It is estimated that 20–30 percent of overall health care spending may be wasteful. Overall, employers and workers spent $244.2 billion to $366.3 billion on six domains of waste: failures of care delivery, failures of care coordination, overuse of low-value health care services, fraud and abuse, administrative complexities, and pricing failure — or when prices migrate from what is expected in a well-functioning market.
In this Issue Brief, we focus on that last area of waste: pricing failure when it comes to lab; imaging; and special medications for conditions such as multiple sclerosis, rheumatoid arthritis, and other inflammatory disorders. This analysis is important because care is shifting from physicians’ offices (POs) to more costly hospital outpatient departments (HOPDs). Compounding this shift in care is the fact that prices for hospital-based outpatient care are increasing faster than physician prices. Ultimately, employers and workers bear the brunt of cost differences when HOPDs perform services that can be provided in less costly POs or in stand-alone lab or imaging facilities.
Key Findings: Location matters...
- To employers and employees:
- Overall, employers and workers would collectively save $11.2 billion if price differentials between HOPDs and other sites of treatment were eliminated for each of the 25 health care services examined in this report.
- Employers would save $9 billion or 80 percent of the total, whereas workers and their dependents would save $2.2 billion or 20 percent.
- When it comes to a wide variety of health care services:
- Employers and workers could reduce their spending on lab services as much as 69 percent, depending on the type of lab service, if price differentials between HOPDs and other sites of treatment were eliminated.
- Savings could be as high as 56 percent for chest X-rays, 49 percent for echocardiograms, and 41 percent for DEXA scanning.
- If site-of-treatment price differentials for specialty medications were eliminated, employers and workers would save as much as 36 percent, depending on the medication.
While $11.2 billion does not seem like much savings compared with the $1 trillion spent on health benefits by employers and workers, the potential for savings is much greater, as our analysis is based on only 25 health care services. Employers could cut costs by 1 percent simply by shifting patients away from more costly HOPD settings or by negotiating site-neutral pricing for the 25 health care services examined in this report.
This study was conducted through the EBRI Center for Research on Health Benefits Innovation (EBRI CRHBI), with the funding support of the following organizations: Aon Hewitt, Blue Cross Blue Shield Association, ICUBA, JP Morgan Chase, Pfizer, and PhRMA.
Figure 14 Figure 7 Figure 12 Figure 3 Distribution of Various Specialty Medications, by Site of Service Aggregate Savings Average Number of Units of Specialty Medication per Patient, 2018 Use of Various Health Care Services, 2018 m I I G W R Fi I Fi Out n th n n th e na g e hil e ne g ur dfe e lly e ic pat e is e rne a a $1 , r a 10 I tlly us re ions ss b ie a 1.2 b senc , , nc l, ue ing nt tAhe a v w e llow B e S e a ill e r rs a v e ie a ion d rof llow e e g rvices e fr , e d a r su a w g A e c roe e c e ha llow d eh ma ly m foc s not cr ha g e a ee m us d d rr s a m g b k C see e e e on ini ha r trfor e st p out w m rv ow g e e a hom e rlik r - st e e y of b d e r ey si - , a p e c m m e Sit ta oc m inf uc us ila hom e k pus h sa e e rloy of t d for ions e ae Se b .m v r yThese lis a a r ngs ount p v b w r ic nd ic a se es ing cs r om a ins som inf v — nic d fa p ur us e da W e ilu s p e e ions rw d e r e rs ha uc d r ight e ov , c w tta oc a ible ide n it c e na c om h d a ur d ly ttP + the p r zin in P e e e a m r cd c roins g $1 a e p Os a in b b nt t25 le tta ur r o ill ot g he tnd a e o e ion sp h HOP nc nga a t Pla he lt e rh ca ic b g + e s. a e e llow D nt D cp A rop s a e iff a s a on tse e e a ie nd d r y n e nt e rhe v fr c nt P s t ic ha xa om a Os, ia e lt hr rs m l for h b g oug c in w e p a la le nd e for im Va ne , h inc hic t w she P fit r ious hil Os. — hs b ra e te in t he llow a y T tN he sed he ron he ee y rd is - e p w sa MRI rw e ic m re e ee r e Endnotes (in millions) Paul Fronstin is Director of the Health Research and Education Program at the Employee Benefit Research Institute Loc 70%ation, Location, Location: Cost Differences in Health 11 65% e p not e c t w rha xc m ot a ans y p e e re g loy nt p a p ve ts t e a ia ions e nly for rl p e ot rs a nc I a m , ra split ic l a y how nd am .ing g llow H e ing w b ow tea fa o e vb e t re Se ilu e w k d olic rv e , e r r a e re v su e s, t m r ic .n th , p c ount a Thes e ph a he ne s ric ................................ e l e s p s te w w ot t a w ra ser ut o a e e s ns nt r oim sett $1 v eia p ic e a 5 in l for e m ings, m rs e u p nc fa ne loy tsa he ll int yne e d v b d is P ings is mu y it o O e he in a it atself a se hr r................................ nd ll c sett ed e ha r $1 a u c ing lc s b g a c 1 in ul h t# e a e w g g 1 te r or ions a . te n found he s St a iet p ill s: la e la a p r, r,b rt te ic .a he b v s our ula D t ious ; o rinc im ug r ................................ bly e a ly ide t a g d e ins na d ing om st nc euff ly s sc ; eina si w r a ic of ibe s is nd eie nt rho e n d . spe b t, tm he a w in r sed e c e ia only inf e w l me don e us uc r ................................ e ser ions ing only daic ls va ic hospit o w t 25 e ions a a b in w s q lhe e a for uit ta l p hic o lt e h ca split r c h ic ond sm e the s ra te ithe llr io , e... n so w s a 13 s (EBRI). M. Christophe Sr pe Roe ciab ltuc y M k eis di P ca re tisi on ds ent and CEO of RxEconomics, LLC. This Issue Brief was written with assistance Berwick, Donald M., and Andrew D. Hackbarth. 2012. "Eliminating Waste in US Health Care." JAMA 307 (14): 1513- 63% In this research, we w ere interested in analyzing outpatient health services that generally met three conditions: Care Services by Site of Treatment — A Closer Look at Lab, 2018 Employer & su e unl t ser ahe st si cim v e h a g highe ic ss nif a e s m ts. c ic eom d aE r nt ul a m a bg t llow ine p iple d gloy iff re d e e g e sc rd w a re s tle it nc c e h p ha c r sa ou e osis rin t la v g ld ings e , n d c he rfo he ue r t int si a um c tllow he g os o n ch a tthom s b the e oid da y p nge c e a or 1 p ha rinf tthr rs ions e gus int rit ec is ion e b e ,nt te nde ha a t w d nd si tid e m d m e ot p not n P tigh o ly he st O tb ha r e y s a ainf e c v sh rce nd la rp ift ue a m a ing tla m m ie tb o a nt astp tm , e or sa r w e t tia y ie o m it l im d h la nt le bis e ss s a or rp b s a d c w v c ros e a t s. a r ty s on tely .p s b fr la Thi s tom he n ie ts a e isponsor ng s of m pna ror ic sie ly g e tr nif si e d cs. a iff s is ost ic t m e aly r nt e im e nt nt H ly p OP . ia or highe For l.t D a Ov nt sett ins re .b r ta O e ia ngs ll, c nc va e us te rhe a or , e ll, r e b y Primary immunodeficiency drug #1 899 from the 1516. Instit ute’s research and editorial staffs. Any views expressed in this report are those of the authors, and 58% Figure 5 1 Figur60 e % 11, AveraHO geP D Allow PO ed CHo ha me rge by Site of Service and Weighted Percentage Price Differential for Various Specialty Employer spending o La n b health insurance premiums includes dire Ecm t p pla oy yme ee nt of E cla mims ploy e in r self-ins Em ure plod y e pe la ns. Im aging, and Specialty Medications 55% w e p ne ca m re ic rg re pe e ot 1. loy is d si ia iff g sh e ting nif rThe es a ift re ic ing si nt nd ahe tnt ia efr - ins a ls ne p om lt rb h ca ur ic ut e e e tP rw r a Os t d r s c l p e e iff Pe rse a e r n H im o ic n rr e a ing m v us nt rOP ic yor ia e e im for D e s ls a m s a rc p c a u ost tom e he n n nd o r k ly d e uni b 25 e d ot ina fhospi i chighly he tihe e tof n ion o ra c si t spe y lta th ca d l o e in t rfs of u c ut v g iaa e r# lt p e lue r2 y m t aserv r te m s of ie -a be nt tam d ic sed ic ut e d es e a e nt ili t pins ion, za a xa rr attur nge ion o m m ra ine e anc nt nging d d re s ( w cin t d ide ost H e OP si fr his ly g om a D n (VB nd — rs e )15 fr .p t he om or Fo – ID 84 r tre . ) 65 for e a p xa nd e e p rm c e ce rrp ont e c nt le fe e 1.nt , ,r 6 ribut in 200 e 1 t nc 8o e e 53 dp 4, a 1 p r si ic ging nif e prp cic re toxi o a nt nt v .m a lP yra ryttic o ee ly should not be ascribed to the officers, trustees, or other sponsors of EBRI, EBRI-ERF, or their staffs. Neither EBRI nor Distribution of Various MRI Services, by Site of Service Medications ................................................................ Savin ................................ gs Savings ................................ Savings ........... 13 February 18, 2021 • NoM . e525 tabolic panel 28% 52% 2 Chernew, Michael, ZackP C rioop marye r im , E muge unod ne ef ic Lia er nsen cy d-rH ug a lloc #3 k, and Fiona51% Scott Mor ton. 2019. "Are H 1e ,a 09 lt3 h Care Services p 94 diff atp ie er e nt e rnt c e cia ost ntls of sh for cahe r ding rm ug ot bthe a esed st ra s p a o ycn th c inf ount us e ions c ehoic d for w ee s t trhe eha a 65 d t m t he p ini eyrst cm e ent r ae k d e a tirn e the g pa hy rb dsi ot ing ctia om ns choic ’of offi e the cof e s ( rhe ange Pa Os lth ca .) , Hb aut rde tb p he y rov 20 aide llow 14r,.e t d ha ct ha pregre c efo ntra d greug had By InsPa titute ul Fro ov oe f rMe ans ll he dic tin, P ine alth ca of h.D., the re c Nost a E tional smplo . Acye adee mie Ben s (2010) efit R ; Se hr sa enk, arch RogIns stadt,itut ande P, aa rend kh (M. 2019) Chris . topher Roebuck, EBRI-ERF lobbies or takes positions on specific policy proposals. EBRI invites comment on this research. HOPD PO Figure 10 Lipid panel 38% 48% 50% 4 Where Do Plan Members Receive Their Health Care Services? Figure 12 Shop , Ave pra ab gle e ?N E um vide Ab ue tnc o ri m of em fr Unit uom ne s of dtihe sea Sp C se onsu e dc ria ugm lt # yp 1 Me tion o dicf Low ation e pre-rL im Pab tie MRI nt, ." 20 N 18 B E ................................ R Working Pap 47% er 321 No. ......................... 24869 (National 14 d t erst op s in la 2. pe d The tb o s b 57 he e p eae n e lt rh ca ce qnt ua r w e l tit se o h a trhe vic c e or as llow rw espond ee re d fr cha e ing qru ge sh ent in P ift ly , tow Os, th buta r not de H p a OP rlw icD a ey sd s, p .iff C eom e re rfo nt pround ia ml w edould ing in HOP t he ha v D sh es. ift b e e in c n c aloser re from to 20 POs t 0 pe o rH ceOP nt, Ds is Ph.D., RxEconomics, LLC 3 Average Allowed Charge by Site of Service and Weighted Percentage Price 70% Total $11,265 $9,020 $2,246 Berwick and HackbarthD (r2012) ug tes.t Figure 13 2% Under VBID, cost sharing is aligned with the value of health care services received. It is built on the principle of Multiple sclerosis drug #1 2,566 The distrB ibut ureion o au of f tr Ec eonomi atment c Re sett sea ings for rch). lab services is shown in Figure 4. Lab services are provided in three types of t m he or fa e c in line t that w pit ric h th es for e ot hospit her lab al -ser bav sed ice sout ’ pr pic ae tie dnt iff e cr ae rnt e a iarls e .g rowing faster than physician prices. A recent study found Sug3. ges ted The ci he tatio alth ca n: rFr e onst servic in, P es a aul lso , a ha nd d M. subCst hr ais nttia op l p he erne Roe trab tion i uck, n o “Loc the atrion, treaLto mce ant tion sett , Lings. ocation: Cost Differences in Differential for Various Non-M 41% RI Imaging Services Percentage Savings From Eliminating Site-of- Figure 13, P 40% ercentage Savings From Eliminating Site-of-Treatment Price Differentials, by Health Care Service ............ 15 Blood 40% co40% unt 26% This study was conducted through the EBRI Center for Research on Health B 12enefits Innovation (EBRI CRHBI), 4 39% lowering or removing financial barriers to essential, high-value clinical solutions. To the degree that the quality of Multiple sclerosis drug #2 909 Winn $ 40 2,0 % e0t 0al. (2018). settings: HOPDs, POs, and stand-alone lab facilities. Metabolic panels and blood counts were nearly equally distributed that between 2007 and 2014, prices for hospit 61% al-based outpatie 61% nt care increased 25 percent, while physician prices Health Care Services by Site of Treatment — A Closer Look at Lab, Imaging, and Specialty Medications,” EBRI Issue Location, LocatTion reatm , eL nto Pc rica e HO t D Pion i D fferPO e: ntCo iaA ls mb , us b lay tory t H Di S eua rglitff ch al C e Pa rr ir ce e eDi S n ffe erenti c rve ial ces in Health with the funding Lsu abV pip taor mt in of D t she cre follow ening ing organizations: Aon Hewitt, Blue Cross 10% Blue Shield Association, Coop60 er % , Zack, Stuart Craig, Martin Gaynor, Nir J. Harish, Harlan M. Krumholz, and John Van Reenen. 2019. "Hospital Supportive cancer drug 4 he Fi Ther Th ga e ur lt second e h ca e w 1a 4rs , eA m se g cor rgr itrv e e e 5ic rg cion w e a onsi s d te oe S st a ae s m s not vnc ings y e a in a ................................ v nt a rtllow yo w highlight it eh sit d cha e rof g ser e ts rve a ic a nd e tm s w ................................ p ernt ic he e , r VB e d iff c Iost D er e c ould nt saia vings o ls b for e us ................................ t phe e pd or v tto a uni re ious nc tieour s MRI ca ur gr e ser e nt pa v ly tic ie ............................... e ent x s e is s t txa o via m 34% see ine eit khe d t rin t r e a pthis rm ice e nt 16 across the three settings. However, lab facilities processed about one-half of lipid panels, drug tests, and vitamin D g rew 6 percent. Ultimately, employers and workers bear the brunt of cost differences when HOPDs perform services 5I Brnt iefr , o no. duc 525 t ion (Febr uary 18, 2021). 55% Cooper et al. (2019). Metabolic panel 1,601 1,030 570 ICUBA, JP Morgan Chase, Pfizer, and PhRMA. 54% 54% Ca $1 r,80 e0 Services by Site of Treatment — A Closer Look at Lab, Prices Grew Subst Auto aint mm iau lly ne Fa dis st ee ar s eTha drun P g #2 hysician Prices For Hospital-B ased Care In 2007 1,6 –0 14 6 ." Health Affairs 38 st in P udy. Os b Aly low low ed e rc ing ha rc gost es fo shra M ring, RI ser relv aic tiv es e w to er H eOP in t Ds, w he $1 he ,100 n ser rv aic nge es a in HOP re recD eiv s, w ed hil ate t ha the t y si t w ee of re cle ass re .t han $600 in POs negotiations or site-of-care shifts. screenings, while POs accounted for one-quarter to one-third of these services, and HOPDs accounted for 15–19 that can be provided in less costly POs or in stand-alone lab or imaging facilities. 30% Spending on health care in the United States reached $3.6 trillion in 2018. About one-third, or $1.2 trillion, was spent Ima Ligi ping d p a vn ia e lMRI 356 262 94 Employer & Employee 6 (2): 184-189. doi:10.1377/hlthaff.2018.0525% 424. DEXA is dual-energy X-ray absorptiometry. Im (Figura e g 9)ing . The , a llow an S eo dd u c rcha S e: rE gp m e pe in HOP loyc eeial BD en s t ew fiy ta R s M er soug eae rchly hd In ic 90 stit u p a te et re cion s etnt im ahighe tes s bars fo edr o4 o n af th dmin ei s5 tra ser tivev ices examined in Figure 9. Copyright Information: This report is copyrighted by the Employee Benefit Research Institute (EBRI). You may copy, perce 50 nt %. Savings on health insurance pre D M m rR uium g I s te ps a s in tend out-of-pocket payments b y employ 97 ers a nd indiv 71 idua 1% ls (Figure 26 1). Employers paid $1,600 H Inow cont eve ra r, st t he to tVB hirId D , crrit ee fe riron enc is p e pe ric rha ing ps t rahe ise s c most ost im shpaor ring tant w . he Idn p enta ify tie ing nts se serv eic k ecs t are ha att hca ev re ta a in he suba st lta h ca ntiarl p e p erne ovtide ration in rs ?37% enrollment and claims data. 20% The only exception was imaging for low-back pain. While MRIs can be used for low-back pain and uncomplicated print, or download this report solely for personal and noncommercial use, provided that all hard copies retain any and 7 19% 1 Figure 4 The Choosing Wisely ca Bmp looa di g cn oup nro t motes conversations betw een clinic 933 ians a nd patients 631 to help patie 302 nts choose care that is $560.7 billion toward he MaR ltIh insu upper r jo ainc nt e premiums for their workers. Workers paid $284 1% .6 billion toward health 20% By Paul Fronstin, Ph.D., La E b mployee Benefit Research Institute, and M. Christopher Roebuck, Fronstin, Paul, and M. Christopher Roebuck. forthcoming. "Persistency in High Cost Health Care Claim." EBRI Issue where the quality or outcome of treatme17% nt is not dependent on the site of care. Under reference pricing, employers or other treatment settings may also indicate that the procedure does not necessarily require delivery in an HOPD. To $1,377 Da head ta a and M ches, X-ray e m tho achin ds e s are often used to diagnose these conditions, which would explain why the allowed charges all copyright and other applicable notices contained therein, and you may cite or quote small portions of the report 39% $1,400 Distribution of Lab Services, by Site of Service supported by evidence aV nd MitR a tru Im lo in w ly e D ne r s jo c cir n e e tsesnain ry g. The campaign includ es reco 130 mmend ations fro 84 m 2% v arious me d 46 ical societies. One ins ur40 anc % e premiums as well. Anot Mehe tabro l$3 ic 75 pan .6 w el as spent on out-of-pocket expe38% nses for he 69% alth care services. Ph.D., RxEconomics, LLC Brief. c ins onfir urem rs tw hiould s, w e p a ha y d a tfix wo ed c lin am icount ians ( or MD lim s)i tr e the vieir w c a ont nd ribut appions rove tour owa fin rd atl list he c.ost of a specific health care service, and ?221% are lower. Despit 11%e the allowed charges for imaging for low-back pain being much lower than for the other services provided that you do so verbatim and with proper citation. Any use beyond the scope of the foregoing requires EBRI’s 35% 35% 11% example comes from theT I m No oa ta g rt lin h g Ame for loric w-a bn acS kp p ine ain Society, which recomm 3 ,e 1nd 18s a gainst the 2 ,0 7us 9 e1% of imag 1 ing ,03 9 of the spine within the Lipid panel HOPD PO Lab 29% Potential Savings 9% 13 healt10 h p %lan members must pay the difference in price if a more costly site of treatment is selected. In context of this $1,154 Data $1 a ,20 nd 0 Study Sample 2 examined in Figure 9, the price differential was much higher: 168 percent higher on average in HOPDs than in POs. prior express permission. For permissions, please contact EBRI at perm 6% issions@ebri.org. However, studies have found that 20–30 percent of overall spending may be considered wasteful. Waste comes from first s 60% ix weeks of an acute Im a eg piis ng o d fo er o uf nc low om-pb lia cc ak te p da h in eain dathe che absence of red flags. More infor0 ma .2% tion can be found at $1,099 Fronstin, Paul, and M. Christophe Dru r gRoe testbuck, “Persistency in High-Cost Health Care Claim12% s: ‘It’s Where the Spending Is, As seen in Figure 3, a total of 25 outpatient health service utilization measures were constructed using relevant Total potential savings for each of the 25 health care services examined in the paper are shown in Figure28% 13. Employers study, a 30% reference price could be set at a level that corresponds closest to the cost of services in a PO. If patients chose ® ® 2% 56% This study makes usIe m of agitn he g 2018 IBM Marketscan Commercial Claims and Encounters Database (CCAE). The CCAE http (1) fa s:/ ilu /w re ws of w.cho caors eing de w liv ise er ly y., o( rg 2) / failures of care coordination, (3) overuse of low-value health care services, (4) fraud Stupid,’” EBRI Issue BrB ie lo f, od no c. ou 49 nt3 (A EmT plo ye A e B ene G fit L Re Asea NrcC h Ins E tit ute, Octob12% er 2019) 0% . procedure (and in some cases also diagnosis) codes from claims. Included were: and w $1o ,00 rk 0ers could reduce their spending on lab services by between 12 percent and 69 percent, depending on the type to receive services in a HOPD, they would pay the differe Fnc igure e in c 9 osts. The expectation is that patients would engage Report availability: This report is available on the internet at www.ebri.org 0% 3 database contains member enrollment information as well as adjudicated inpatient and outpatient medical and Ima Cgi olo ng no v sicao p N yon-MRI 403 372 32 and abuse, (5) administrative complexities, and (6) pricing failure. Overall, employers and workers spent $244.2 billion Vitamin D screening 21% ?172% 8 https://www.ebri.org/docs/default-source/ebri-issue-brief/ebri_ib_493_highcostclaims- Primary Primary Primary Autoimmune Multiple Sclerosis Multiple Sclerosis Supportive Cancer Autoimmune Average Allowed Charge by Site of Service and Weighted Percentage Price Fronstin et al. (2020). of lab service, if price differentials between HOPDs and other sites of treatment were eliminated. with the recommending physician and choose the site of treatment based on not just recommendations but what it pharmacy Im cla muim nods. efi cW iene cy cons Immut nr ouc defitce ied nc yan a Immna unoly det fiic ciea nl d cy aDis taeset of ase Druga #d 1ults (a Drug ge #s 18 1 –64) Drug who #2were cont Drug inuously Dis enr easolle e Drud g # in 2 50% 20% C Ch oe lo sn t o X s-c ra oy py 681 432 2% 249 48% t o $366.3 billion on these six domains of waste. $760 • Five laboratory tests: metabolic panel, lipid panel, drug test, complete blood count, and vitamin D test. $800 24oct19.pdf?sfvrsn=ced83c2f_6. D iffer 47% ential for Various MRI Services Drug #1 Drug #2 Drug #3 I t is estimated tha $707 t 20–30 percent of overall health care spending may be wasteful. Overall, employers and workers would cost them in out-of-pocket costs. In fact, research has shown that the influence of referring physicians is larger 9 employer-sponsored insurance in 2018. Members in capitated plans were excluded. A total of 10.9 million individuals Author estimates from M C the h Re I s 20 stp X i20 n -r e aAnn y ual Social and Economic Suppleme 391 nt to the Curre 285 nt P7% o pulation S 106 urvey. Imaging • Five imaging services that are either always or sometimes conducted via magnetic resonance imaging (MRI): When it comes to imaging, savings could be as high as 56 percent for chest X-rays, 49 percent for echocardiogram, and Notes: HOPD=hospital outpatient departments; PO=physician offices. 14 Tabl spent e $2 44 of .2 b Co illnt ion to $ ents 36 6.3 billion on six domains of waste: failures of care delivery, failures of care coordination, than the influence of cost sharing even when out-of-pocket costs are significantly high. met these criteria. MRI upper joint 207 147 60 Mammogram 14% P$6 ricing 00 Source: Fail Authors ur ’ an eal ysis of IBM MarketScan administrative enrollment and claims data. Fronstin, Paul, and M. Christophe Cor lo Roe noscb ouc pyk, "Reference Pricing for Health CareHO Se Pr Dvices: A PO 15% N Pre icw e Di Tw fferenti ist al on the Defined 10 spine, upper joint, lower joint, imaging specifically for low-back pain, and imaging specifically for 41 pe 10 rc %ent for DEXA scanning. For MRI-related imaging, savings amounted to 39 percent for imaging for low-back pain, Koller and Khullar (2019). overuse of low-value health care services, fraud and abuse, administrative complexities, and pricing failure — or when Introduction ......................................................................................................................................................... 5 M Ec Rh I o lo cw ar e dri o jo gir n a tm 414 286 3% 128 40% Chest X-ray ?158% 56% $1,40C 0ontribution Concept in Employment-Based Health Benefits." EBRI Issue Brief, no. 398 (Employee Benefit 37% uncomplicated headache. ?35% Sample averages for the following chara 37% cteristics are reported in Figure 2: gender, age, relationship to policyholder, 34 percent for imaging for uncomplicated headache, 28 percent for MRIs of the spine, and 25 percent for MRIs of One H ow of Do the lim All ita o tions we d Ch of refear renc ge e s p rDiff icing e isr t ha by S t hea ite lth ca orf eTr p $378 rov eat idem rs e bnt elow ? the reference price may increase prices to 11 ?92% p rices migrate from what is expe ?234% cted in a well-functioning market. Ac $4 cor 00ding to Berwick and Hackbarth (2012), pricing failure occurs when “ 35% prices migrate far from those expected White and Whaley (2019) M DE am X .A m s oc ga ra nm ning ?89% ?91% 711 707 2% 3 MRI spine 28% Data andRe Me sea thod rch I s ................................ nstitute, April 2014). ht ................................ tps://www.ebri.org/h ................................ ealth/publications/iss................................ ue-brie$295 fs/content/refe................ rence- 6 and plan type. • Seven other (non-MRI) imaging procedures: colonoscopy, chest X-ray, mammogram, echocardiogram, DEXA upper and lower joints. Savings for mammography would be as high as 17 percent, and for colonoscopy they would be $259 33% ?139% at or close to the reference price, offsetting some of the savings. This ma$256 y happen to the degree that there is a Allowe 0% d charges were higher in HOPDs than in POs and other sites of treatment for all 25 health care services ?92% 32% in well-functioning E C m c aa h ro r ok tc ie d a t ra s, drito e t g rha r ya m s t te isn, otshe is sa cc re tua enil c ngosts of product 1ion p ,476 lus a fair 1 p ,1r2 ofit 9 1% .” In prior 347 research, we $207 31% MRI upper joint 25% 12 6 pricing-for-health-care-services-a-new-twist-on-the-defined-contribution-concept-in-employment-based-health- See http sc:a // nni vbid ng, MR ceI S nte cpi ar. r not eorg id /fre artq eue ry ntly st MR e- nosi a I Upp ske s sc er J d-q oi rue e ntesnin tions g, /a nd MRe I L le oc wer troc Joia nt rdiogra Im m ag ( ing EK for Lo G) s w B cre ac eknin Paig n spe Imagi cifi ng cfor ally Un for compl low icate -r disk $1,200 $159 15 In th pe is r cIe ss nt ue . Brief, we focus on that last area of waste: pricing failure when it comes to lab; imaging; and special com Dp ata and e $2 ti00 tive m Stud arke yt Sa and mp trle an ................................ sparent pricing inform ................................ ation. ........................................................................ 6 examined in this study. Findings on price differentials for lab services are shown in Fig29% ure 8. Two types of data points 29% Imaging for low-back pain 273 184 89 e 30% xamined paymentE d Kiff Ge src ernt eein ails ng for for onc low-olog risk pya tm iee ntd sications by site of treatment for t4% he 37 most prescribed Fig $1,111 ure 2 $100 Headache $1,108 MRI lower joint 25% $68 benefits-5378. patients. $49 $1,068 m 13edications for conditions such as multiple sclerosis, rheumatoid arthritis, and other inflammatory disor 26% ders. $29 This are shown: the actual allowed charges by site of treatment and the price differential in percentage terms. To calculate Fronstin and Roebuck I( m 2014) aging. for uncomplicated headache 85 62 23 Outp med atient S icationserv thaic t es tog ................................ ether c M aa pm tur me og dr a 92 m pS e ................................ r acm ent pl e of Cth ot aarl c ach te em ris ot t................................ ihe csr,a 2 p0 y1 us 8 e and spending ................................ 17% (Fronstin, Roebuc .............. k, and 7 I f site-of-treatment price differentials for specialty medications were eliminated, employers and workers would save $0 Anot • he r E thin ight g t non o con -onc sid olog er is y , w inf hic us h p eda spe tient cia s w ltyill m be ed sensit ications ive : t po rim the ar y com imb mina unod tion o eficf p ienc ricye dtrrug ans# pa 1r , ep nc rim y a arnd y cost sharing analysis is important because care is shifting from physicians’ offices (POs) to more costly hospital out $976 patient Notes: HOPD=hospital outpatient departments; PO=physician offices. Numbers do not add to 100% because sites of service coded as the p$1 ric ,00 e 0differential,S a pe D bc E le iX and A lt y s e c M da e n pdi n riic n cg e a tfor ion s non-HOPD settings w as us e 178 d. The blende 142 d pric e was we 36 ighted by the Stuart 2020 Col) on . os W ce o py found t Ch ha es Etc , t X hha o -Ra cd a y rd hospit iograMa m al o mmut ogp ram atient E dce hp oca ardi rtm oge ram nt (HOP DEXD A) S uni cannti ng prices 49% Ca m roa tit dc A he rted ry phy Esi KG cia Scn o reen ff ing ic e for Fr 14 onstin, Paul, M. Christopher Roebuck, and Bruce C. Stuart, "Location, Location, Location: Cost Differences for Chernew et al. (2019). betAn wealysis en 4 "p oth ................................ eer" rc e are nt no at nd sho36 wn. percent ................................ , depending on the ................................ medication. ................................................................. 8 changes. As shown in this paper, the cost of lab services and certain imaging are relatively low, while the cost of MRIs immunodeficiency drug #2, primary immunodeficiency drug #3, autoimmune disease drug #1, multiple 19% Stenosis Screening Low-Risk Patients departments (HOPDs). Compounding this shift in care is the fact that prices for hospital-based outpatient care are p erce 20% ntage of services p C Pr a riov rm ota ide id ry a d irm t e in e m ry u s na to e c d n h e o fis sett ciis e n sing. c cy re d e rn uig n g#1 141 112 2018 0.0 1% 29 Source: Authors’ analysis of I Im BM a g M in ag rk e fo tS r cla on w a -b dm ac ink is trati pain ve enrollment and claims data. 39% 18% (PO) prices, holding drug mix and treatment intensity contact, third-party payers would have saved $9,766 per Oncology Medicines Based on Site of Treatment," EBRI Issue Brief, no. 498 (Employee Benefit Research Notes: HOPD=hospital outpatient departments; PO=physician offices. a 15nd specsc iale ltyr osis med d ic ra ug tions #1 , am reult reiple lativ sc ely le rhigh. osis dE rv ug en w #2it , h i supp nfor orm tiv ae tion ca, nc pe ar tie dnt rug s m , aand y not aut see oimk m out une t he dis e low ase est dr c ug ost # 2. increasing faster than physician prices. Ultimately, employers and workers bear the brunt of cost differences when W he Fron res D tin o a Pnd lan M Roe eb m uc bk e r E (s Re 2019) KG s cc e;riv eFron e en iTheir n15% g s tin, for H lR oe w oa -e rlt b ish C u kc p ka a , tra ie end n Se ts Srtua vicrt e s? (2020) ................................ . 5 ................................ 4 1 ................. 9 Source: Authors G ’P an ernd im alya e srr is y of im IB m M uMa nod rke etS ficc ie an n c ad y m dir ni usg tr ati #2 ve enrollment and claims data. 0.01% Imaging for uncomplicated headache 34% use of these medicines in 2016, a savings of 45 percent. In the aggregate, employers and workers would collectively save $11.2 billion if price differentials between HOPDs and $80I 0nstitute, January 2020). Figure 6 Figure 8 providers if the cost sharing savings is low. Hence, the potential for price transparency and higher cost sharing may be HOPDs perform serviceT s t Po rM itha m aa la lt e r yc a im n b mue n o pdre ov ficide iend cy in l drue gss #3 costly POs or 4 in st ,964and -alone 3 ,la 86b 1 48% 0o .r 0 1im %aging 1,1fa 03cilities. DEXA scanning ?168% 41% How Do Allowed Charges Differ by Site of Treatment?........................................................................................... 11 other sites of treatment D w istri ereb e ulim tioina n o te f d V ari (Fio gu ur s eN 1o 4n ).- M Em Rp I loy Imag ers w ingould Ser sa vices, ve $9 b b y ill Si ion o te orf 80 Ser pv eirce cent of the total, Figure 11 These specialty med A ic va erag tions a e A rlel ous we ed d tC o h tr arg eate cb ond y Site itions o f su Serv ch ais m ce an ultd iple W eig scle h rted osis , Percen rheuma tag toid e Pri artce hritis, and $657 other limited. And, we have argued in the past that high-cost claimants may not be sensitive to changes in cost sharing Female 52% 10% Autoimmune disease drug #1 0.09% Caro tid artery stenosis screening 37% Fronstin, Paul, M. Christopher Roebuck, Jason Buxbaum, and A. Mark Fendrick, "Do People Choose Wisely After Average Allowed Charge by Site Figu of Ser re v 1ice and Weighted Percentage Price whereas workers and t$582 heir dependents would D sa iffer ve en $2ti .2 alb fo illion o r Lab r $582 Ser 20 p ve ices rcent. Workers and their dependents would inflammatory disorders. They are usually admini HO stP eD red PO by a he Amb au lt la h ca tory S re ur gp icral ofessional through an intrave15 nous (IV) $565 Key Pote Fin nt$6 ia00 l Sa dinv gings s: L o ................................ cation matters... ................................................................................................................ 14 because many of these patients reach not only their deductible, but also their out-of-pocket maximum. However, Spe M cu ia ltlitp yl e M se cd le ic ro as tiis o n ds rug #1 0.02% EKG screening for low-risk patients 3% Satisfying Health Plan DeductD ible iffer s? en Evti ide al nc foe r V Frar om io u the s Sp Use eci of alL ty ow M-ed Vailue cati H oen as lth Care Services." EBRI Issue National Health Spending: Employer and Individual Spending r ealize 33 percent of the savings in labs, 22 percent of the savings in imaging, and 3 percent of the savings in specialty HOPD PO Lab Price Differential injection or infusion. 100% Age, years $472 42.0 $300 ?78% 96% reference pricing mightP a rid m darre yss im tm his un o lid m eit fic ait eion. ncy d The rug #c1ost sha ring paid 173 by a patient 169 who e xceeds the 5 reference price can Multiple sclerosis drug #2 0.02% Conclusion and Implications for Employers and Insurers ................................................................ $10,078 ........................ 16 Brief, no. 516 (Employee Benefit Research Institute, October 2020). https://www.ebri.org/docs/default- on Health Insurance and Health Care, 2018 med • ic $4 $11,000 aT 00 to ions em . ployers and employees: 18–24 16% exceed the in-network P m Sru a im pxi p am o ry rtum iiv m e m cout a un nc o -e d of re d - fip c ru ioc e gnk ce yt d lim rugit # 2 as that amount 201 is con sidered 196 out0 -.of 0 8-% network a 6nd does not have to 0% Specialty Medications $400 HOPD PO Home Pric (in bil e Diffe li rons) ential source/ebri-issue-brief/ebri_ib_516_lowvalhealth-22oct20.pdf?sfvrsn=56693a2f_6. Use of 90t% hese health care services varied considerably. Some of these services were quite common, while others were Re ferences ......................................................................................................................................................... 18 Metabolic Pane 25–34 l Lipid Panel Drug Test Blood Co 17% unt Vitamin D Screening $5,647 P Aru im toa im rym im un m eu d nio sd ee afs ic ei e dn ru cg y # d2 rug #3 40 390.0 3% 1 be counted in any accumulation related to the maximum out-of-pocket cost allowed for in network benefits, as Primary immunodeficiency drug #1 35% National Health The aggregato e saOv vings in erall, e spe mpcloy ialteyr s maend dicw ato ion rke s for rs w ould work e colle rs acnd tive the ly ir sa d ve ep $ e11 nde .2 b nts is illion if Ho ru ese lah p to iv rld ic eCo e ly d nlow tiff ribu e.r t ion e Tha nt $5,444 to ia t ’ls s b be ec tw aus eeen ?65% quite rare. For instance, 38 35–44 percent of the sample had lab work for a lipid panel. In co 21% ntrast, only about 1 percent of $350 $6,500 $250 Expenditures, $3.6 trillion Autoimmune disease drug #1 723 680 43 prescribed by the AffS or ou dra cb ele : E C mare ployA ee ct B . e nefit Research Institute estimates based on adminisE tmp ratiloy veer -Sponsored Primary immunodeficiency drug #2 14% I nstitu 80 te % of Note Me sd : ic HO ine PD= of hostpi he tal ou Na tpa tiona tient de l Ac para tm den em ts; P ies. 201 O=physic 0. ianTh offe ic es H .ealthcare Imperative: Lowering Costs and Improving $178 individuals on specialty medications are typically part of the small percentage of the population who use a large $200 HOPDs a 45–54 nd other sites of treatment were eliminated for each of the24% 25 health care services examined the sample had imaging via MRI, and very few people used any of the specialty med Pic riva ate tions Heal te h xa Insu m rance ined. This is no Source: Authors en’r o an llal m ye sin st of a IB nd M cMa laim rketS s d ca an ta ad . ministrative enrollment and claims data. Multiple sclerosis drug #1 $224 267 259 8 Primary immunodeficien? cy59% drug #3 4% Outcomes. Edited by Pierre L. Yong, Robert S. Saunders and LeighAnne Olsen. Prem Wiu ash msin , $284.6 gton, DC: The National 55–64 23% p Figur ercente ag se of the health care. in this report. surpr$is 30 e0, as these medications are used to treat chronic conditions that are often rare, such as autoimmune diseases 70% Fi nally, instead of introd Muc ultin ipg le fin scla enc rosia isl inc drug e nt #2ives, which ma y or ma y 567 not work, em 553 ploye rs and ins 14 urers could move Autoimmune disease drug #1 16% Academies Press. and multiple sclerosis (MS), which has a prevalence rate of about 0.1 percent in the United States, and not all MS $200 patients fr $0 om HOPDs to Sot uphe porrt iv si et e ca s of ncert rde ru ag tment by removing the HOP 906 Ds fr om the 892 ir net work. Suc h a 14 n arrangement is Multiple sclerosis drug #1 26% Figure 1, National Health Spending: Employer and Individual Spending on Health Insurance and Health Care, 2018 ..... 5 The distribution of treaP tm ere son nt C sett ove ings for red MRI services is $248 shown in Figure 5. MRI services were provided mainly in only o Employers would save $9 billion or 80 percent of the total, whereas workers and their dependents 60% MRI Spine MRI Upper Joint MRI Lower Joint Imaging for Low-Back Pain Imaging for Uncomplicated patie$2 nt50 s are on such medications. Autoimmune disease d 55% rug #2 55% $173 306 55% 292 14 m Ana ostlys com ism on in staff-model health maintenance organizations (HMOs) but can be applied 54% more generally to any Koller, Christopher F., and Dhruv Mu K lthul iplela sc r. le20 ros19 is . dr"uThe g #2Com Emp me loy rc er ia & l Differential for Hospita 36% l Prices: Responses From Headache Policyholder 59% 2 settings: HOPDs and POs. Those claims that were coded as “Other” were likely miscoded. POs accounted for 54–61 would save $2.2 billion or 20 percent. Figure 2, Sample Characteristics, 2018 ................................ Worke ................................ r .................................................... 6 Total 3 ,183 3 ,080 104 Notes: HOPD=hospital outpatient departments; PO=physician offices. network plan. Providers could 48% respond by lowering their prices so that they may return into the network. This strategy Supportive cancer drug 32% States and Employ Spe or us." se Journal of the American Medical Association E1-2. doi:10. 25% 1001/jama.2019.9275. H 50 igh % Us ?84% ers of Health Care Services $199 Spending on For each of th 46% e 25 services, we calculated the proportion of utilization (claims or units for medications) delivered in percent of MRIs for 4 of the 5 services examined. However, HOPDs accounted for the majority of imaging for It is importS aou nt r ct e: o Anot uthors e ’t an ha alty ssi isx ofof IBM the Ma rhe ketS aclt an h ca admrie ni sse trati rv ve ic en es e rollm xa enm t an ine d cd lai m in t s da hi ta. s report — vitamin D testing, imaging ?15% $1 $250 00 Health 43% Autoimmune disease drug #2 24% ha s its limitations ? as w 531% ell.C I hti ld m /oa th ye not r dep w enor dek n twell in areas with limited provider choices. 15% Figure 3, Use of Var 42% ious Health Care Services, 2018 ................................................................................................. 8 • When it comes to a wide variety of health care services: each of the following sites of care: HOPD, PO, laboratoryInsu , am rance bula and tor y surgical center, home, and other. Of course, the unc omplicated headaches. ?143% specifically for low-back pain, imaging specifically for uncom 39% plicated headache, DEXA scanning, carotid artery stenosis 38% Source? : A 62% uthors’ analysis of IBM MarketScan administrative enrollment and claims data. 40% Health Care Shr ank, WIlliam H., Teresa L. Rogstad, and Natasha Parekh. 2019. "Waste in the US Health Care System: Estimated In prior research, we found that 10 percent of the population accounts for 70 percent of health care spending 36% Allowed charges for three sites of treatment for non-MRI imaging services are shown in Figure 10. While the allowed 34% a lternative places of service to HOPDs differ across procedures, for example, medications administered in HOPDs vs. Figure 4, Distribution of Lab Services, by Site of Service ................................................................ ?60% ........................... 9 screening, and EKG screening specifically for low-risk patients — are considered to be of low value. The Choosing $150 Type of Health Plan Costo s a nd Em Pp ot loy ent eia rs a l for nd Sa wor vin kg es." rs cJA oul MA d r 32 ed2 (1 uce 5 the ): ir 15 spe 01-nding 1509. on doi:10.1 lab ser 00 vic 1/j es am as m a.201 uch a 9.1s 39 69 78 p . ercent, depending (Fronstin a $101 nd Roebuck 20 So 1u9) rc.e :W Ait uhin thor s th ’ a is n a gly roup sis o , f 50 IBM –60 Ma p rkeertc Secnt an not adm only inistr a re tiv a ec he enrd o lltm he en irt deductible, but they also Similarly to the other imaging services we examined, those not provided via MRI were also provided in only two charges varied significantly by type of imaging, allowed charges for HOPDs were consistently above allowed charges for 7 ?246% POs, or lipid panels in HOPDs vs. labs. We then derived the average cost using the allowed amounts for each service by $118 30% Wisely campaign — which identifies wasteful health care spending — suggests tha ?402% t any use of these services is of little $100 HMO/EPO 12% $108 and claims data. on the type of lab service, if price differentials between HOPDs and other sites of treatment were Figur re e a5c, hD ed is ttrhe ibut ir ion o maxim f Va um ri ous MRI out-of-p oc Sekre vtic lim es, b it. yA Sit s ae r of esu Se lt, rv m icuc e h o ................................ f the healt $103 h care the ................................ se people use are su............ bject to 10 P sett Osings, . And w in t ith o hene ca e se xcof epc tion. olonosc Colonosc opy, a $99 op llow y ser ed v cic ha ers w ges for ere s p Hlit OP ne Ds arw lye r ee v e anly bov b ee a tw llow een e dH c OP haD rg s a es for nd a m am bb ula ula tor tor y ysu su rg rg icia ca l l ?73% 8 $79 site of care. Average HOP $93 D costs were subseq $93uently divide $1,220.9 d $77 by average non-HOPD costs to derive the utilization- ?60% t W o hit no e $1 , c00 lin C., ica and l vaC lue . W . ha To let yhe . P20 P O d19 e /P gO . re S Per ic tha es P t H aOP id tD o s c Hospit harga els m b or y eP r for iva tthe e H se easer lth P vicla ens s t ha Arn o e H 55% tigh her Re sitla es of tive c to arMe e, d the ica se re se and rvicVa es ry eliminated. low or no cost sharing, which lowers the savings potential for these plan members from any reduction in price c C eo ntnc er 20 s ( lus %Figion and ure 6). Only I m 8 p pli ercca entt of ions colonos for c op Em ies w ploy ere e prrov s ide and I d in Pns Os.ur Ote he rs rw ise, chest X-rays and mammograms centers as well. Overall, allowed charges were 35 to 234 percent higher in HOPDs than in POs. weighted average in all other sites of care. Hence, results under 100 percent reflect lower HOPD costs relative to non- Figure 6, Distribut$66 ion of Various Non-MRI Imaging Services, by Site of Service ........................................................ 10 HRA $64 15% would fall into two categories of $62 waste: ov $61 eruse of low-value health care serv $57 ices and pricing failure. Widely: Findings from an Employer-Led Transparency Initiative. Santa Monica, CA: Rand Corporation. $54 $59 $54 $53 differentials. Employers would continue to realize savings if these plan members switched from high-cost to low- were more likely to be provided in HOPDs than POs, while echocardiograms, DEXA scanning, and carotid artery stenosis O ur findings have implications for both employers and insurers. There are a number of actions these third-party payers HOPD, whereas numbers H eS xc A-e ee lig diing ble h10 eal0 th p pe larn cent represent higher HOPD costs relativ16% e those in the non-HOPD settings. $50 $50 8% $36 doi:10.7 $2,428.5 249/RR3033. $34 o Savings could be as high as 56 percent for chest X-rays, 49 percent for echocardiograms, and 41 10% $35 Fi The guraellow 7, D ed is tcrha ibut rgion o es shf Va own i rious Sp n Figur ee c ia 11 lty for Me spe dica ctia ions lty , mbeyd ic Sit ae tions of Se rerflect vice ................................ the per-unit price of the ................................ medication. The 11 cost providers, because employers pay for nearly all of these claims. $31 screening were more likely to be provided in POs than in HOPDs. EK $22 G $21 screening for low-risk patients was ne $21a$21 rly always c an take. First, they can exert pressure on hospitals to shift from discounted charge contracts based on a multiple of Source: Employee Benefit Research Institute estimates based on administrative Out-2% of-Pocket All Other Spending percent for DEXA scanni $19 ng. largest price differe $15 nce per unit was for primary immunodeficiency drug #1, which was 84 percent more expensive We also estimated the ep not role lm nt eia ntl a ang d g cr la eig ma st e da sa ta $13 .vings to the U.S. health care syst Pe riv m ate unde Business r a scenario whe Heal re tnon h -HOPD provided in onP Hea Os. lth Insurance $11 $11 Fi Me gd ur ic ea r 8e , A to ve som rage e ot Allow her epdr os Cha percgte iv e b yc a Sit se e rof ate Se . E rv m icpeloy and ers c Would eight a els d o Pe erxe cernt t su agceh Pprric ee ss D ur iff $9 ee on rent he iaa l for lth p Lla ab ns Se to rvdic o etshe ...... sa m 11 e Winn, 0% A. N., N. L. Keating, J. G. Trogdon, E. M. Basch, and S. B. Dusetzina. 2018. "Spending by Commercial Insurers $0 EBRI Issue Brief is registered in the U.S. Patent and Trademark Office. ISSN: 0887 -137X/90 0887 -137X/90 $ .50+.50 Spending, Contribution to Employer- and Health Care when administered at an HOPD vs. at a PO or at home. It is important to recognize that for many of these medications, Co Prim lon ary oscopy Prim Chary est X-Ray Primary Mammogram AutoimmE un ch eocardi Mog ultir pl am e Sclero DE sisXAM S ulctia pl nn e Sc ing lerosis Ca Sup ropo tidrti A vre teC ry ancerEKG Auto Scirm een mun ing e for p rices prevail for all 25 health services. More specifically, we “repriced” all HOPD utilization using average-weighted $375.6 Sponsored Private Health with th $0 e hospitals in their networks. A coalition of employers across Illinois, Iowa, and Wisconsin, is already trying on Co he mIot f sit hee ra -of py- tB re aa sed tm eon nt Sit pric e eof diff Ca erre e." nt JA ials MA for Onc spe olog cialt yy E m -1, E edic-a 2. tions were eliminated, employers and Immunodeficiency Immunodeficiency Immunodeficiency Disease Drug #1 Drug #1 Drug #2 Drug Disease Drug #2 Stenosis Screening Low-Risk Patients Figure 9, A© ve2 ra 0g 2e 1 ,A E llow mplo edy C ee ha B renef ge bit y Resea Site of rc Se h In rvic se tit a ute nd - W Ee du ight cae tio d n Pe arnd ce nt Resea ag Insu e rc P rance ric h F e P un D remiu iff d. er ms All r ent , ia ig l for hts Va resrer ious ved. MRI When it comes Me to tabsi oltie c P of anetlreatment, spe Lipid P cia an lty el medications c Dr aug n b Tes e tfurther group Be loo d d Co into unttwo categor Vita ie mi s: n D S the cr eten hrie nge primary multiple units will be used per year (Figure 12). For example, the average patient using m ultiple sclerosis drug #1 Drug #1 Drug #2 Drug #3 non-10 HOPD prices and tN he otn a e: Hg Mg O re =g he aatle thd m the ain te on baser ncev o erd g asp nize anding tion; E P dO iff =e er xe cnt lusia ive ls p for rovi da ell indi r organ viidua zation ls ; in the sample. We then this. $560.7 workers would save as much as 36 percent, depending on the medication. Notes: HOPD=hospital outpatient departments; PO=physician offices. Numbers do not add to 100% because sites of service coded as Services ................................................................................................................................................ 12 im munodeficiency medP ic Pa O tio =pns ref ea rrnd ed p the rovi dot erhe org r afiv nize a tm ion e;d Pic Oa St= ions poin.t The of sertvhr icee ;e H R im Am =hunod ealth re efi im cie bunc rse ym m ene t dication infusions were 9 — which wNote as p s: rHO iceP d D= only hospi tal $1 ou 3 mor tpatiene t de wpa he rtm n ta entsk ; P en O=a ph t ya sn ician H OP officD es v . s. at home — took 2,566 units per infusion session. In e xtrapolated S Note ou to rce: st: he HO National n PD= ationa ho sHeal pital l le tou h v Ex tpa el usin penditur tient de gpa es, a rn e tm hten tps: st ts/im ; P /wO w aw =tph e .cm yof ss. icg i12 an ov /8 of Resear .5 fices m . ill chion -Stata istd ics ult -D sat w a-it and h e -Sm ystp ems/ loye Sr tat -sponsor istics-Trene ds d- and ins-urance in 2018. "other" are not shown. Source: Authors’ analysis of IBM MarketScan administrative enrollment and claims data. arrangement; HSA=health savings account. R Sep ouort rce: s/N Aatio uthors na’l H an ea ally thE sisx of pe IB nd M Data/ Mark NetS atio cna an lH ad ea mlitni hA str ccou ative ntsHi enrol st lm orica ent lan d claims data. often admini Sou st rc ee: re Ad uth in a ors’ an pa alt yie sisnt of’ IB s hom M Mare ketS (40 can – 63 adm p inie sr trc ati evnt e en of rolt lm he en t tan im d ec) lai (m Fsigur data. e 7), whereas the other five specialty contrast, patients using the much more costly supportive cancer drug only used 4 units per infusion session. e e e e e e e e e e e e e e e e e eb b b b b b b b b b b b b b b b b br r r r r r r r r r r r r r r r r ri. i. i. i. i. i. i. i. i. i. i. i. i. i. i. i. i. i.o o o o o o o o o o o o o o o o o or r r r r r r r r r r r r r r r r rg g g g g g g g g g g g g g g g g g IIIIIIIIIIIIIIIIIIs s s s s s s s s s s s s s s s s ss s s s s s s s s s s s s s s s s su u u u u u u u u u u u u u u u u ue e e e e e e e e e e e e e e e e e B B B B B B B B B B B B B B B B B Br r r r r r r r r r r r r r r r r rief ief ief ief ief ief ief ief ief ief ief ief ief ief ief ief ief ief A re • • • • • • • • • • • • • • • • • • se Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb Feb arcr r r r r r r r r r r r r r r r r rh u u u u u u u u u u u u u u u u u u a a a a a a a a a a a a a a a a a a re r r r r r r r r r r r r r r r r r ry y y y y y y y y y y y y y y y y y p o 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1r8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 t ,,,,,,,,,,,,,,,,,,f r 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 m 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 t h • • • • • • • • • • • • • • • • • • e N N N N N N N N N N N N N N N N N N Eo o o o o o o o o o o o o o o o o o B..................R 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5I2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 E 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5d ucation and R esearch Fund © 2021 Employee Benefit Research Institute 19 17 16 10 18 15 13 14 12 11 9 7 8 6 2 3 5 4