’s marketplace. The totalcost of providing parity-level benefits is lessthan the increase of benefit expansion claimedby recent actuarial studies.”

• In August 2001, PWC projected that the Mental Health Equitable Treatment Act of 2001 wouldcost employers 1 percent or $1.32 per enrolleeper month. Parity in this Act is required for in-network services only, where providers havetypically agreed to discount their fees. TheCBO estimated that this Act would raise healthinsurance premiums by 0.9 percent.

• In February 2002, Mathematica submitted a report on the California Mental Health ParityLaw to the California HealthCare Foundation.The California bill, effective in July 2000,included SMI and SED (serious emotionaldisturbances in children). Mathematica found“the law did not appear to have had anyadverse consequences on the health insurancemarket to date, such as large increases inpremiums or decreases in health insuranceofferings by employers. Although employersfaced premium increases of 10 percent to 20percent in 2001, little of the increase was attrib-uted to parity.”

• In 2004, the CBO modified their estimate of the expected cost impact of national mental healthparity to 0.8 percent of total health care costs(down from their prior estimate of 0.9 percent).This reflects the aggregate expected impact onall states given the current status on mentalhealth parity by state.

The combination of actual state-specific experienceunder various parity programs with the reviseddownward projections of several key organizationsnarrow the expected cost impact of national mentalhealth parity legislation to a reasonable range. Thebottom line is that evidence now exists supportingthe argument that mental health parity laws havevery little impact on the overall health care costs.Offering mental health benefits at the same level asmedical benefits may be an efficient, affordableway to improve the quality of the insureds’ livesand protect them from catastrophe.

It should be noted that mental health benefits inhealth insurance policies typically include servicesprovided by specialty mental health providers suchas psychiatrists, psychologists, masters-level socialworkers, and other approved mental healthspecialists. Services provided by primary carephysicians and psychotropic drugs are consideredto be medical benefits and are not restricted bylimited mental health benefits. The use of these andother medical services to treat behavioral healthconditions have soared in recent years. This isdiscussed in greater detail in the preceding articlein this edition, “The Bottom Line on BehavioralHealth-Care Costs.” h


MENTAL HEALTH PARITY | FROM PAGE 17


talent and resources currently in place in bothgroups that Health Section members and all prac-ticing health actuaries will benefit in the mannerintended by this change in structure. The bottomline is this—whatever structure we use is nothingwithout the great volunteer support we have—theSOA is just trying to organize that volunteer baseso that it is utilized in the most efficient mannerpossible.

What does this mean? HSC members will serve ascoordinators for the Health Section's role in thefollowing areas of activity:

• Secretary/Treasurer (Bill Lane)

• Communications and Publications (Lisa Tourville)

• Basic Education (Damian Birnstihl)

• Continuing Education (John Lloyd, Craig Kalman, Lori Weyuker)

• Research (Bryan Miller)

• Professional Community (Mark Billingsley)

The HSC as a whole will be engaged in otherimportant activities, such as identifying key issuesfacing health actuaries in their current professionalenvironment. Others (invited advisors, liaisons,friends, etc.) may support the HSC in these activi-ties by providing additional perspectives andvolunteer muscle. (To volunteer, see new opportu-nities for the health section participation on page27.)

Please note that Lori Weyuker is also serving asvice chair of the HSC this year.

At this point, the plan is that these coordinatorswill rotate every year to ensure that new perspec-tives are introduced on a regular basis. Alonger-term underlying volunteer structure willensure that the necessary continuity will be in

place. Thanks to all of these HSC members as theywork to support all of us in their respective areas.

I could say a lot more about all of the changesunderway and all of the work being done tosupport the section membership, but I'll save it fora later time!

Two things I'd like to ask of the Health Sectionmembership over the coming year:

• Please be patient with us as we work through this transition. We're trying to provide higherlevels of service, and I believe we ultimatelywill; however, there may be some "bumps in theroad" in the short term. Please contact me if youthink anything important to you might be slip-ping through the cracks.

• Please volunteer!!! There are many areas where more volunteers are needed. Are you interestedin information regarding opportunities for serv-ice? Please contact myself or Kara Clark of theSOA.

Final thoughts:

• Many thanks to HBSPAC members, and we hope you'll continue to serve the SOA/Section membership as you have been.

• Thanks to all of the SOA staff for their support, but especially Sue Martz for keeping us organ-ized, and Kara Clark—without her, much of what you see wouldn't happen (or at least not during my tenure).

• A special thanks to Lois Chinnock, who will be missed very much!

Please feel free to call or e-mail me questions/comments at (317)580-8661 or [email protected]. I look forward to agreat year! h

CHAIRPERSON’S CORNER | FROM PAGE 3

CHAIRPERSON’S CORNER

pharmaceuticals to cover. Formularies have beenimplemented, relying upon differences in cost shar-ing to steer members and their physicians towardless costly or more cost effective choices. A criticalrole in this process is deciding which pharmaceuti-cals are to be covered and at what level of costsharing. Pharmacy and Therapeutics (P&T)Committees typically make these formulary deci-sions, and to do this effectively, they need goodinformation. To make the best drug coverage deci-sions, a P&T Committee should study efficacy,safety, effectiveness, and pharmacoeconomic data.

Since the FDA requires extensive efficacy and safetydata in order for a drug to gain its approval for salein the United States, pharmaceutical companieshave this type of information readily available.

Efficacy research is designed to prove a drug’sscientific value in an ideal setting. However, thissetting will not be seen outside of a specificallydesigned and controlled experimental environ-ment. Use of a drug in a typical health careenvironment, where compliance may be less thanperfect and patients may have concurrent medicalconditions, is more apropos. A drug’s usefulnessfor treatment in the latter environment is calledeffectiveness. Effectiveness gives a better idea as tohow pharmaceutical use will impact patients in thereal world. However, effectiveness studies are lesscommon because they are not necessary to gainFDA approval for a drug and they are expensive toconduct.

More rare are studies of the pharmacoeconomicproperties of pharmaceutical use. Such studiesattempt to show the costs associated with using adrug. Costs are typically assessed in one of severalways, which will be discussed later in more detail.

Standard PharmacoeconomicAnalysis MethodsAs a field, pharmacoeconomics is fairly young andhas a very academic feel. Much of the researchdone on the cost impacts of pharmaceuticals usestechniques adapted from the field of economics,including cost-minimization analysis, cost-effec-tiveness analysis, cost-utility analysis andcost-benefit analysis. These methods are describedbelow.

Cost-minimization analysis (CMA) is the simplestof the methods listed. Cost-minimization analysisidentifies the least expensive option among severalwith equivalent effectiveness.iv For example, a

cost-minimization analysis would conclude thatthe less expensive of two equally effective aceinhibitors is the preferred choice. CMAs are rarelydone because few clinical trials result in the conclu-sion that a drug is equal to its comparator. Mostaim to show its superiority.v This method neglectsother important variables such as the cost orunpleasantness of possible side effects.

Cost-effectiveness analysis (CEA) is done to deter-mine the cost per unit of effectiveness, resulting ina cost-effectiveness ratio. This ratio can be stated asthe cost per unit of outcome, or units of outcomeper dollar spent. Outcomes are measured in termsof clinical events such as heart attacks, hospitaldays avoided or life-years saved. A lower cost-effectiveness ratio (cost per unit of outcome) isassociated with a preferred treatment choice. Thepreferred choice is not necessarily the least expen-sive one, however, since the health gain of theoptions can vary as well. CEA can be a robustanalysis, taking all associated costs and savingsinto account. However, there is considerable varia-tion across CEA studies with respect to types ofpatients examined, measures of effectiveness andcosts used, and the way in which cost-effectivenessratios are calculated and reported, which can maketheir interpretation and comparison difficult.vii

The most meaningful CEA analysis is the calcula-tion of the incremental cost effectiveness ratio(ICER) between two alternative treatments. Thisrequires data from head-to-head trials or at leastfrom different trials that were fairly similar instudy population and methodology. Since mostclinical trials are sponsored by the manufacturer ofone of the drug products within the study, theyrarely provide all the direct comparison dataneeded to answer the questions a health plan isasking. A rare exception is the recently publishedPROVE IT study, which compared two cholesterol-lowering drugs, Pravachol and Lipitor. Althoughthe maker of Pravachol funded the study, itshowed that Lipitor was better.viii This outcomeprobably makes it is less likely that other drugcompanies will want to fund head-to-head trials inthe future.

Cost-utility analysis (CUA) is done to assess thecost per outcome unit that is adjusted for patientvalue placed on those outcomes.ix Rather thansimply assessing life-years saved, for example, theCUA would assess the cost per Quality-AdjustedLife-Years (QALYs) saved. For example, a patientwhose work requires a lot of standing and walkingmight assign more utility to an orthotic device (agait-correcting shoe insert) than a patient who doeslittle standing or walking. Critics of CUA maintain


John R. Watkins, Ph.,

MPH, is a pharmacy

manager in formulary

development at

Premera Blue Cross

in Mountlake Terrace,

Wash. He can be

reached at (425)

918-5146.

Jill Van Den Bos is a

consultant at Milliman,

Inc. in Denver, Colo.

She can be reached

at (303) 299-9400 or

jill.vandenbos@

milliman.com

THE FORMULARY DECISION PROCESS | FROM PAGE 1


that it is difficult to use and compare because thereare numerous different ways to assign healthstatus, no agreement upon what constitutes thegold standard, and whose preferences are meas-ured—patients, providers or public—affects theresults.x

Cost-benefit analysis (CBA) measures the cost peroutcome where outcomes are translated intodollars. In the example above, the patient whosework requires a lot of standing and walking mightbe willing to pay more for an orthotic device thanthe patient who does little standing or walking.The cost to buy the device can be assessed againstits value stated in dollars. This method has thedrawback of having to obtain assessments of themonetary worth of health outcomes. In evaluatingpharmaceuticals, CBA is often used to compare thecost of a more expensive drug with the expectedsavings from reduced need for other medical costssuch as physician visits, hospitalization or emer-gency room care, thereby sidestepping thisdrawback.

Whereas studies using the methods above may beavailable to P&T Committee members, their resultsare not well suited to the needs of a health plan.Such results may help to determine which of thedrugs compared in one study seems to be the betterchoice from a cost perspective, but they do little tohelp health plan decision-makers quantify howand where drugs will have an impact on the overallbudget.

Pharmacoeconomic analyses often rely upon infor-mation from multiple sources, with potentiallycomplicated study designs, making them difficultto perform and analyze. A study of submissionsreviewed by the Australian PharmaceuticalAdvisory Committee found that 67 percent of 326pharmacoeconomic analyses had serious flaws.xiiThe resources available to make that assessmentwere considerable, possibly beyond the capacity ofmany individual health plans. While this maycontribute to a health plan’s reluctance to use suchinformation, to avoid doing so misses a real oppor-tunity to add value to the formulary decisionprocess.

The Development of the AMCPFormat for FormularySubmissionsIn an effort to counter some of the problems withavailable research on pharmaceutical costs, theAcademy of Managed Care Pharmacy (AMCP)

developed and disseminated the first Format forFormulary Submissions in 2000. Version 2 of theFormat, released in 2002, incorporates user feed-back.xiii,xiv The Format is a guideline that specifieswhat information health plans want to see fromdrug manufacturers in order to help them makeinformed, evidence-based, drug coverage deci-sions. This information includes data on efficacy,safety, effectiveness and economic impact of a newdrug. The Format puts responsibility on pharma-ceutical manufacturers to provide all informationavailable in a standardized format.

Since the release of the latest guidelines, AMCPreports that adoption is spreading at a rapid pace.To date, no large studies exist on the impact of theFormat on patient outcomes.

Proponents of formulary guidelines maintain thatthe Format makes great strides in leveling the play-ing field between manufacturers and health plans.The Format creates a standard for constructing,presenting and critiquing models. Early experiencesuggested that manufacturers were unwilling tocomply with dossier requests. However, recentinformation has suggested that most are nowsubmitting dossiers, but they are frequently incom-plete.

The P&T CommitteeTo better understand the process used by P&TCommittees in formulary development, one of theauthors did some informal observations of P&TCommittees and their decision processes. Anotherof the authors is a formulary manager and leadingmember of a P&T Committee. This section willdiscuss P&T Committee features, relying to a largeextent upon these observations.

A lot of research is gathered in preperation for aP&T Committee meeting where drug coveragedecisions are made. As discussed above, informa-tion is gleaned from pharmaceutical manufacturerdossiers, published research, FDA analysespublished on their Web site, and possibly modelingand analysis done by the health plan itself. Thepharmacy staff normally conducts a search for rele-vant primary literature using MEDLINE andpossibly other databases. Secondary sources suchas Cochrane reviews may also be consulted.Summaries of the information from these sources,and sometimes research articles themselves, aredistributed to P&T Committee members prior to ameeting.

THE FORMULARY DECISION PROCESS


Jonathan Shreve, FSA,

MAAA, is a principal at

Milliman, Inc. in Denver,

Colo. He can be


9400 or jon.shreve@

milliman.com

Kristen Reed is a

healthcare consultant

at HealthGrades in

Lakewood, Colo. She

can be reached at

(303) 716-6509 or

kreed@healthgrades.

com

The P&T Committees we have seen are comprisedof primarily physicians and pharmacists. Othermembers included a psychologist, an osteopath,registered nurses and employer representatives.Two of the committees explicitly noted that onlymembers not employed directly by the health planwere allowed to vote on formulary decisions. ThePBM committee profiled allowed one vote perclient, a representative of which sat on the P&TCommittee. These committees ranged in size from11 to 25 people.

In each meeting, a pharmacist or team of pharma-cists gave a presentation of information about thenew drugs under consideration. These presenta-tions were brief, the details having been suppliedto the members prior to the meeting, includedformulary recommendations, and were followedby discussion from the group in general. Thediscussions were very interactive, with many ques-tions and dissenting points of view. In everymeeting observed, at least one recommendationmade by the presenting pharmacist(s) was notaccepted.

If needed, experts outside of the standard P&TCommittee were asked to give relevant opinionsand observations. The pharmacists who prepareformulary reviews usually consulted with one ormore such experts prior to writing their recommen-dations.

Most of the discussion during the meetingrevolved around drug safety and effectiveness.Information for this included both research find-ings and observations from clinical practice. Costs

were not discussed much, although the price of thedrugs and patient copays were mentioned severaltimes. Cost offsets and total budget impacts werenever discussed during meetings. One groupexplicitly avoided the subject of costs, focusinginstead on selecting the most effective and safedrugs from a class of drugs and narrowing that listdown to the best few. Once that list was deter-mined, final formulary placement was determinedby the deals that could be negotiated with themanufacturers.

While the subjects discussed were pertinent to theparticular drugs under consideration, several inter-esting and fairly animated discussions occurredaround the following topics.

• During the educational component of one meet-ing, a presentation was given on special featuresof biotech drugs, their anticipated utilizationand costs as a class, and strategic initiatives toappropriately plan for their influence on treat-ment and the pharmacy budget.

• One group brought up a perceived connection between the FDA and the pharmaceutical indus-try and the expected impact on the FDA’s abilityto provide impartial expert opinions on prod-ucts reviewed.

• Another discussion involved the desirability of covering drugs that provided no unique bene-fit to patients other than convenience. Anexample of such a drug is Seasonale, a new 3-month course of oral contraceptive that allowsthe user to restrict menses to four times per year.

• Concern over the convenience and cost to patients when using the pharmacy benefitsurfaced in several meetings. For example, somenew drugs combine two drugs that are alreadyavailable separately, but having them combinedunder one copay would save members money atthe pharmacy.

• In another meeting, members expressed concern that patients might be confused when requiredto obtain prior authorization for an injectabledrug and then have to write a large check at thepharmacy when this was not required for otherdrugs. The up-front payment requirement coulddiscourage some members from filling prescrip-tions.

• Most meetings included some discussion of manufacturer strategic maneuverings. Theseincluded acknowledgements that drugs like



Clarinex which is slightly different fromClaritin, or Nexium which is slightly differentfrom Prilosec, or new formulations such asWellbutrin XL (once per day) are developed tocapture market share from another product fromthe same manufacturer that is about to losepatent protection.

• One meeting included a discussion on usingclinical trial and other data to approve a drugfor the formulary when much of the anticipatedusage of that drug, such as the epilepsy drugTrileptal, would be off-label psychiatric use forwhich data was not available.

• Only one committee (a large PBM) specifically talked about rejecting pharmaceutical manufac-turer models in favor of doing its own economicanalysis. Other groups discussed costs of thedrugs or copays, or mentioned when economicresearch was not part of the dossier (evidentlynot uncommon).

With increasing public attention to pharmacy bene-fit management processes, health plans shouldimplement formulary decision making processeswith the goals of improving clinical outcomes andreducing overall cost of care rather than simplymaximizing rebates and minimizing drug expendi-tures. These strategies may also help to alignincentives for health plans, physicians, pharmacistsand patients. xix

Formulary Decision Making—What Do We Know About theProcess?Health plans, PBMs and hospitals follow the samegeneral process when evaluating a new drug forformulary submission.xx,xxi,xxii Guiding principlesfor clinical decision-making have been defined asfollowsxxiii:

• Assess the findings of peer-reviewed medical outcomes research and pharmacoeconomic research,

• Employ published practice guidelines, devel-oped by an acceptable evidence-based process,

• Compare the efficacy, effectiveness, value and therapeutic interchangeability,

• Compare drugs on patient compliance, and

• Do a thorough evaluation of benefits, risks and Adverse Drug Reactions(ADR).

In practice, P&T Committees examine safety andclinical effectiveness first, then the incrementalvalue of a drug compared to existing alternatives.If a drug has superior clinical properties and has noequal counterparts, then it is added to the formu-lary. If a drug is inferior to an alternative on theformulary, then it is not added. If the drug showseffectiveness equal to a drug currently on theformulary, then costs are considered in the adop-tion process. If there are unanswered questionsabout the product’s safety, the decision is usuallydeferred until more data are available.

Most sources of information, including manufac-turer dossiers, published literature and FDAdocuments, focus on clinical and safety issues.Economic information is sometimes available.Current evidence suggests that pharmacoeconomicinformation is not widely used by decision makers,however.xxiv,xxv Some reasons are listed below:

• Health plan decision makers are skeptical ofinformation provided by drug makers.

• Decision makers report being uncomfortable with the extensive use of assumptions in phar-macoeconomic analyses. They prefer observeddata.

• Health plan decision makers have a general concern about the aggregation of health benefitsinto a single index such as Quality–AdjustedLife-Years (QALYs) saved. They prefer to exam-ine independent components.

• Impacts on the budget are often missing. When included, the cost of a new drug is oftenconfined to its effect on the pharmacy budgetalone. This misses the impact in other treatmentareas.

• The information is not presented in language used by health plans. They want to know theeffect on overall cost per member per month oftheir benefit, rather than the cost to prevent ahospitalization or cost per QALY gained.

• Pharmacoeconomic information typically lacks head-to-head comparisons with the most rele-vant treatment alternatives.

• Health plan decision makers need to know how a particular drug is going to affect their ownpopulation. Concern about transferability ofmodel results is a barrier to their use.




How Actuaries Can HelpThe FDA does not have a mandate to evaluate adrug’s cost-effectiveness as a part of the New DrugApplication (NDA) process. Although an NDAsubmission includes a literal truckload of data, theFDA review focuses entirely on safety and efficacy.An expensive drug with only marginal clinicalbenefit may be approved if the reviewers concludethat the reported efficacy outweighs the potentialtoxicity, regardless of cost. Therefore, P&TCommittees must do their own economic evalua-tion of new products if they are to weigh value intheir decision-making.

Pharmacoeconomic research currently available toP&T Committees, although much improvedfollowing the dissemination of the AMCP Format,is not fully meeting their needs as indicated above.While conducting economic research is not particu-larly actuarial, modeling is. This seems to be anarea where actuarial methods can fill a need.

The primary area in which the pharmacoeconomicmodeling falls short is in the inability to specifyand quantify any medical cost offsets associatedwith the use of a drug. While the AMCP Formatcalls for quantification of budget impacts in themodels requested, health plan decision makershave expressed dissatisfaction with this element ofthe dossiers received. An informal review ofdossiers submitted to one health plan over the pastthree years showed that no more than 15 percent ofthem contained useful disease-based models.

When a reasonably constructed model is submitted,the health plan may still need to adjust the manu-facturer’s assumptions to get a relevant estimate.

Furthermore, economic models typically comparethe manufacturer’s own drug to a single compara-tor or to placebo. A more useful model wouldincorporate all the relevant treatment options forthe medical condition of interest in a single head-to-head comparison.

Models could be made more useful by the use ofdynamic population modeling typically used byactuaries. Pharmaceutical company models areoften based on populations studied in clinicaltrials, or on populations that come from canneddatabases rather than (a) reflecting the populationof the health plan, and (b) allowing the user tomanipulate the population mix. Population consid-erations should include features unique to the typeof payer, such as commercial, Medicare, Medicaidor TRICARE populations.

An ideal model would incorporate these capabili-ties, reflect the prescription coverage benefitdesign, medical condition incidence and preva-lence, the rate at which the new drug will enter thesystem and replace or supplement other treat-ments, utilization and costs associated with themedical condition and side effects of the treatmentoptions, expected compliance rates, and the level ofhealth care delivery management expected in thesystem. Estimates of parameters in this model can



HOW ACTUARIES CAN BRIDGE THE GAP

Clinical Trials

METRICS

ACADEMIC

CLINICIANS

MORE ACADEMIC MORE PRACTICAL

PHARMACO-

ECONOMISTS

FORMULARY

COMMITTEES

EMPLOYEES &

FAMILIES

CONCERNS

• Clinical endpoints

• Adverse events

Does it work?Is it safe?

EconomicModels

• $/Outcome• $/QALY• Cost offset• Target pop.

It is cost-effective?

MCOs

• $ PMPM

• Total Budget Impact

What’s mybottom line?

Members

• Premium• Copays• Affordable

Care

Can I afford theout of pocket$$?

Actuarial Modeling



be obtained from the medical literature, expertopinion about reasonable clinical pathways, studyof prior claims data and other expert judgment.

The people best qualified to create such a modelare in the actuarial area. It would not only be avaluable tool for the formulary decision process,but would have much more broad usability withinthe organization. Economic outcomes expressed inper member per month claim costs could bereviewed and used by actuaries when monitoring

experience and preparing for pricing. Specificationand quantification of medical cost offsets, orincreases, that result from the use of drug treat-ments could be useful to people in caremanagement and utilization management roles.Ultimately, pharmaceuticals are an integral part ofgood medical care and their costs should beviewed as part of the total budget. As biotechnol-ogy drives up the average cost of new drugs, astrong partnership between actuaries and pharma-cists is crucial to the success of a health plan.h

i Neumann, P.J., Sandberg, E.A., Bell, C.M., Stone, P.W., Chapman, R.H. Are pharmaceuticals cost-effective? A review of the evidence. Health Aff (Millwood). 2000;19(2):92-109.

ii Milliman USA. 2003 Intercompany rate survey.iii The National Institute for Health Care Management Research and Educational Foundation. Prescription Drug Expenditures in

2001: Another Year of Escalating Costs. May 6, 2002.iv Drummond, M.F., O’Brien, B., Stoddart, G.L., Torrance, G.W. Methods for the Economic Evaluation of Health Care

Programmes, 2nd Edition, Oxford University Press, 1997.v Newby, D. Use of pharmacoeconomics in prescribing research. Part 2 – cost minimization analysis – when are 2 therapies

equal? Journal of Clinical Pharmacy and Therapeutics. 28:145-150.vi Drummond, M.F., O’Brien, B., Stoddart, G.L., Torrance, G.W. Methods for the Economic Evaluation of Health Care

Programmes, 2nd Edition, Oxford University Press, 1997.vii Mullins, C.D., Blak, B.T., Akhras, K.S. Comparing cost-effectiveness analyses of anti-hypertensive drug therapy for decision

making: mission impossible? Value in Health. 2002;5(4): 359-71.viii Cannon, C.P., Braunwald, E., McCabe, C.H., Rader, D.J., Rouleau, J.L., Belder, R., Joyal, S.V., Hill, K.A., Pfeffer, M.A., Skene,

A.M. Comparison of intensive and moderate lipid lowering with statins after acute coronary syndromes. New England Journal of Medicine. 2004; 350. Pre-publication posting at www.nejm.org. (To be published in April 8, 2004 issue.)

ix Drummond, M.F., O’Brien, B., Stoddart, G.L., Torrance, G.W. Methods for the Economic Evaluation of Health Care Programmes, 2nd Edition, Oxford University Press, 1997.

x Brinsmead, R., Hill, S. Use of pharmacoeconomics in prescribing research. Part 4: is cost-utility analysis a useful tool? Journal of Clinical Pharmacy and Therapeutics. 2003; 28(4): 339-346.

xi Drummond, M.F., O’Brien, B., Stoddart, G.L., Torrance, G.W. Methods for the Economic Evaluation of Health Care Programmes, 2nd Edition, Oxford University Press, 1997.

xii Hill, S.R., Mitchell, A.S., Henry, D.A. Problems with the interpretation of pharmacoeconomic analyses. Journal of the American Medical Association. 2000; 283:16: 2116-2121.

xiii Academy of Managed Care Pharmacy. Format for Formulary Submissions Version 2.0. Academy of Managed Care Pharmacy, October 2002.

xiv Fry, R.N., Avey, S.G., Sullivan, S.D. The Academy of Managed Care Pharmacy Format for Formulary Submissions: an evolving standard—a Foundation for Managed Care Pharmacy Task Force report. Value in Health. 2003;6(5):505-521.

xv AMCP, The Year In Review, 2002-2003. Available at www.amcp.org. Accessed February 17, 2004.xvi Neumann, P. Evidence-based and value-based formulary guidelines. Health Affairs. 2004:23(1): 125-134.xvii Wang, Z., Salmon, W., Walton, S. Cost-effectiveness analysis and the formulary decision-making process. Journal of

Managed Care Pharmacy. 2004; 10: 48-59.xviii Otompke, J. A new format for making more cost-effective drug coverage decisions takes off. Business and Health. January 2002.xix Chung, R.S., Taira, D.A., Noh, C. Alternate financial incentives in multi-tiered formulary systems to improve accountability for

outcomes. Journal of Managed Care Pharmacy. 2003;9(4):369-71.xx White, J. Making pharmacoeconomics in formulary development a reality. Managed Care Magazine. 2001; 10(2).xxi Grabowski, H., Mullins, C.D. Pharmacy benefit management, cost-effectiveness analysis and drug formulary decisions.

Social Science Medicine. 1997; 45(4):535-44.xxii Saprong, D.F., Application of Pharmacoeconomics and Outcomes Research In Formulary Decision Making. Drug Benefit

Trends. 1999; 11(8):53-57.xxiii Pharmacy Benefits Management Strategic Healthcare Group Department of Veterans Affairs, Principles of a Sound Drug

Formulary System, available at www.vapbm.org/PBM/formularyprinciples.pdf, accessed March 1, 2004.xxiv Wang, Z., Salmon, W., Walton, S. Cost-effectiveness analysis and the formulary decision-making process. Journal of

Managed Care Pharmacy. 2004; 210: 48-59.xxv Drummond, M., Brown, R., Fendrick, M., Fullerton, P., Neumann, P., Taylor, R., Barbieri, M. Use of pharmacoeconomics

information–report of the ISPOR Task Force on use of pharmacoeconomic/health economic information in health-care decisionmaking. Value in Health. 2003; 6(4): 407-16.


IAA Health Section Update

Sections are becoming a key area of interest tothe IAA (the IAA is an association of membernational actuarial associations), and IAA’s

leaders are actively encouraging their developmentin order to bring together actuaries from aroundthe world. IAA sections are grassroots organiza-tions that largely set their own agendas, much likethe very successful SOA sections.

The IAA Health Section Committee (HSC),which was formed in 2003 to lead the newlyapproved Health Section (HS), has been verybusy planning how best to serve the needs of theinternational health actuarial community. Ourthree most important recent activities are: adopt-ing a planning document entitled “Way ForwardProposal” (this important document can beviewed in the Health Section portion of the IAAWeb site at www.actuaries.org), forming nine TopicTeams, and organizing a health track for the ParisInternational Congress of Actuaries in May 2006.

In late November the IAA’s HSC held aconference call and, among other matters, agreedto form nine Topic Teams. These teams aredesigned to bring our HS members together towork on problems and areas of interest to a broadcross-section of our members—so they are vital tothe HS’s success. The new Topic Teams are asfollows:

PRODUCT TEAMSLong-Term Care Insurance (Avi Bar Or*, Israel)Income Protection InsuranceCritical Illness Insurance (Sue Elliot*, U.K.)Private Medical Expense InsuranceVoluntary Health Insurance (to supplementsocial insurance benefits)

PROCESS TEAMSHealth Risk Adjustment (Lori Weyuker*,U.S.A.)

POLICY TEAMSHealth Systems in Developing Nations (AlvaroCastro*, ILO - Switzerland)Public-Private Health System Partnerships Nontraditional Medicine (Heather McLeod*,South Africa)* Team Leaders

The goals for Topic Teams are to create aninternational community of actuaries and others(membership is not limited to actuaries) who seek

to gather, discuss, research and disseminate infor-mation to our members. The HS (through the IAASecretariat) has many communications tools tohelp the Teams to do their work: our Web site,listserver capabilities, online and teleconferencingcapabilities, our Online Journal, the IAA’s ASTINBulletin and periodic health colloquia and meet-ings. Teams are not limited to simply making useof these existing capabilities. They could, forexample, decide to hold an international confer-ence or join with other organizations to work onprojects. What teams do is largely up to theirmembers to decide. The IAA’s HSC's role is tohelp teams to accomplish their goals. The “WayForward Proposal” provides a more in-depthdescription of Topic Teams and their importanceto our future.

Over the past few weeks, many IAA HealthSection members have responded to communica-tions from the IAA’s HSC and volunteered to jointhe new teams. However, all of them still needmore members, and many teams are in need ofleaders.

Following on the success of our first twointernational health colloquia, one in Cancunduring ICA2002 and the other earlier this year inDresden Germany, we are actively planning ourthird international meeting, which will be held asa part of ICA2006 in Paris. We have formed anorganizing committee headed by ClaudeFerguson to coordinate our efforts with the ParisOrganizing Committee to develop our scientificprogram. We plan an extremely interesting two-day health track which will provide healthactuaries with opportunities to hear from expertsaround the world and explore many topics ofinterest to SOA and Academy members from aninternational point of view. Our HS OrganizingCommittee will provide members with moreinformation about the program and how toparticipate by writing a paper or speaking on apanel. The HS Committee approved a 500 Europrize for the best health paper submitted toICA2006.

Our Topic Teams are certain to be importantresources for all of us; however, they simplycannot function without your help. Please let meknow if you are interested in joining a Topic Team or helping to organize the health track atICA2006 by e-mailing me at [email protected].


I also encourage those of you who have notalready joined the IAA Health Section to go to theIAA Web site (www.actuaries.org) and sign-up byclicking on “How to Join” under the IAAHS(Health) tab. 2005 dues will be collected by your

national actuarial organization some time nextyear. h

Howard J. BolnickC H A I R M A N, IAA H E A LT H S E C T I O N

New Opportunities for HealthSection ParticipationAs you may be aware, the SOA has been making some changes in order to strengthen its grassrootsconnections through the special interest sections. To that end, the Health Section will be broadening itsrange of activities over the next several months. To support this expansion, the Health Section Councilhas or will be establishing a number of activity-based volunteer teams. In the near term, four teamswill be focused on Communications and Publications, Continuing Education, Research andProfessional Community (a.k.a., external relations).

The Health Section Council has been working with the health practice area volunteers to define thebreadth of its activities for the future and this new organizational structure. At a meeting in Chicagoon Dec. 3, both groups discussed the transition of the Health Section Council to this new role withinthe SOA’s organization, as well as key issues facing the health actuarial practice in 2005. The key issuesinclude:

• Health care affordability and financing;• Outcomes measurement and cost/benefit studies to drive the efficient use of health care resources;• Data requirements and mining to support such analyses;• Professional visibility in the health care industry

These issues will provide some focus for the Health Section’s activity-based teams in the upcomingyear. If you are interested in finding out more or participating as a member of one of these teams,please contact Kara Clark, the SOA’s Health Staff Fellow, at [email protected]. More information about theteams will also be included in the monthly editions of the electronic Health E_News. h

Availability of the DisabilityChart BookThe SOA and the America’s Health Insurance Plans (AHIP) have just released a jointly developedconsumer Chart Book entitled, Disability Insurance: A Missing Piece in the Financial Security Puzzle, toeducate the general public on the need for disability insurance coverage. The chart book is a graphicdepiction, supported by narrative, of the fundamentals regarding disability risk, the financial risk ofdisability, and the availability and limits of public disability income programs. The chart book isdirected primarily at media outlets and secondarily at consumers, employee benefits decision makers,and policymakers. The chart book was funded by the Actuarial Foundation.

The SOA’s Disability Chart Book Task Force was chaired by Thomas R. Corcoran. Members included:Thomas M. Ciha, Kara L. Clark, Peter M. Crockett, Patricia J. Fay, Scott D. Haglund, Delaine B. Hare,Emily Kessler, Kenneth M. Latus, Debra Sue Liebeskind, Allen D. Livingood, Charles H. Meintel, AnneG. Mitchell, Alex N. Moral, Matthew R. Naughton, Lori A. Nelson, Kari C. Powell, Ellen J. Retz,Forrest Richen, Susan R. Sames, Robert E. Schneider, Bruce D. Schobel, Eric L. Smithback, Douglas W.Taylor, Amy Thompson, Maria N. Thomson, Charles M. Waldron, Carl A. Westman, and Thomas F.Wildsmith.

An electronic version of the chart book is available on the SOA Web site at:http://www.soa.org/ccm/content/areas-of-practice/health/publications-downloads/disability-insurance-a-missing-piece-in-the-financial-security-puzzle/

If you would like a hard copy of the chart book, please contact Susan Martz at [email protected] or 847-706-3558. h


Overview of Sessions at the2005 Health/Pension SpringMeetingIn addition to a significant number of highly relevant, practical sessions in the traditional 90-minuteformat, the Health Section is sponsoring a number of featured embedded seminars within the SOA’sHealth/Pension Spring Meeting program. The seminars include:

• Affordability: The Market Response (3/4 day);• Affordability: The Regulatory Response (3/4 day);• An Introduction to Care and Disease Management Interventions and Their Implications for

Actuaries (1/2 day);• Financing Chronic Care (3/4 day);• Electronic Medical Records: Impact on Providers and Health Insurers (1/2 day);• MMA – The Biggest Changes Since 1965 (1/2 day).

The SOA’s Health/Pension Spring Meeting will be held in New Orleans, from June 15-17, 2005. Formore information on these seminars and the other Health Section-sponsored sessions, please refer tothe SOA’s Web site at http://www.soa.org. h

health section news · as a result of the soa's governance audit we've been hearing about...

Documents