One of the strongest trends in employment-based health benefits has been the adoption by employers of high-deductible health plans (HDHPs). As a result, the percentage of individuals with private insurance who were enrolled in an HDHP increased from 17.4 percent to 46 percent between 2007 and 2018. Further, deductible levels have been increasing more generally regardless of whether someone is enrolled in an HDHP.
Recent research found that many high-cost claimants not only reach their deductible but often reach their maximum out-of-pocket limit. In essence, for these individuals, use of health care services early in the year affects what they pay for health care services at the end of the year. Because deductibles are often satisfied early in the year by high users of health care services, it can be argued that it is the end-of-year price that matters most when individuals are deciding whether to use a health care service even before the end of the year is reached. This phenomenon raises questions about the effectiveness of deductibles in controlling spending. But it also raises questions about the relative effectiveness of different types of cost sharing later in the year — such as copayments and coinsurance — once an individual reaches his or her deductible.
The purpose of this paper is to examine the differential effect of copayments and coinsurance on use of health care services. Our findings have implications for the effectiveness of deductibles, particularly once they are satisfied. Why would we expect to find differences in behavior by whether an individual faces copayments or coinsurance after they reach their deductible? Prior studies often found that individuals are not forward-looking when it comes to prices for health care. The same may be true when it comes to type of cost sharing. Before an individual reaches their deductible, they will know what the price of health care is if they face copayments after reaching their deductible. Copayments are well-defined and known before health care services are used. Coinsurance is less well-defined when it comes to knowing the price of a health care service in advance of using the service. The uncertainty of coinsurance relative to the certainty of copayments may mean that coinsurance has a differential impact on use of health care services compared with copayments.
Key Findings:
- Coinsurance reduces use of inpatient care and specialist physician office visits more than copayments do.
- However, while most employers already use coinsurance for inpatient care, only 44 percent use coinsurance for office visits.
- This suggests that employers seeking to manage use of health care services and spending — especially among high users of health care services — may look to moving from copayments to coinsurance for office visits as a way to do so.
Figure 2 Figure 5 Figure 8 Average Annual Employee-Only and Family Deductible, Among Workers in Statutory Out-of-Poc ¨ket Maximum, The Fi Because we prescript W Healt Ref W Quan, H., g e hen ure 3, co re did ere ntrol hie it wnot e come ere d ion dr n Distr st A e P. H fo ces duc on x r s amine how ib ug alfon Sun ly a few re ge to the tibl ution s Most ne f es a o ral und that of ire o i mpac health Sen nst Heal dara co ft an en payments t of raj sit th Spen stat ces wh re an, A. ive t cos w ache en us via incl htco ere -di o C sd ea a Fo ha insur n coi nd coins g, A ng, ri os rly in the ng nsurance ance B usion of t Shar mo by . B ng uranc urnan h a Indivi ealth nd copa ing year by hi the he a d, J. stat d C a dF ff uals With Em a l harlso iym ecte g C. us, u arg rents e gh users Luthi, w d presc er i 7 n e h Com h m ave L. av pac ri e or p t D. Saund of ption dru loym t than a bi he he numb dis ty alth ent ame Index, copaym er o -ers g us Bas care ex , fpe e e C. A. B e d we s . ents x cted ervices Hea pe Tca he BPD ctat lth ( o nnot eFi ut ck , io gu Cov -, T it ca ns. of re d rul - .e po o E. 1 F n e rage, Co es not 0) ck irst F out b. e eas et cos , ar other Am we by, gu provid on ntinu e e ts and W. A. d g 18 x p , pt di o ect the hat e ous tential a cost – be 39 ittes -i-s ?? = ?? + ?? ?? ?? + ?? ?? + ?? ?? + ?? (3) Paul Fronstin is DD irector ema of nd t R he eHea spo lth n s Res esearch a to Cond Edu st Sha cation Pro ring, by gra Ty m p at th e of e H Em ea pllth oyee Ca Br ene e Se fit R rv esear icech Institute Managing Use of Health Care Serv ?? ?? ?? ?? ?? ices After People Satisfy Private-Sector Establishments With a Deductible, 2002–2019 6 by Type of Coverage, 2013–2018* patien confoun year s the least haring in e - h xpect ol ts ea ds, coin de G wer lthy t Enro form hali rs ed . e les . 200 end N o b llment ation f severtheless urance ha se - l of 5. " ikely to less - i or year n 201 Codin p sensi rd esc pr r , 7 a time g efi ice tive to cos g ription ................................ Al llreat p that gorit -invar resc er i d matte hms ru riptio m iant tgs. pac s fh or rs unob ns aring t than mos Def wh s,ining erva en regardl t copaym whe ................................ they bl Com n indi e ess charact faced ents or vi of biditi du on coinsura werist als heth es in inpat are ics er they I d are elimi ince, as com CD ent ec ................................ -id 9 days -iface C ng wh M and ICD , nate co spe pa et payments d cher red iali usin s-to t of with 1 g linea 0 A ufice s or e dmini thos ................................ a c visits r health c oinsurance. fix es w e trat d ,ith and effect ive cop are ser Daym ata. s.P as vice " et nts .. . 6 (EBRI). M. Christopher Roebuck is President and C CEO os of t Sh RxEco arinnomics g, 201,8 LLC. This Issue Brief was written with assistance As mentioned above, we measure health status using the Charlson Comorbidity Index (CCI). The CCI is a weighted Their Deductible: What Do Copa Employee-Only Coverageyments Family Cov an erage d Coinsurance Do? Aron-Dine, Aviva, Liran Einav, Amy Finkelstein, and Mark R. Cullen. 2015. "Moral Hazard in Health Insurance: Do even psy research chothera before Medi has f py/counseli tcal hound e Care end that o 42 f ng vi the t (1 hes 1): 1 ysear i e its pat . 130 sAm ients ar reache -9. ong 4 e no d0 . –F 64 t only ig-ur ye ear 4 likel -ol sho dys ws , to s coins that atis urance fy their average had d o edu ut a -cti great ofbl -poc e bu erket s it mpact are pen also than din l g ike c (i.e. opaym ly to reac , theents prh ice t on thei hat u r out se - from the Institute’s research and editorial staffs. Any v 2013 2014 ie20 ws e 15 xpress 2016 ed 20 17 in this 2018 report are those of the authors, and index $1 th6,0 at pr 00 edicts risk of death within one year of hospitalization for patients with specific comorbid conditions. It is Figure 4, Average Out-of-Pocket Spending Among PPO Enrollees for Selected Health Care Services, by Calendar Month, Overall Elasticities Coefficient Dynamic I ß is the nce mar ntives gina Matter? l impact of " T he the R coinsuranc eview of Ec e onom rate on ics Y and , hold Stat ing ist p ics co 97 nst (4): ant. In 725E -7 qu 41 ation . 3, t captures By Paul Fro 3 nstin, Ph.D., Employee Benefit Research Institute, and M. Christopher Ro i ebuck, of patien of-eme pock ts rge et ma pay) ncy xamo imum. depa ng rtm indiv Secon ent vi iduals d, sits re . e gardles nrolled s iof n a he pre alth ferre stat d provider us, we ex org pect t aniz hos ation e wit (PPO h coi )n fell f surance rom Jan to be uary mo throu re sensitiv gh Decemb e to er should not be ascribed to tD he ed of uficers, tr ctible ustees, or other sponsors of EBRI, EBRI-ERF, or their staffs. Neither E $14,700 BRI nor widely used in the extant literature as a gauge of general health status. Medical conditions such as diabetes, cancer, $4,000 2018 ....................................................................................................................................................... 7 November 19, 2020 • No. 519 $14,300 Copayments (Model #1) Coinsurance (Model #3) Result Me StataCorp. thods s 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC. indivi Da dual ta fixe Sour d effe ce ctss , ? it is the idiosyncratic error term, and the double-dot accent on ? distinguishes between cos in Ph. 2t018. D sharing t ., RxE In ot con han her wor omi thosds, ecs, with copaym pat LLC ient prices ents fa . ll throughout the year. If consumers are sensitive to those prices, it is Employee-only coverage $1,045 $3,655 EBRI-ERF lobbies or takes positions on specific policy proposals. EBRI invites comment on $13,700 this research. and heart disease are included. Overall, the CCI currently consists of 17 health conditions. We augmented the CCI $14,000 Brot-Goldberg, Zarek C., Amitabh Chandra, Benjamin R. Handel, and Jonathan T. Kolstad. 2017. "What Does a ® ® $13,200 Interestingly, amo Inpang tie indiv nt Daid ys uals under age 18, with the exception -0.05 of chiropractor visits — -0 wh .18ich is *** a relatively rare com Figure pan 5ion , St e atuto stimates ry Out in- Equat of-Pock ions 1 et Maxi anm d 2. um, By T ype of Coverage, 2013–2018 .......................................................... 9 For enro This study llees in ma plakes ns w us ith e of copaym the IBM ents , co Mark stets shar can ing Com for a g mer ive cial n sCla ervic ims e and Encou is a fixed dol nters lar amount Database (c (CCAE ). Under ) as w coinsu ell as rance, possible that the price sensitF ivity am ilis d y ciffere overant at dif ge ferent times of the year as well. $ T 1his raises ,858 a qu $3,396 estion $3,392 as to the $12,700 with an indic $3,500 ator for depression (CCI&D) and examined 3 groups of patients: those with CCI=0, those with a CCI=1, $12,500 Deduct ® ible Do? T® he Impact of Cost-Sharing on Health Care Prices, Quantities, and Spending Dynamics." The Sample Statistics W occ e urrence find exactl — y wh we fin atd we w that c ould oinsurance expect to find amo had no impa ng p ct o en ople with use of healt chro h nic care con service ditionss.. T Copaym hose patie 4ents n ts also who did wer not e affe leasct t the IBM Mar Eme kets rgecan ncy D Be epne artm fit e Pln an Des t Visits ign Database (BPD). 0.0 3 The CCA E database contains mem 0.004 ber enrollment however, member out-of-pocket cost is the product of the coinsurance rate (r) and the price (p) of the service. The two generalizability of the widel Iy used npatie e ns t timat Serve icof es -0.2 for the elasticity of demand for health care from the RAND Health Suggested citation: Paul Fronstin and M. Christopher Roebuck, “Managing Use of Health Care Services After People $3,069 and Mana those wit gih ng Use of Health a CCI=2 or more. Clearly, those Care with a CCI of Se zrvi ero us ces e very A few fter People health care services iSa n anyt gi isfy ven year. Figure $1 6 2,0 , S 00 ample Characteristics, 2018 .................................................................................................................. 12 Quarterly Journal of Economics 132 (3): 1261-1318. http://www.nber.org/papers/w21632.pdf. inp Introduc healthy atient, wer eme e tio less rgen n sensitiv cy depart e to cos mentt, or s sharin peg ci, w alist off hether ice i t vis be its copaym . However, ents or coinsuranc they were fou end (Fi to gure redu 11) ce. p Cons rimary ist care ent w vi ith sits our , Statinf ist orm ical ation Infere as w nce ell as adjudicated medical (inpatient and outpatient) and pharmacy claims. The BPD database is % with coinsurance 98% D cos emogra t-sharing r phicseg Pr of imes im the ar y s sho am Cauld not pl ree Ph for ysbe 2018 ic ia ex n are pe Offi ct in c ed eF to Vi igs n ur its ec e ess 6 an arily p d informat -0 ro.du 19ce *ion ** ide on ntical deman cost sharing d i responses $2,871 s i-0 n .Fi 11 gu **r *e , si 7.nce Asu you nder can see, Insurance Experiment (HIE) and other papers cited in Cutler and Zeckhauser (2000). More recently, there are a Satisfy Their Deductible: What Do Copayments and Coinsurance Do?,” EBRI Issue Brief, no. 519 (November 19, 2020). $3,000 Those with a CCI=1 use some health care services. And those with a CCI equal to two or more comprise the bulk of Their Deductible: What Do Copayments and Coinsurance Do? intended to provide numerical values for the main elements of health plan benefit design including deductibles, Fi ov ph gerall fin ys ure ical 7, thera Cos ding t py sSharin in Fig visits g, ure 8, , an 2018 d psy we fin ................................ chothera d that c py oinsurance visits. Cop ................................ ha aym d aents gre h ater ad th ime pact on larg ................................ est re imp duac cing t on inpati useent of................................ chir leng opracto th of sr serv tay and ices . 13 % with copayment $2,640 2% One coinsof urance the s, trcongest tr onsumer end cosst in sharing employ is men a futnction o -based fh ealth varia b tion i enef n its pr h ice as( p be ). en In th theory e adopt , members ion by em co plul od be yers of mohig re h or - less numbe cost sharing c r of pape om rs es in thatma look ny d at the ifferent quest sha ion pes and as to sizes. whether consumers Generally, afte are more r reaching ave or lessrage de myopic du incti terms bles of of $1 price ,04 5whe for n it Of critical concern Spec was th ialist Ph e fy act that sician O de ffis cp eit Vi e s havi its ng data on mil -0lio .03 ns of individuals, the pla -0 n desig .08*** n measures — the E the ndnotes 20 percent of the population who use around 80 percent of the health care in any given year. Buntin, Melissa Beeuwkes, Cheryl Damberg, Amelia Haviland, Kanika Kapur, Nicole Lurie, Roland McDevitt, and M. $10,000 $2,491 coinsurance rates, copayments, and maximum out-of-pocket (MOOP) amounts. This information was not 1 specialist office visits among patients with 5 a CCI=1. Among patients with a CCI equal to two or more, coinsurance had Copyright Information: TA his re verag port is c e coinsopyright urance ed by the Employee Benefit Resear 18% ch Institute (EBRI). You may copy, deductible health plans (HDHPs). As a result, the percentage of individuals with private insurance who were enrolled employee responsive -only unde coverag r coinsuranc e and $1 e com ,85par 8 for ed f w amily cover ith copaym aent ge,s i.ndiv iduals face a mix of cost-sharing arrangements. When it By com Pau es to use $2,50l 0 Fro of nsti hen alth , Ph care .D., services Empl.oye They f e Ben ind larg efit Rese e variation i arch n e In las stitu $2,322 ticities te that , an are d M. often Christop of a mag her Ro nitude h eb iguck, her than regressors of interest — only varied at the plan/year level. Ignoring this would have led to an overconfidence in our Physical Therapy Visits -0.25*** -0.23*** Figure 8, Demand Responses to Cost Sharing, by Type of Health Care Service ................................ ® ......................... 14 Susan Marquis. 2006. "Consumer-Directed Health Care: Early Evidence About Effects On Cost And Quality." $2,220 Overall, both coinsurance and copayments have a greater impact on individuals ages 18–39 than on those under age abstracted from specific plan design documents but was instead generated by IBM via a plan-specific statistical a greater impact than copaym Av ents erag e on co em paerge ymency d nt epartment visits and specialist o$304 ffice visits. print, or download this report solely for personal and noncommercial use, provided that all hard copies retain any and in an HDHP increased from 17.4 percent to 46 percent between 2007 and 2018 (Figure 1). Further, deductibles have cam the -e 0.2 to RAN inpat Dient HIE elast services, icity. mo st plan enrollees (98 percent) faced coinsurance instead of copayments. For those with est Ph. imates D., RxE of tcon he im omi pact o cs,f m LLC ember cost sharing on health services utilization. Unfortunately, neither the CCAE nor BPD 1 Chiropractor Visits -0.33*** -0.16*** HDHPs a Health A re health plans ffairs w5 wi 1th 6-w5 dedu 30cti . bles of at least $1,400 for individual co $1,975 verage and $2,800 for family coverage in 2020. analysis of claims and enrollment data. For example, in attempting to determine the member cost share for a 18 or ages $8,00 40 0 –64. Figure 9, Use of Health CarE em Se ervic rgees, ncy b D y e Age, 201 partme8 n t................................ .............................................................. 14 $7,350 C all opaym copyrige ht nt and Regi oth me er a Model pplicable notices contained therein, and you may cite or quote small portions of the report been increasing more generally regardless of whe $7,150 ther someone is enrolled in an HDHP. Among individuals with a coins$2 urance, ,000 the average was 18 percent. Copayments averaged $304 per stay. Figure 4 contain a plan-specific identifier. As a solution — albeit an imperfect one — we assumed that everyone in a given year $1,761 $6,850 primary car Ps e p ychys hoth ician erap (PCP) y/Cou visit for nseling a Vis sp its ecific employer, -0 the .09 d **ata * base developer wo -0 uld .10 analyz *** e all adjudicated 2 $6,600 $1,658 % with coinsurance 52% Buntin et al. (2006) examined some of the early evidence from a number of studies that looked at, among other things, the $6,350 provided that yoA u v do er $6,250 so ver age Ou batim a t-of-Po nd wit cket h Sp pro enpe dirn ci g tat Amo ion. Any u ng PPO seEn beyo rollnd ees the for s co Sel pe ected of 16 th e for Heale th goin Car g re e quires EBRI’s deductible, the average deductible amount increased fro Fi m $446 gure 11 to $1,931 from 2002 to 2019 among those w$1,931 ith Charlso who had n, id M. E entical ., P. valu Pom es for pei, K. L. Ales, a all benefit desi nd C. R. gn m MacKen easures zie. 19 were e87 nro . "A N lled in ew thMet e same hod o pf lan. Class Us ifyin ing this g Prognos approac tic Com h, we orbidity To Figure 1 examin 0, e De th mand Res e impact of pon m sember es to Cos out t Sh -of- aring pocket cos A ,T by A A ge t o n and h G ealt Typ Lh e A servic ofN Health e C s uE tilizat C are ion un Service de r a ................................ copayment regime,.......... we 15 $1,846 claims for that service within that plan. Since the claim line includes distinct fields for deductible, coinsurance, Figure 10 $1,808 Source: Authors’ analysis of IBM MarketScan administrative enrollment and claims data. impact o $6f movi ,000 ng to HDHPs on he % walth care s ith copaym pend enting. It found one-time spending reduct48% ions ranging from 4 percent to 15 $1,351 Services, by Calendar Month, 2018 prior exp $1,500 ress permission. For permissions, please contact EBRI at permissions@ebri.org. $1,696 employee-only coverage (Figure 2). And it increased from $958 to $3,655 among those with family coverage. Demand Responses to Cost Sharing, by Health Status Among Adults Ages 18–64 created specify t a he in s f ynt Lo oll ngit hetic owing linear udinal plan Studies: ide fn ixed $1,232 tifie Dev -re for ffects eeach lopment model: year (t and his corr Validation. esponds to " Journal g iof n Ch thero mod nic Di els sabove) ease 40 . We t (5): 373 hen -re 83. as signed and copayment (in addition to the plan paid and allowed amounts), it is possible to infer whether or not a $1,541 percent. In a more recent study Demand $1,143 Respo , Brot nse -Go s tld o berg et a Cost Sha l. (2017) ring, foun by Ad tha ge at nd sav Ty ings ra pe of nged He from 12 percen alth Care Sert to vic 14 pe e rcent, while Figure 11, Demand Responses to Cos Average c t Sh oins aring uran,c b ey Health Status Among Adults Ages 20% 18–64 ..................................... 16 *p< 0.05, **p<0.01, ***p<0.001. $1,079 Inpatient Emergency Department Primary Care Physician Visit Specialist Visit Prescription Drug individuals to their most recent plan identifier for all prior years to force panels to be $1,353nested within cluster. In all coinsurance $958 rate or fixed dollar copayment prevailed for PCP visits as well as its corresponding numerical value. Conclusion Fronstin and Roebuck (2013) found a 25 percent savings in year 1, but by year 4, savings $1,273 amounted to only 6 percent. Average copayment $162 One of the strongest trends in employment-based health benefits has been the adoption by employers of high- Report availability: This report is available on the internet at www.ebri.org Empl$1 oy ,00 $4 ers ,00 0 h 0 ave been increasing deductibles because it is one of the easiest plan design changes to adopt in order to $18 Cutler, David M., and Richard J. Zeckhauser. 2000. "The Anatomy of H $1,167 ealth I17 nsurance." In Handbook of Health $1,123 In this example, the developer may determine that th Fie s gurtat e 6 istical mode (e.g., copayment=$20) was the most multivariate models, we clustered standard errors by this new plan identifier. Of course, compared with not clustering, $1,025 Overall Elasticities Use of health care services O and utp ass atie ociate nt Ofd fic se pend Visiing ts are far from evenly distributed across the population. As a general de 3 ductible health plans (HDHPs). As a result, the percentage of individuals with private insurance who were enrolled in manage the cost of providing health benefits. When an employer increases the health plan deductible, it only has to Fronstin and Roebuck (2019). $917 Economics, by A. J. Culyer and J. P. Newhouse, 563-643. Ams Oterdam: verall ElaEls stievier. cities likely member cost share for a ?? PC = P ??visit +for ?? $869 that plan. ?? + ?? Oth ?? +er tec ?? hniques would be (1) used to derive deductible and p-values were substantially higher, but we ?? b Sa elieve m ?? ple thi C sh ca ons raervat cteriive a sticp sproach to , 2018 be more appropriate. % with coinsurance 44% rule, 20 percent of the population accounts for about 80 percent of health care spending. Many of these high-cost an HDHP increased from 17.4 percent $714 to 46 percent C bet opaween 2 yments 007 a (Mod ned l # 21 018. ) Furth Coer, dedu insuranc ctibl e (Mo e levels del #3 hav ) e been $500 chang $1 e 6one number. It is much easier to increase deductibles than to do things like change insurance carriers, alter $2,000 $652 4 maximum out-of-pocket measures. Of course, the accuracy of this approach depends on the size and fidelity of The elasticity of demand is t $573 he measure of responsiven Coess pay of co ments nsu (Mo mers to a chan del #1) ge in Coia product’s nsurance (Mo pricd e. An elastici el #3) ty of $518 2018 % with copayment 56% In claimants short, c rea oinsurance ch their re dedu ducti ced bluse e. B of ecause inpatide ent ductib care les ar and se pe ofte cialn istr p eache hysician d earof ly i fice n th visi e y tsear b more tha y high n u cs opaym ers of health ents. care increasin g mo $446 re generally regardless of whether someone is enrolled in an HDHP. T Deyo, R. able A., D. of Contents C. Cherkin, and M. A. Ciol. 1992. "Adapting a Clinical Comorbidity Index for Use with ICD-9-CM networks, mo CCve to a I and D h ep igh res -p serformin ion = 0 g network, move to a limited network, change the formulary, offer health-risk the underlying data. For this reason, BPD developers set confidence thresholds (opaque to end users) below demand for health care of -0.2 means that a 10 percent increase in health care prices results in a 2 percent decrease is use of Elasticities of Member Cost Sharing Under A G ge e n 1d 8er Average coinsurance 20% Otherwise services$0 , it ca , con payments be argue a d nd coins that it is the uranc e e h ndad -ofa -year bout a pri n e ce qu that al e matters ffect on the u most when se of the indi vid other uals are healt decidi h care services ng wheth er to $14 Administrative Databases." Journal of Clinical Epidemiology 45 (6): 613-9. assessments, offer financial incentives for biometric screenings, audit the plan for non-eligible dependents, etc. Introductio which d n In ata ................................ pa val tienues w t daysere set to miss ................................ ing. As one would ................................ expect -0.1, the 3 degree of ................................ missing data is sm -0.72 aller .......................... for PCP and 4 health care ser $0 vices. 2002 2003 Mal20 e 04 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 49% 2016 2017 2018 2019 Average copayment, primary care physicians $26 Recent Postestimat resear Iion npch f aof tieou n Etqu nd that da ations ys many 1, 2, and 3 high-cos , we trans t claimants form ed no t only coef -0 ficient .r0 each 8 s th ß eir ded , ß , and uctib ß le but into elast often 0 iciti .1 reac 3es calculate h their ma d at xim the um 1 2 3 examin Indi use vi a du health ed als ar . Ch ce are service iropractic indexed care by e v i en befor ( w from 1 as Sel th f-O to N) e exce ne ly Cov the end erag an ption e d of time , t whe he b yea r ye we t (r i from year sfoun reac d he th 20 d. at copayme 13 Thi tshroug ph Oenom the nts h 20 r T ha had en 18 non Se a )l. F f- l O rais ar nl or yge e es Cov ac r effect erage que h osf tions the than s ab even coin out t surance. he specialist physician visit variables, and larger for emergency department visits and inpatient services measures. Emergency department visits 0.019 0.05 5 Female 51% Data Sources ......................................................................................................................................................... 7 These studies include Aron-Di Ane e verat al. ge c (2015) opaym; Bro ent, ts -Go peldber cialisg et al. ( t physic2017); Ei ians chner (1998) $41; Ei14 nav, Finkelstein, and Schri 2 mpf out-of-pocket limit. In essence, for these individuals, use of health care services early in the year affects what they pay mean o Dor, Avi, a f allnd reW gress illiam Enci ors. nosa. 2010. "How Does Cost-Sharing Affect Drug Purchases? Insurance Regimes in the Increasing de Em du erg cti ebl nes cy d he as pa be rtm en e nas t vsis ociat its ed with a reduction in use 0.008 of health care services and ov 0.01 erall spending. relative de pendeef ntfecti varveness iables ( of Y) , diffe we rent typ estimate mod es of c els of ost s an haring lat y use, condit er in the ional year, use such , and as ov copa erall us yme. ents In an did vid coinsurance, uals enrolle o d in nce the an Mor $12 eover, individuals in the CCAE database in large-employer plans are less likely to have missing data in the * As required by the Patient Protection and Affordable Care Act (ACA), out-of-pocket (OOP) maximums are set by the U.S. Source: Medical Expenditure Panel Survey Insurance Component (MEPS-IC). Primary care physician office visits -0.13*** -0.06 (2015); Ellis, Martins, and Zhu (2017); Keeler and Rolph (1988); and Kowolski (2016). Source: Authors’ analysis of IBM MarketScan administrative for health care services at the end of the year. Because deductibles are often satisfied early in the year by high users Little Impac Private Mark t on You et for nge Pres st and criptioOlde n Drust Indivi gs." Journal dua of ls Econom ic Management and Strategy 19 (3): 545-574. Ho s Study ind ame g wever, ividu Sampl rou al reac D Fro p pl ep eart nst ................................ an hes hi min en (de tan of snoted d H or ea Ro h lth and e e by bu r de ck g H d ) uman i uctib (2 n013) year ................................ S le. erv ( fiound ces. t) w Oill have that OP max the long i the mums same ................................ are -term sav lowbe er n for efi H it ng S de As -el was siig gibl n. e heal l The imite ................................ th coeffic d plan to s,individua an ient d ar ß e set rep lsby in res the th ........................ ents e mi the ddle of the 7 BPD data. Pri Note m F : Data a in ry ally, i ca not re col t p lec is wort hted ys iici n 20 ah n 07 notin o . fficeg vtihat sits 0 values were -0 never .16**inferr * ed for any plan desig-0 n e .11 l1 ement and were Age (years) 34.5 Internal Revenue Serv ein ce. rollment and claims data. Specialist physician office visits -0.07 -0.12 6 15 of health care services, it can be argued that it is the end-of-year price that matters most when individuals are deciding health Dor an instead $1 care 0d Encin sp set to end osa ing di mis (2010) sing stri . , des bution prior to pite the pos the sibili adop ty of tio hav n of ing tr the HDHP. ue 0s (e.g., $ They 0 also PC P foun visit co d that pays the , $ re w 0 ded as n uctible). o long- term marginal impact of the member cost share, copayment (c), on Y. The model also includes the following vector of Use of health Sp e care vari cialist pes by hysici a age. n office As see visits n in Figure 9, older indiv 0.07iduals generally use more health -0.01 care than younger Study Variables .................................................................................................................................................. 8 Under 18 ?? 23% ¯ ¯ ¯ Eichner, Matthew J. 1998. "The Demand for Medical Care: What People Pay Does Matter." American Economic Review Copaym ents Ph ayre well sical th -de era fine py v d isan itsd known before health care serv -0.29*ic *es * are used. Coinsurance-0 is. l 2e 5ss well-defined when it whether to use a health care service even before the end of the year is reached. This phenomenon raises questions ?? = ?? * ?? = ?? * ?? = ?? * co spe var ndiates ing re (X duc ) w tion amon ith ? param g h eig ters h-cos to be t claim estimated: ants. Mor age e re -sq cent r uared e;s earch Charlso found that n Comorb id ma ity ny high Index- ;cos Midwes t claimants t, South, not o or nly West 7 ?? ?? ?? ?? ?? ?? people and the nature of health care use ¯ d also varies. For exampl ¯e, individuals ages 40–64 ¯saw specialist physicians an The BPD co Ph ntyains a sical t single o herapy v utpa isitstient co ??insurance rate variable that pres -0.14** ?? * umably applies to all o ??utpa -0.2tie 3 nt services. 18–39 31% Methods ............................................................................................................................................................... 9 88 (2): 117-121. comes to knowing the price of a health care service in advance of using the servic 3 e. An individual may know that they about the effectiveness of deductibles in controlling spending. But it also raises questions about the relative $8 Chiropractor visits -0.33*** -0.18*** s Mor atise fy spe their cifi dedu cally, cti unl bless e but indi often vidual also s face rea d deduct ch their ibles maxi of mum out at least -of $12,0 -pock 00 et l , we imit can . conclude that everyone in the 10 (Northeast as reference); met deductible; met maximum out-of-pocket cost; and year indicators, with 2013 as the Study Variables average of 2.7 times per year, whereas those ages 18–39 saw specialist physicians an average of 1.8 times per year. 40–64 46% 8 Chiropractor visits -0.48*** -0.28*** In about 8 percent to 13 percent of cases, both copayment and coinsurance for a given service contained non-missing wil l pay 10 or 20 percent of the cost of the health care service, but they usually do not know what the allowed charge Study effCopaym ectiveness Sample ent of Re dgi ifferent me Mo ty de pe l ................................ s of cost sharing later................................ in the year — such as ................................ copayments and coinsu ................................ rance — once an .. 9 refer perce ence. nt of F the Ps iny ally, p ch op o tulation a he ca rapture py th /coat uses u s nind seivi lin 70 du g v al is pe f itis x rcent of ed effect thse , an hed alth e -0 car is th .0e 8e *satis idiofsied their yncratic error dedu term. ctible, be -0.caus 16 e everyone in this i it Furthermore, whether a person reaches his or her deductible and/or out-of-pocket maximum may impact their overall Einav, Liran, Amy Finkelstein, and Paul Schrimpf. 2015. "The Figure Respo 1 nse of Drug Expenditure to Non-Linear Contract In addition to the benefit design measures previously described, other annual variables were constructed and utilized in values. To al Ps low for a straightforward an ychotherapy/counseling vialysi sitss of copayment vs. c -0oi .1nsu 1* rance, these individuals were dropped from o -0.003 ur for the service will be until the claim has been adjudicated by the health plan. The uncertainty of coinsurance relative individu $6al reaches his or her deductible. Elas Coinsur ticity a enc 1 re e Re portsduces the expe Dem cted perc and for enta Special ge chanist ge P in hysi Y from cian a 1 per Office cent i Visi ncreas ts e in copayments. e2 is the expected pop Coinsurance ulation used Regime at leas Mod t $12,000 o els ................................ f health care (Figure ................................ 3). Furthermore, amo ................................ ng these high users............................... , 50 to 60 percent 9 Construction of the analytical dataset involved several steps. First, each year from 2013 through 2018, we included all demand for health Per T cy are ser p cen e otag f vices Ce ov oe . f rPer a Ol gde e so r indivi ns Un du der als A (g ag e es 40 65 En –6 ro 4) lled wer in e abo an H ut D3 HP* time or s as l CDH ikely to P,** Amo reach ng their out-of- the analysis. ensuin Design: g analys Evid es,e nce fro including m Me gend dic er, are Part age, ty Dpe ." of The Quarter coverage ly (individ Journal ual of vs Ec . fam onomics ily), rel 130 (2 ations ): 8 hip 41 code -899(.po licyholder, to the certainty of copayme ® nts may mean ® that coinsurance has a differential impact on use of health care services than percentage change in Y from a 1 percent increase in member cost-share under a coinsurance regime. e3 is the Coinsurance Regime Models reache individu d als their m in the aximum o IBM Mark ut-ets of-pock can Th et l o Be se im nef W itit Pl . ith In es an Pr Des s iv ence, f ate ign H eal Database or these th Insu hi (rgh BP an-D ce, use ) file 2007 indi who vid –2018 uals had , u a n se on o-f mis health sing care value services for (1) ei early ther We Stat find su istical bst In antial ference var Em ................................ iatio ployn ee in - the only e clast oveic raity of dema ge................................ nd for health care ................................ for different types ................................ of 20% services in Figure ......... 8. 12 We 10 pocket maxim CCum I an d than Dep those ressi u onn = de 1r age 18. As such, it can be argued that older individuals are less responsive to any spouse, or other dependent), region (Northeast, Midwest, South, or West), and the Charlson Comorbidity Index. We 9 The purpose of this paper is to examine the differential effect of copayments and coinsurance on use of health care copaym The BPD $4ents incl A.g u edes se s 18–3 ts of co 9 7,pay 8 ment and coinsurance variables for both in-network a9 nd out-of-network services. However, expected percentage change in Y from a 1 percent increase in the coinsurance rate holding price constant under a in the year affects what they pay for health care services at the end of the year. This phenomenon raises questions copayment or coinsurance for an in-network primary care physician (PCP) visit, (2) either copayment or coinsurance Ellis, Randall P., Bruno Fa Marti mily cns, ove an rag de Wenjia Zhu. 2017. " HDHHealth Ca P* CDHP**re Demand Elasticities by 80% Type of Service." Journal of Under c also form of find oinsur cos a few t sharing ance, ou differ.ences Ho r com we in ver, it ca parab the elast le ln inear icity also b fix by e e argue ty d-p effect e ofd cos s t s hat pe t spar cif harin ication ents g. ar For ie s: less inst sensitiv ance, coin e to th surance has e cost of ahn ealth effect care on us whe e n of i t Elasticities of Member Cost Sharing ................................................................................................................... 10 Inpatient days 0.02** -0.13*** also created a dichotomous variable indicating whether the member had satisfied their deductible (either individual or services. Our findings have implications for the effectiveness of deductibles, particularly once they are satisfied. Why values for the latter are routinely missing. Therefore, we used the in-network measures only. coinsurance In re pa gi time. No ent days te that the difference, e2 – e3, represents -0.14 t **he price elasticity componen -0.50t* of ** member cost share for an about i tn he ef -netfe work ctiv specialis eness of dedu t phys cti ician bles visit, and 3) in controllinboth g spe indivi nding. But dual an it d also family raises dedu qu cti est ble ion am s ab ounts out. the We re retai lative ned inpatient Health Econo health care, while mics 232 copaym -243.ents doi: 10 do.1 n016/j ot. Th .jheale e coin co. surance elas 2017.07.007. ticity of demand for inpatient 13 health care was com es to their children as compared with themselves. family) Poss b iby le yea Em Confo re en rgund e d. This ncy ers d e ................................ p wa artsm accomplishe ent visits d by summ ................................ ing the d 0.e 0du 3 cti................................ ble field values from al ................................ l -0 claims .02 — at both th ....... e 10 And indeed, we find that coinsurance reduces use of inpatient care and specialist physician office visits more than would we expect to find differences in behavior by whether an individual faces copayments or coinsurance after they $2 10 under 50 coinsur % ance. eff It is w ectiveness orth n Em of ot eing tha rg different eP ne cy r t beside s do e ty n p a C pe rto m s s li v of e en kely be r cost e t v d isit sh sing co aring rrelated wi later in tth he plan size year 0., 0 such 0 , the as missi copayment ngness of s an inpatient an d coinsura 0.05d e nce mergen , once cy an copayment and coinsurance variables for both inpatient services and emergency department visits, the values for which -0.18, meaning a 1 percent increase in price leads to a 0.18 percent decrease in utilization. This is close to the widely individual and family levels — and comparing those values with the prevailing benefit plan deductibles levels. An copayments. In contrast, copayments reduce use of primary care office visits more than coinsurance. Otherwise, we Results reach t heir ................................ d Pri em du acti ry bl ca e? re P ph rio ysr studies ici ................................ an offi of ce te vis ni tfoun s d that i ................................ ndivid -0uals are .14*** not forwar ................................ d-looking whe -0.0n 6*............................... it * c * omes to prices for 11 46.0% 10 departmen Fronstin, Paul, t cost an -sha d M. ring me Christ asure opher s may Roe also no buck, "Healt t be at ra h Car nde om due to Spendin g plan aft- er Ad level hea optin lth st g a atu Full s. F -Re or ex placement, Hi ample, empl ghoyees -Deductibl of a e Policyholder ~ 46% were indivi miss dual reac ing fhes hi or some indivi s or her du deals. ductib Fi le. nal ly, since the BPD’s maximum out-of-pocket (MOOP) measure was routinely Primary care phys ??ician = of ?? fice + vi ?? sits ?? ?? + ?? ?? -0 + .1 ?? 3*** (2 -0 ) .06*** cited -0.20 estimate from the RAND HIE. We find no effect for copayments. This may be due to the fact that such a ?? ?? 43.7% indicator 45% for met maximum out-of-pocket costs was also derived using the same approach with the deductible, found that copayments and coinsurance had about an equal effect on the use of the other health care services Since health h $0 care. ealth T services he same uti may lization be tr differs ue wh in en i bot t comes h type and to tyfp requ e ofency cost by shari age ng a . nd hea Before lth s an individu tatus (amo al rng other eaches the ir Specialist physician office visits -0.03*** -0.07*** mining company might, on average, use more urgent care, thereby increasing the amount of usable data for BPD developers Sample Health Statist Plan ics With ................................ Spousa H e ealth Savings Accou ................................ nt: A Five-Year ................................ Study," EBRI Issue Brie ................................ f, n18% o. 388 (Employe............. e Benefit 11 missing, we assume the statutory out-of-pocket maximum level prevails (Figure 5), recognizing that the true value can small percenta JSp anue ary ge cia o lif F seb tour p ru h ary sam ysicia pl Ma ne ro h ch ffa ice d cop viA spril iaym ts ents Ma f yor inpa Jun tient health e -0.05* J* ul*y care Aug (2 perc ust Sep ent) tember andO -0 be ctob .1 ca e2 ruse *** Not vhere ember m De ay cemb not er be copayment, and coinsurance claim fields. examined in this paper. deductible, they will know what the price of health care is if they face copayments after reaching their deductible. charact to infer p erist lan de ics)sign v , we a alues. lso re -estimate all models for three age groups: ages under 18, ages39.4% 18–39, and ages 40–64. Our Figure 9 11Physical tC he hra ildp /d y e vp isein tsdent -0.22*** -036% .23*** Research Institute, July 16, 2013). 40% be Coinsurance lower. Reduces Demand for Specialist Physician Office Visits ...................................................................... 12 sufficient variation within person over time as relied upon by fixed-effect modelling. Neither copayments nor Physical therapy visits -0.28*** -0.28*** ® 36.9% 36.7% Source: EBRI analysis of IBM Health Analytics MarketScan Commercial Claims and Encounters Databases. Copaym models ar ents e als are well o estim -de ated fine fo d r d aniff d er known ent lev be els of fore h healt ealth h s care serv tatus as defin ices are u ed by sea d. Coi combinat nsurance ion o i f sth le e C ss w harlso ell-de n fine Com d orbid when ity it 11 Use of Health Care Services, by Age, 2018 where the me Chmber iropract cos or t vs is hia tsre is the product of coinsurance -0 rat .31 e *( *r *) applicable to service Y-0 , an .14 d* pr ** ice (p) of each unit of The BPD does not contain any variables for prescription drug benefit design. Seven coins Litturance le an Imp nual health act o were foun n You service d ng to est h ave utili and zatio an Ol de impact n scou t Innt measures on emerg dividuals ................................ enc were y de const partment ructed for use visits ................................ . as dependent var ............................... iables. These were 13 Most employers already use coinsurance for inpatient care. However, only 44 percent use coinsurance for office visits. 33.9% 18 20.8% Chiropractor visits Figu -0 re .3 3 6*** -0.28*** 35% ® ® comes to knowing the price of a health care service in advance of using the service. The uncertainty of coinsurance Fronst Index an in, Paul, d dep an resd sion. M. C All hrist anopher alyses wer Roebu e co cknd , "Pe ucte rsist d us ency in ing Stata High 16 -Cos .1MP t Health Ca . re Claims: 'It’s Where 18.2% the Spending Is, Charlson Comorbidity Index (#) 0.25 Next, using the IBM Marketscan Commercial Claims and Encounters Database (CCAE) enrollment files for each year s ervice Y. We calculated p as the mean allowed amount per unit of service Y, using all adjudicated claims for service Y Psychotherapy/counseling visits -0.12*** -0.13*** 12 the numbers of: If an employer is seeki D ng is to trib manage ution of H use ealth of S he palth endicare ng, A 31.1% s m ervices ong In and s dividupe alsn W diin th g E — m es plo pe ym cial enly amon t-Basedg high users of health Charlson et al. Healthiest Are (1987) Most Se , Deyo, nsitiv Cherk e to Cos in, an t Shar d Ciing ol (1992) ................................ , and Quan et al. (20 ................................ 05). .................................... 15 Psychotherapy/counseling visits -0.12*** 15.5% -0.20*** relative to the certainty of copayments may mean that coinsurance has a differential impact on use of health care Stupid,'" EBRI Source Iss : ue AutBr hoief rs’, no ana. 4 lys9i3 s o (Em f IBM pl o M yee B arketSca enen fit aRe dmsiearch In nistratives e titut nroe ll, Octob ment aner d cl 24 a, im 2s 0 19 da)t. a. from 2 The pu013 thr rpose oough f this 2pa 018 pe,r we select is to exami edne allt full he -dtim ifferenti e empl al e oyees ffect, alon of copay g with their ments and coi spouses and nsurance the on u ir dep se en of de health nts, w c ith f are ull- 29.1% Under Age 18 18–39 40–64 W withi hen n it g come and ts . T to he PCP coe off fficice ien visits t ß repr , weesent find that s the co mar payme ginalnts impact had of nea the rly memb twice th er cost e imp act on t share paid he nu unmber der the of co visinsu its as rance 2 13.3% 30% 13.3% Health Coverage, Continuous Enrollment in 2017 care services —moving from copayments to coinsurance for specialist office visits could provide a viable alternative. In the remainder of the paper, we focus on the elasticities derived from Equations 1 and 3, e and e . The elasticities 13 1 3 Conclusion .......................................................................................................................................................... 16 services compared with copayments. We assu https: med t //ww hat pha w.erm bri.org acy expenditures were included. /docs/default-source/ebri-iss ue-brief/e 11.7%bri_ib_493_highcostclaims- s ervices. Our findings have implications for the effectiveness of deductibles, particularly once they are satisfied. We year (365 orI 366 days npatient H)o e sli pgi ita blili izty ati i on n a non Admis-c sapi iontat s (p ee d rP 1PO 00)plan. PPO 1. me 4 mbers compri 5.2se about two-third 4.4 s of all CCAE regi coins me • urance. on Inp Y atie aThe elast ndnt h can be ospicity ital compared d of ays deman . to d ß1 for . In PCP Equ of ation fice vis 2, d its i w capt asures i -0.11n di for coinsuranc vidual fixed eff e, w ects hile it , ? itwas is the -0.i19 dios for yncratic error 25.3% CCI and Depression = 2 or More from Equation 2 are available upon request. 10.8% 25% Percentage Who 24oct Inp 1a 9.p tiedf nt?s Ho fvrs spin=ced tal Day83 s (c p2f er_ 1 6. 00 ) 9.8 20.6 20.3 use d ind 14 ividu ata als. on A Fb g oetwee e r our s 40– n 1.9 a6 nalys 4 and 3 is, we. i 6 smi olate llion d indivi PPOs since du 9.2% als, ot de hpe er plan nding ty on pe th s ( e e.g., year,HM who Os , E either ha POs) likely ve co di pfaym fer on ents uno or bservable References .......................................................................................................................................................... 17 term copaym , an ents d th.e t In ildcontr e accent ast, on whil ? 22.5% e di copaym stinguishe ents s h be ad no im tween co pact on t mpanion he ede stimand for s mates in Equ pec aiali tion 1 st visit . s, coinsurance had nearly The findings presented in this paper are based on overall use of health care services. The findings based on the models for Key FindingIs n:p atient days -0.06 -0.01 Minimum Reached Their Out- • Emergency Emergen de cy partment Departme visits nt Vis.i ts (per 100) 29.7 36.3 30.2 coinsurance for various health care service 7.7% s. Why would we expect to find differences in behavior by whether an plan design characteristics (e.g., breadth of physician networks), which may in turn be correlated with member cost- any use (i.e. In , t phe e atien xtensive m t d 19.3% ays argin) and conditional use (i.e., the inten 0.04 sive margin) are available up -0.0 on reques 3 t. the same effect as it did for PCP office visits (-0.08). Endnotes 20% ............................................................................................................................................................ 18 Possible Confounders Percentage Percentage Median Spending Average Spending Spending of-Pocket Keeler, Emmett B., and John E. 6.6% Rolph. 1988. "The Demand for Episodes of Treatment in the Health Insurance Em 17.4% ergency department visits -0.01 -0.02** Primary Care Physician Visits 2.2 1.4 2.2 Of key interest in the present study, ho wever, is the impact of coinsurance on health services utilization. In Equation 2, ind sharing arran ividual face gs ements copayment . After s or coinsuranc merging with t e af he previous ter they rea ly de ch the scribir ed de BdPuc D f tibl ile, th e? The s e inters tudi ection o es cited f th abo e tve ofte wo datas n foun ets fd o rmed that 15 Emergency department visits 0.03 -0.02** • Coinsurance reduces use of inpatient care and specialist physician office visits more than copayments do. The mix For simpl bo et ic f w E ity, n een r o we o llco eeinsuran mit s 5.2% o an f S index for hea pce an endin d gcopay lth s p ments er ervices, but Pe rwas son muc ch more , r, and per e p P ven pertai ers o whe n n to the specific service n it cam per e Pe to rshealt on h car Y M being model e prov aximuid me *d in ed.a n Experiment." Journal of Health Economics 7 (4): 337-367. • Primary care physician visits. By design of Pri So p m ur ea ciry analyti al ica stre Ph cal p yh sy id c satas iia ci na V n et, individ i s oiftfs ice visuals coul its d have ei -0 ther c .13*1 *oins .*2 urance or co 1p .8 ayment -0for .02 a given *** 2.7 service each both We look r 15 an %ed d at a num p4 a .5re sources % ber of o o f v ther typ ariation i es of n m office ember visits cost f sor hare. That outpatient is, c sehanges rvices to see in eith ifer they w variabl ere e pro mor du e ce or l changes ess impa in cted their individuals are not forward-looking when it comes to prices for health care. The same may be true when it comes to the final sample of between 1.9 and 3.6 million individuals, depending on the year, over the six-year study period (i.e., • HoPri wever, mary whil caree pmo hys sit em cianpl oo ffiyers ce v al isitrea s dy use coinsuranc -0.13e **for * inpatient care, only 44 -0.pe 05rce ***nt use coinsurance for emergency department. 1% About one 28%-half of plan enroll $120,500 ees had coins$ ura 168 nce ,500 (52 perce $80 nt) ,00 an 0 d copayments 7 (48 0-80 p %ercent). 16 25.5% Physical Therapy Visits 0.6 0.8 1.5 25.2% year. The be Individ nefit de uals can, sign variables i however, have ncluded in th a chang is proc e iness included individua cost-sharing regimel an (coins d faurance mily deduct vs. copay) ibles, and coi over time. nsuran Ice a n our nd produ by typct. e of More cos Spe ov t ci ser, harin alisrt a p g. nd hy s Th p ici ey in can e an ocl ffve iuded ce n vmo isphy itv se s in ical oppos thera ite py divis rectio its -0, chi .0 ns. 2* Eq ropra * uation cto 23r visits .62 % impl , and p icitly ass syc 23.9% umes -0 hother .05* t *hat mem *apy/couns be eli rsng vis respond its. an type un obal f cos ance t sd haring panel datas . Beforet e )an . individual reaches their deductible, they will know what t 23.4% he price of health care is if 22.2% Figures office visits5% . 56% $41,500 $65,315 $23,000 60-70% Kowols • ki, Ama Specia nd list p a. 2016. hysician "Censor visitsed . Quantile Instrumental Variable Estimates of the Price Elasticity of Expenditure on Coinsurance 10% Sp ae verage cialist d p h 20 pe ysicia rcent a n office nd cop visits ayments aver20.3% aged -0 $ .0 162 5**for emergency departme -0nt ser .07***vices. 19.9% copayments fC or i hirnpa oprtient, actor emergen Visits cy room, primary care physician, and sp 0.2 ecialist physi 0.ci 5an. We required nume 0.7 rical values analytical dataset, this occurred for approximately 5.1 percent of individuals for outpatient, 14.1 percent for emergency The similarly to r copayme elnt an ative cha d coing nsur esance in r as elast the icit 17.6% y d ieos to relat were abo ive c ut t hang he ses ame in w p. W ith r e espect to should not phys expect ical thi thera s to py be the and case. Arguably, they fac e copayments after reaching their deductible. Copayments are well-defined and known before health care Physical therapy visits -0.12*** -0.04*** 15.9% 10% 70% $23,500 $41,300 $12,000 50-60% • Medi This su calgg Care. ests" that Journal empl of oy B er us s seeking to m iness and Econom anage use ic Studi of e shealth 34 (1): 107 care se -117. rvices and spending — especially among 14.1% (and whether or not Ph 12.9% ysical va the lues we rapy vis re missing) its to exactly match. -0.23*** -0.20*** Psychotherapy/Counseling Visits 0.7 0.8 0.6 Fi Indi room gure 1, vi, and 4 duals i Perc n H .8entag pe DH rcPs ent e w of for e P re ers in excluded ons patien Und t. f We d er rom Age 6 t ecid he 5 e an Enro d not to alys lled i is for mak n two an e HDHP use reas of ons or CDH this w . First it P, Amon hi , n be -pe cars use g on variation T d hos edu e With P ctibles and a rivat re h inse tead igh Hea est mo lth for dethis led psy at the chothera time of py/counseli making ang vi heals th its care . Ho decis wever io, the co n, the mem payme bent e r is more lasticity likely for chi to know ropractic their care coinsuranc was twic e e rate tha that ofn the th e price of services 5% are used. Coinsurance is less well-defined when it comes to knowing the price of a health care service in 20% 84% $12,700 $24,900 $5,400 30-40% The mix • b Pet hys C w heen ical iropra th coct era insuran or py visvis itce an sits. d copayments was also evenly -0.21 spl ***it when it came to outpati -0ent .03*office * visits. A little high users of health care services — may look to moving from copayments to coinsurance for office visits as a Met Deductible by Year-End 21% 19% 22% 17 Chiropractor visits -0.32*** -0.10** group, few Insurance, satisfy their 20 ded 07–2018 uctible, ................................ which means we do not ................................ have cost-sharing in ................................ formation above th ............................. e deductible for a 5 the two regimes separately. A key issue in our study is that because individuals selected their own health plan, coins the A to servic urance tal oe. S f 2 e ,119 un everal lasticity ique pla factor . s ns might enha were detected across all nce or diminish t yea his p rs (426 in 2018). rice uncertainty . For example, the patient may have advance of using the service. An individual may know that they will pay 10 or 20 percent of the cost of the health care N ewhouse, J. and the Insurance Experiment Group. 1993. Free For All? Lessons from the RAND Health Insurance less than one-half of plan enrollees had coinsurance (44 percent), while a little more than one-half had copayments (56 Psychotherapy/counseling visits -0.08*** -0.12*** way to do so. Met Maximum Out-of-Pocket by Year-End ® 0.3% 0.5% 1% 0% large unmeasure Sourc percenta e: EBR d Ps Ir a y eas ge ch na o lo ons yts f his e th f ra oor t is sa fp IB y/ha M co mple. Se H t choice ue na slte hl iA nn g that con a v ly is td, iic a ts s whe re al Mark rse we e o corr tScado hav nela Cted with om -0 e m .cos 1 e4 rc *t h i* a -*s lea har Cllt ah ing imsservices inf anorm d En at use co io un, w nteoul near rs D d likely -0 a ly all H .t1 a5 b* a*s*leead DHP e s. to bias nrolle ed es have • Chiropractor visits. experience with the service and its associated cost-share. Conversely, the price of an emergency department visit may s 18ervice, but they usually do not know what the allowed charge for the service will be until the claim has been percent). Experim Coi 20nsurance 07 ent. 20 Cambrid aver 08 age ge 20d 09 (MA) 20 pe : Harva 20 rc 10 ent, rd an 20 Uni d 11 co vers payity Press ments 2012 avera . 2013ged $26 2014for pri 20mary 15 care phy 2016 sician 2017 office 20 vi 18sits and Fi g S ure 2, tataCorp. Aver (2019) age Annu . al Employee-Only and Family Deductible, Among Workers in Private-Sector Establishments With Source: Authors’ analysis of IBM MarketScan administrative enrollment and claims data. Source: Authors’ analysis of IBM MarketScan administrative enrollment and claims data. * Based on a subset of individuals with data on out-of-pocket maximums. coinsurance above their deductible. The sample with copayments above the deductible was not large enough to be estimates of the impact of member cost sharing. It is for this reason that we segregated models by regime. This is not be unknown So to the urce: n Au aïve thors p’a a tient. nalysTo is oiso f IBM late Mthe arke effect tScan of ad rm , we inisenter trative p e n in roEq llm uation ent an 2 d t cl o f aiorm ms d E aqu ta.ation 3 as follows: adjudicated by the health plan. The uncertainty of coinsurance relative to the certainty of copayments may mean that This study was conducted through the EBRI Center for Research on Health Benefits Innovation (EBRI CRHBI), $41 for speci a D * HDHP al edu istcti = of hibl gh fic e, -e de 20 du vis c0 ti its bl 2e – . 20 heal 1 th 9 pl................................ an with no health savings accou ................................ nt or health reimbursement arr ................................ angement. ......................... 6 EBRI Issue Brief is registered in the U.S. Patent and Trademark Office. ISSN: 0887 –137X/90 0887 –137X/90 $ .50+.50 *p< 0.05, **p<0.01, ***p<0.001. • Psychotherapy/counseling visits. to s usable ay, f h or o ow **ever, that ur CDHP analys = conis. chang sum er-dires in ected he co alpayme th plan, an nt l HDHP evew ls or coi ith a health nsurance savings ac rate count s or hea thelms th rei elves c mbursem annot also ent arrangemen be t. endogenous. While coins with urance the fun * has p< 0 a .ding 05lar , *ge * sp u<0 r pp im .ort 0pact o 1 of , ** *the pn <0 u fo .s 0e llowing 0 of 1. healt organiz h care ations services : Aon tha He n witt, copaym Blue ents Cros . P s B rior researc lue Shield As h on sociatio cost sharin n, ICUBA g for , © 2020, Employee Benefit Research Institute –Education and Research Fund. All rights reserved. Source: Figure 11 in https://www.cdc.gov/nchs/data/nhis/earlyrelease/insur201811.pdf and Figure 3 in JP Morgan Chase, Pfizer, and PhRMA. https://www.cdc.gov/nchs/data/nhis/earlyrelease/insur201306.pdf ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org ebri.org Is Is Is Is Is Is Is Is Is Is Is Is Is Is Is Is Iss s s s s s s s s s s s s s s s sue B ue B ue B ue B ue B ue B ue B ue B ue B ue B ue B ue B ue B ue B ue B ue B ue Bri ri ri ri ri ri ri ri ri ri ri ri ri ri ri ri riA res ef ef ef ef ef ef ef ef ef ef ef ef ef ef ef ef ef • • • • • • • • • • • • • • • • • Nov Nov Nov Nov Nov Nov Nov Nov Nov Nov Nov Nov Nov Nov Nov Nov Nov earch rep ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 ember 1 or9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 9, 2020 t from the E • • • • • • • • • • • • • • • • • No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. No. BRI 51 51 51 51 51 51 51 51 51 51 51 51 51 51 51 51 519 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 Educ ation and R esearch Fund © 2020 Employee Benefit Research Institute 18 10 14 12 17 15 13 11 16 9 3 5 2 6 8 4 7 ???? ???? ???? ???? ???? ???? ???? ???? ???? ???? ???? ???? ???? ???? ???? ????