With an ever-growing number of workers covered by health savings account (HSA)-eligible health plans, how does use of health care services used to measure quality compare with traditional managed care? New research from EBRI finds a mixed bag: Depending on the health care service, HSA-plans could result in higher or lower use, and in some cases no change whatsoever.
- This study reports use of health care services related to health care quality over five years among over 18,000 individuals from a single large employer in the Midwestern United States that adopted an HSA-eligible health plan for all employees. It represents one of the longest observation periods reported with a full-replacement CDHP, and it is one of the few studies with a matched control group.
- The introduction of the HSA-eligible health plan had a negative impact on office visits for annual physicals, well-child visits, and preventive visits in the year that the plan was adopted. In the second year, office visits increased for HSA-eligible health plan enrollees, but were mostly unchanged for the comparison group. By the fourth year in the HSA-eligible health plan, office visits for annual physicals, well-child visits, and preventive visits were down slightly relative to the comparison group.
- Rates of LDL testing for adults with cardiovascular disease were reduced only in the first year of the HSA-eligible health plan. However, the introduction of the HSA-eligible health plan had a negative effect on medication monitoring for adults on select maintenance drugs not only in the first year that the new health plan was introduced, but in the following three years as well. • The HSA-eligible health plan reduced avoidance of both antibiotics for adults with acute bronchitis and imaging services for adults diagnosed with low back pain. Both services are often considered unnecessary.
- Adoption of the HSA-eligible health plan was associated with a reduction in breast cancer, cervical cancer, and colorectal cancer screening in year one, although screenings for breast cancer and cervical cancer rebounded in year two. By year four, breast cancer screening was higher among enrollees in the HSA-eligible health plan than in the comparison group. In contrast, cervical cancer screening was lower among HSA-eligible health plan enrollees than the comparison group in year four. Throughout all of the study years, colorectal cancer screening was lower among HSA-eligible health plan enrollees than in the comparison group.
- The HSA-eligible health plan was not associated with a change in the percentage of adults receiving HbA1c testing until the fourth year. LDL testing was lower as a result of the introduction of the HSA-eligible health plan in all years.
3 Health Savi Contribu The prescript endorsed by There Appendix Herzlin prior ? are, ger, Preventive i s on ho Re tudi dr tion to the Literature twever, ngs Accounts (HSAs) h gina. ugs. In e e Nationa s ap visi "Let peare several total, s t. ’s Pu l d t Qu limitations o t Consum ality Forum. p exami ending ne on ers I mos to he The this study. nalth car t Char ly HRA-bas intro ge Of Health C e ducti servi Firs eo d t, n of cpl it draws o es avera ans. On the are." HSA gly t ed n H - eli arvard B the experien w $2,4 o studies gible 07 for healt usiness Review (Pa indi h ce pl rente, of a sin van r iduals o esulte Fel g 80 n le, d th ma d , geogra in e n n an HSA-e o slightly . 7 ( d Xu, phi Jul ligi cally y 200 hi 2 bgher le 01 h 0 2e ) ) use alth : References This Issue Brief was conducted through the EBRI Center for Research on Health Benefits Innovation (EBRI CRHBI), Quality of Health Care After Adopting a Full-Replacement, of anti plan concentrated, manufactu (Charlto and $2,498 44- biotic 55. n et pr al., 20 escriptions for th 11) e e co xring ami in t mparison n h employ ed HSA-eligible h e first (20 grou er wit 0p. Tota 7h ), third (20 a large, el em alth pla plo ca 09 pt y n Figure 1 )er sp ,iv s. and fo e work Wit eh n e din urth force ng rollm was $ (2010) y with r ent i 1,94 n e HRA e gio ars. In 8 n fo al, pla r the socio-demo cont ns contractin HSA-eligi rast, ther gr ble aphic, g e was healt and en and co no im h pla rollme npa mpany an ct on nd t in with the funding support of the following organizations: American Express, Ameriprise, Aon Hewitt, Blue Cross Blue There A Propensity S have been on core Matching technique ly a few studies on CD waHPs and thei s used to crea r te a com impact on p h aris ea oln gro th care up. As qualit re y measures. I commended n by Austin deed, the (20 au1 thors of 1), a Figures Individuals enrolled in an HSA-eligible health plan are eligible to open and contribute to an HSA on a tax-free basis. High-Deductible Health Plan With a Health Savings Account: Austin, P. C. "An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Adult Chronic Disease Measures imagin $2,001 for t HSA-eli cultural g g charac ible for lo hh e comparison w back eteristics poten alth plans pain gr in group, or owing an the tially in first ye flu abo d e ex nci ar u pected to cont tn (20 80 percent g beh 07). Ho avior u we in inue n ver, use o de each to r the HSA-eligi gro grou fw im p. (Fronstin, aging associat ble-h20 ealth 13), studies usin ed -pla with n ch low ange. back pai g Consequently, data n from H was high S these A-er in Baseline Sample Characteristics (in 2006) Shiel this 1:1 nearest Issue Br d Associa -n ieeigh tfion, found only bor matc Boeing, seven stu h Deseret was per Mutua d forme ies oln d w , this to Feider thin al one pic. I Reserv -fnifth of ge e neral Employe a standard , thee studies Benef deviatio its ide System, nn tifi of the ed hav Genera pro e e pxe l amin n Mills, sity score. S ed HED Health IS w pays, ec qual ifically, ity JP Employers may offer HSA-eligible health plans and individuals may enroll in such plans without opening or contributing Figure 1, Base Studies." Mu line Sampl ltivariate B e Ce hhavi aracteristics (in oral Research 2006) 46, no. 3 .................................................................................................. 9 (May 2011): 399-424. Ne A Five-Year Study whouse, J. and the Insurance Experiment Group. Free For All? Lessons from the RAND Health Insurance Experiment. the HSA findings are n eligible h -eli ealth gibo le h plans shou t necessarily ealth pla ld fi n gene group relat ll this voi ralizad b le iin t ve to to the h e lit broa eratur der populations. compari e. This study son grou p i en xa y m eines data f ar two (200 rom an HSA 8), and use con -eligible tinue hea d to lth incre planase . Morgan Chase, Mercer, and Pfizer. HSA* Comparison measures associated wit a pro ? b it mo LDL cholester del of grou op me l testin h preventive mbers g forh a ip d off ults (i.e., icdiagnose e s visits, canc tudy vs d w . com ith c er screeni parison) was ardiovn ascular disease. gs, diabetes car estimated on e, a chil large set of pr dhood immunizatio e-period ( ns, a 2006 nd ) to an HSA. September 2014 • No. 404 Figure Cambridge (M 3 shows mean A): Harvar baseline val d University Pr ues for thess, 1993. e 11 health care quality measures examined in this study. Again, statistically Group Group thereafter. medicatio variables. nThese regr monitoring an essors included: d managem egen nt. der, age, age times gender, indicators for spouse and child, years of tenure, By Paul Fronstin, Ph.D., Employee Benefit Research Institute, and M. Christopher Roebuck, Figure Bundor2, M f, M.e K an ate. "Co Healthn Car sum ee Use and Spen r-directed." Research Synthesis ding among HSA-Pla Rep n an ort N d C oo . mparis 24 (Robert on-Gr Wood Johnso oup Membersn , by Foundat Type of ion), October Health (N=18,419) (N=18,419) signific Second, Prior studies ant altho difmostly examin fugh erences estimatio weren of ed o not detecte the nly on impac e y d e betwee tar of of the HSA dat n th a aft e -po ee ligi r the pulation ble intro healt edu n hroll ction of plan fo ed in t ur years the he HSA CDHP. On - post-a eligible e stu do h ption ealth dy (Row is a key stre plane et grou al p a ., 20 ngth of nd th 08) e Where the world turns for the facts ? Medication monitoring for adults on select maintena on U.S. employee benefits. nce drugs (ACE inhibitors, ARBs, digoxin, diure tics, and Employers are allowed to contribute to a person’s HSA, but are not required to do so. In 2014, people with individual and number of members in the household. The Charlson Comorbidity Index was added to the model as a general Ph.D 201 ., RxE 2. Care S conomics ervice, 2006 .................................................................................................................................... 9 Paul Fronstin is director of the Health Education and Research Program at the Employee Benefit Research Institute Parente, Stephen Measure T., Roger Feldman, and Yi Xu. "Impact of Ful Mean l Replacement Mean with Consumer D % Bias riven p-value Health Plans on Figure 5 comparison this analysis, i examined thre grtoup, as t e years o also requir his was the i f dat es pla a after n partici the nten im p da epleme n dts to have effen ct as a tation m r of the aintaine esult of CDH match d co P, nt while inuous e ing. anot ligibil her (Charlto ity throughout n et al., 20 the study 11) ex peri amin od. ed anticonvulsant Preventive Offsi) for ce Visits— at least Wil180 days. son et al. (2008); Parente, Figure 6Feldman, and Xu (2010); and Charlton et al. (2011) all Figure 3 coverage were allowed to make annual contributions of $3,300, while those with family coverage could contribute as measure of health status (Charlson et al., 1987) (Deyo, Cherkin and Ciol, 1992) (Quan et al., 2005). Age: (EBRI). M. Christopher Roebuck is president and CEO of RxEconomics, LLC. This article was written with assistance Health Care Cost and Use of Preventive Predicted Mean QualiServices." ty Measures Insur , bya Y nce Mark ear & Heal ets a thn P d Com lan Ty pp ani e,e 2s: Analyses an 006 ?2010 d Actuarial This e two years o ligibility f data. This criterion may Issue Brief also limit t exam he ingene es five years o ralizability of th f data: on e result e year of s. data prior to the adoption of an HSA- Quality Measures Among HSA*-Plan and Comparison-Group Members, examined the impact of a CDHP on the use of preventive office visits. The findings of these studies were mixed. Wilson Retirement Predicted Mean Quality Measures, by Year & Health Plan Type, 2006 and health benefits are at the heart of workers’, employers’, and our nation’s ?2010 Figure much as Buntin,3, Q Meli $6, unda Be 5 al50. A ity Measures a euwk dditionall es et ym al."He , indivi ong HSA d althcar ua-ls Pla who en S an phave r en d Com ding e a p ac ar nhed age 5 d Pr ison-Gro eventive up 5 a Me C nd are are in mbers not Hi , by yet gh-De Tyenrolle pd eucti of Health ble d in M and C Ca edicare o re Serv nsumer-D may m ice, irec ake an ted Average (years) 30.6 30.5 0.90 0.38 from EBRI’s research and editorial staffs. The authors wish to thank Paul McAuliffe and Sarah Scholle for their valuable Adult Chronic Disease and Overuse Measures About on Computations e-half (46 1 –,4 7 n operc . 1 (ent) of 2010): indiv 4-14iduals ha . d a well-child office visit, adult physical, or a preventive visit. eligi Quality of Health Care After Adopting a Full-Replacement, ble health plan, and four years of by Type of Health Care Service, 2006 data post-adoption. Overuse Me economic security asures . Founded in 1978, EBRI is the most authoritative and objective source of et al. To control (2008) found that C for indivi <18 duals’ ba DHPs resulted seline healt i hn c hi are consum gher use Cancer Screenings o pti f pr on, eventiv the 33% num e of ber an fice visits, d cost o 33%butf the study each o -0.10 f th ee xa fol ml0.93 ined data owing were from also Introduction Health Plans." 2006 ......................................................................................................................... American Journal of Managed Care, March 2011: 222-230. ............................ 11 additional $1,000 catch-up contribution. Any interest or other investment earnings build up tax free. comments and suggestions. Any views expressed in this report are those of the authors and should not be ascribed to Quality Measure (Dependent Variable) 2006 2007 2008 2009 2010 Third, deductible size matters in utilization (Newhouse, 1993 HSA* ), and in this study most Comparison plan participants chose the HSA 18–24 3% 3% 0.80 0.45 information on these critical, complex issues. employers t included in th hae model: i t provided first npatient -dolla hospitali r covera zations, hospita ge for such serv l da ices. Pare ys, emergency nte, Feldma depan, a rtment visits, phy nd Xu (2010) s fician’s of ound that C ficeD visits, HPs Quality Measure (Dependent Variable) 2006 2007 2008 2009 2010 High-Deductible Health Plan With a Health Savings Account: ? Antibiotic prescription fill for adults with acute bronchitis within three days for diagnosis. A consumer-driven h LDL tese ting alth (c pl ardian (CDHP ovascular d) is is eas a combi e) nation of health insurance coverage with a deductible, generally in Among a Quan, H. det al ults . "Co with c din ardiovas g Algorithms for cular disease, Defining Comorb 81–82 percent idities in receiv Group ICD ed LD -9L -CM and ICD testing. A Group bout -10 three-quarter Administrativs (71 e Data. ?74 " p Me erce dica nlt) the officers, trustees, or other sponsors of EBRI, Employee Benefit Research Institute-Education and Research Fund plan with the high 25–34 er deductible amounts, which were higher than 16% national avera 16%ges for HSA 0.90plans at the time (Cl 0.37 axton Charls HSAs are i Figure o4, n, M. E Breast cancer screening (females, age>=40) Predi ndiv.icted M dually et al. a ow e"A anned Qu New Met ality Measur and fully hod porta of C es, by Y las ble. sifying While ear Prog & H dist enostic Comorbidity i al ributions th Plan Ty from pe, 2 an HSA 006 n Lon -2 can 01g 0, it be taken Offi udina ce l Stu Visits f at a dien s: Deve oyr Adults time, t lopme and hese n t and reduc and prescri ed prev ptie on dr ntive vis ug fil its ls. B , buaseline t the s (200 tudy did 6) values o not have a f th co e 11 ntrol gro health car up. C e qua harlton lity measu et al. (res listed belo 2011) also fou w nd t (i.e., t hat th he e CDHP HSA Group (N=18,419) 85% 74% (N=18,419) 78% 78% 76% excess of $1,000 for employee-only coverage, and a savings account option that can be used to pay out-of-pocket of adults Care with 42, nselected chro o. 11 (Novem nb ic con er 20d 0itions r 5): 1130 eceive -9. d medication monitoring. About 34–35 percent of adults with acute A Five-Year Study a 35–44 19% 21% -4.90 0.00 (EBRI-ERF), or their staffs. Neither EBRI nor EBRI-ERF lobbies or takes positions on specific policy proposals. EBRI Data and Methods et al., 2007). [[not in refs]] HSA Group 65% 57% 61% 61% 60% EBRI focuses solely on employee benefits research — no lobbying or advocacy. Validation." Measure ChildrenC Jo ...................................................................................................................... om urnal o parisonf G C rou hron p ic Disease 40, no. 5 (1987) 81% Mean : 373-83 8.2% Mean 79% % Bias 77% p-value ........................... 11 79% withdrawals are excluded from taxable income if they are used to pay for qualified medical expenses as defined under resulted in depen E ? Bd Rent Diagnostic I Em low var ploye ie ables r use e B ima ene of used in fit R ging prev ese(x-ray, m a entive visits, rc the e h Instit nsuin utagnetic e Isg a sue B but the nalysis) also resona rief (IS e Sffect nce ima N 0887 ewas n ?te 13 gred 7 in bar Xg ) is th e (MRI), or ly sta pu e pro blishted m p istically sign ens computerized tomo ont ity hly s by core the E ifi model. Fin cant, was mployee B graphy bas en ally, efit e R d o s (CT) s q es uare n ear ach small d and can) for Institut cube sample e, d 45–54 24% 22% 4.50 0.00 costs for health care services. Savings account options are typically of two varieties: health reimbursement bronchitis were not prescribed an antibiotic, and 84–85 percent of adults diagnosed with low back pain did not receive invites comment on this research. Comparison Group Difference-in-Differences (relative to 2006) 65% -62% 0.11** -0.61% 05 -0.0360% -0.07 58% 110 Adults and Children 0 13th S EBRI t. NW, S stand uite 8s 78, alone in em Washington, Dpl Coyee , 20005ben -4051, efits at $3 rese 00 per ar ych ear or i as san inclinde uded as pend part eof nt, a m nonprofit, and no embership subscription. npa P rtisan eriodi- This study utilized medical and prescription-drug claims data, as well as health insurance enrollment information from a By Paul Fronstin, Ph.D., Employee Benefit Resea Internal Revenue Code (IRC) Sec. 213(d). An individual need rch Institute, and M. not be covered by an HSA-eli Christopher Roebuck, gible health plan to withdraw adults diagnosed with low back pain within 28 days of diagnosis. (n= terms of a 696), and ll co prevent ntinuous var 55–64 ive visi iab ts were les listed subjec abov t to the e were a deduc lso in tcluded. ible in 6% the After pla estimatin n examine 6% g th d. e pro-1.60 bit model,0.14 the sample was Rowe, John W., Tina Brown-Stevenson, Roberta L. Downey, and Joseph P. Newhouse. "The Effect of Consumer- arrangements (HRAs) or health savings accounts (HSAs). Employers have now been offering CDHPs to workers for over imagin Fourth, cals g. althou postage gh t ra Me te dih pai ce HSA atd i ion n m W o-a ni el sh to ig in riibl ng gtoe n, D heC alth , and pl aan cohort dditional mail was ing o ffi matched to ces. POSTMAS comparison members from TER: Send address changes to: a EBRI second Issue Br emplo ief, 11y 00 er as Well-child or adult physical visit Difference-in-Differences (relative to 2006) 47% -0.05*** 46% 0.00 1.70 0.01 0.10 0.03* organization. It analyzes and reports research data without spin or underlying agenda. All findings, Charlton, Mar Figure 5, Predi ycted M E. et ael. an "Eff Qu ects ality Measur of Healthes, by Y Savings e A ar cco & H ue nalth t Eligibl Plan e P Ty lape, 2006 ns on Utili ?2 zati 010 on an , Adult d Ex Chro pennic ditures." Disease American and large manufacMale turing employer located in the Midwest. On Janua 52% ry 1, 2007, it a 53% dopted an HSA- -0.80 eligib0.43 le health plan for all money from Ph.D., RxE 13th St. N t Wc h , S onomics e HSA. uite 8 a78, W ashington, DC, 20005-4051. Copyright 2014 by Employee Benefit Research Institute. All rights reserved. No. 404. reduced to the area of common support, which yielded the final study and control groups, comprised of 18,419 Directed Health Plans on the Use of Preventive and Chronic Illness Services." Health Affairs 27, no. 1 (January? Preventive visit HSA Group 77% 47% 66% 46%68% 1.70 68% 0.10 67% Cervical cancer screening (females, age>=21) a decade. By 2013, 23 percent of employers with 10–499 workers and 39 percent of employers with 500 or more a control, thewheth possibility r er on fi em nan ains that cial data, deroption ived eff s, ects mi or trends, ght still are reve be bias aling ed if an unobserv d reliable ed var — the re iables a aso ssociated w n EBRI informatio ith the n is Journal o Overuse Meas fHousehold size Manageures d Car ............................................................................................................................... e 17 (January 2011): 79-86. 3.1 3.1 1.70 0.10 .... 12 a Copyright Inf Cancer Scree ormation: nings —Bunt This rein port et al. is c(2 opyright 011); P earente, d by the Fe Emp ldmloyee an, an Bene d Xufit (2Resear 010); Havi ch Institut land et e al. (EBRI). (2011); Charlton It may be et active workers and their dependents. Both before and after that date, the employer offered a preferred provider Cancer Scre Adults enings Comparison Group 71% 68% 70% 70% 72% HSA Group 1 individuals each. 60% 52% 54% 55% 49% WheFebr n it came uary 20 to preve 08): 113 ntive -120. cancer scre enings, 63–65 percent of women ages 40 and older received breast cancer the gold standard for private analysts and decision makers, government policymakers, the media, and workers offered either an HRA-based or HSA-eligible health plan. As a result, CDHPs covered about 26 million people in HSA-eligible health plan were also correlated with health-services utilization. There is also the possibility that HSA- Policyholder 43% 43% 0.20 0.84 LDL testing (cardiovascular disease) Difference-in-Differences (relative to 2006) 82% -0.08*** 81% -0.08*** 1.60 -0.08*** 0.87 -0.11*** used without permission, but citation of the source is required. organiz Health Rei al. (201ation 1); an mb (PPO) pla d Ro ursement Arrangements (HRA we et a n. Pr l. io (20 r to the 08) all aexa dopm tion of ined the the use HS s) of ca A-eligncer ible heal screenin th plan, gs such deduc as mammogra tibles ranged f phy, ce rom $5 rvical c 00–$1, ancer, 500 Comparison Group 61% 58% 56% 55% 53% AT A GLAN CE screenin gs, 6the 1–62 publi percent c. of women ages 21 and older received cervical cancer screenings, and 27–28 percent of ? Mammogram for breast cancer screening for women ages 40 and older. Spouse 22% 22% 0.00 0.99 2013, representing about 15 percent of the privately insured market, with most CDHP enrollees being in an HSA-eligible eligi Claxton, Figure ble6, -hG., ealth Predi et- al. cted M pla Avoi n "Healt d e anc n erolle e o anh B f Qu ant ee siality Measur may have bi nefits I otic pres nc 200 ript alt ies, by Y on 7e s :red the Premi (acute b ear u timin rm In onc & H hitc ie g sreases Fal )alth of consum Plan Ty l To pt pe, 2006 ion of An Eight-Y hea ?2010, lth s ear Cancer eLow, While rvices at t Scrh e Of e e nings end of fer Rates An .................. 2006 (either d 14 Medication monitoring 74% 71% 5.60 0.16 Difference-in-Differences (relative to 2006) -0.05*** -0.01 0.00 -0.03** Except colorectal c for geo ancer, an graphid c r prostate canc egion, which er scr was n eenin ot used in gs. Th th e st e udies co matching proc nsistently fo ess due un to th d that CDHPs e study employ led to lower er’s ex use treme of Wilson, Amy per person, an R.d mem et al. "Comparing Quality ber co-insurance ran The Em plo of Care ged yee from Benefi Be10 to twe t Research Institut en 20 a Consume percent for r-Dir e in-net (EB ected Hea RI) was founded in 1978. Its work services, and 10 to 40 lth Plan and a Traditio mission is to nal perc Pla ent n for : An a Child 34% 34% -0.20 0.84 adults a Prior to g ado e 50 ptiand o ng HSA lder -el rec igibe le ived co health lorectal plans, some em cancer screen ployers offered H ing. RA-based plans. In contrast to HSAs, HRAs are Recommended C Avoidance of antibiotic prescription fill (acute bronchitis) HSA itation: Group Paul Fronstin and M. Christopher 36% 35% Roebuck. 33% “Qua 34% lity 37% of Health 0.90 30 C % are Af0.88 te2 r 7Ado % pting a Full- health plan (Fronstin, 2013). delaye Enrollment d or acce Re lerate main Sta d) in an ble. ticipation " Health A of th ffairs e impen 26, nd oin . 5 (Septe g change to t mberh -O e ctobe HSA-eli r 20 gible 07):h 14 ealth 07-1 plan. 6. First-year effects may Colorectal cancer screening (age>=50) contribute to, to encourage, and to enhance the development of sound employee benefit out-of-net concentration preventive ca worncer scre k Years of tenure in th services. e Mienin dwest, covariate gs. Cancer scr bal eenin ance gs w was ere quit lower e goo both when they 12.1 d. The percent12.0 were s bias wa ubject s well 1.80 to below the deductible th 0.08 e commonly (Charlton cited Analysis of HEDIS Measures Related to Management of Chronic Diseases." Population Health Management 12, no. ? Avoidance of imaging for low back pain Cervical ca a Com ncer screeni parison Groupng for women ages 21 and older. 34% 85% 39% 84%39% 2.60 35% 0.57 32% Figure EBRI explo 7, Predicted M res the breadth of emplo ean Quality Measures, by Ye yar ee ben & Health efits and related issues. Plan Type, 2006 ?2010, Adults Wit h Diabetes ............... 14 offered, owned, funded, designed, and controlled by employers, who have a tremendous amount of flexibility in Replacement, High-Deductible Health Plan With a Health Savings Account: A Five-Year Study.” EBRI Issue Brief, no. Who we are HSA Group 29% 18% 21% 22% 21% be ove ? rstated This study reports u if HSA-eligible-healt se of health care ser h-plan programs and so enrollees und public policy v had al ices relate tered d the timi through objective to health car ng of use of h e quality over f resear ech and alth care ser educati ive years v on. EBR ices as a r am I is the only ong over esult of Region Cancer Screenings Difference-in-Differences (relative to 2006) -0.07* -0.04 -0.06* -0.06* 10 et al., perc 20 ent t 11) a hrnd eshold o when th f tol ey erwere ance. I not n oth (Hav er wor ilandd et s, the al., co 20ntrol gro 11), (Row up e et clo al., 20 sely res 08 e). mb Ro led th we et al. e study (20 gr 08 oup ) was t prior to he o nly Among a 2 (April dults 200 with 9): 61 dia-6 betes, 7. 86–87 percent had their HbA1c levels tested, and 83–86 percent had their LDL tested. EBRI studies the world of health and retirement benefits — issues such as 401(k)s, IRAs, retirement Proponents of CDHPs contend that providing participants with a pre-funded account and subjecting claims to high Davis, Kar designing h Comparison Group een. al "Consum th plans that er-Di in rected Healt corporate thi private, nonprof h sCare design. The : iWill t, nonpar It I 26% m amount of mo tisan prove Healt , Washington, DC-b 24% hney availa System ased organi Performance?" 24% ble in the HRA, t zation com 23% Health S hm e lev itted exclusivel e e21% l o rvices f the Research y to 404, September 2014. any anno 18,000 individuals from uncement regarding health a single large employer in th plan changes. e Midwestern United States that adopted an HSA- Breast cancer screening (females, age>=40) ANortheast^ voidance of imaging f or low back pain 65% 0% 63% 28% 3.60 -85.40 0.14 0.00 As noted previously, claims data were available for analysis for one year prior to the adoption of the HSA-eligible health ? Colorectal cancer screening for individuals ages 50 and older. income adequacy, consumer-driven benefits, Social Security, tax treatment of both retirement and health adoption of the HSA-el study identified to look at t igible h h e e ea ffect lth p of lan the . CDHP beyond the first year and found that while use of cancer screenings public policy research and education on economic security and employee benefit issues. Difference-in-Differences (relative to 2006) -0.09*** -0.06*** -0.04*** -0.03** deductibles before in asurance benefits are triggered will induce enrollees to make more cost- and quality-conscious deductible, t 39, no. 4 h (A e compr ugust 20 ehe04 nsiveness of t ): Part II: 12 h1 e9 health -34. insurance, and the account’s portability upon job separation are all Cervical cancer screening (females, age>=21) Midwest^ HSA Group 84% 62% 79% 76% 61% 16% 71% 1.30 160.00 73% 0.50 70.00 0% eligible health plan for all employees. It represents one of the longest observation periods reported with a plan (2006) and for four years post-adoption (2007-2010). Although the HSA-eligible health plan was the only plan Report Availability: Wilson, Amy R benefits, . et al. "Mor cost This e rmanage e Preventive port is ment availa Car , ble e worker a , an on d th Fe e nd wer Intemploye eEmer rnet gency at rwww attitude Ro .eom Visits an bris, .org policy reform p d Prescriptio ropo nsals, Drugs: Hea and pelth C nsion are assets Source: Employee Benefit Research Institute estimates based on administrative claims data. EBRI’s membership includes a cross-section of pension funds; businesses; trade associations; Impact of t were lower inh the first year, e HSA-Eligib there le Health were no Plan on Use longer term e of Quality Measures ffects. Colorectal cancer screening (age>=50) Comparison Group 84% 28% 79% 27%77% 2.00 81% 0.41 81% health care decisions (Her South^ zlinger, 2002). Skeptics, on the other 18% hand, caution t 44% hat patients lack -60.10 comprehe 0.00 nsible, timely, determi Finally, the ned by diff the erenc employer. e-in-diffe Th rences ere econo is no st m atutory re etric metquir hod reli ement t es on t hat a he nassumption th HRA be paireat un d with derlyi a hing trends gh-deductible in qual hea ity lth full-replacement CDHP, and it is one of the few studies with a matched control group. Notes:and funding. There is wi labor un despread ions; health recog care prov nition iders and insur that if employee be ers; government org nefits data exist, EBRI kno anizations; and service firms. ws it. To estimate the impact of the HSA-eligible health plan on use of health care, a difference-in-differences (DiD) modeling option, Adults With Diab Utilization worker in s were offered a cho a Consumer etes -Driven Healt ice of tw h P o deductibl lan." Bene es: fits Quarterly 24, no. 1 (First Quarter 2008): 46-54. Adults With Diabetes Difference-in-Differences (relative to 2006) -0.04 -0.06* -0.08** -0.11*** West^ 3% 12% -32.50 0.00 Deyo, R.a A., D. C. Cherkin, and M. A. Ciol. "Adapting a Clinical Comorbidity Index for Use with ICD-9-CM Administrative Figures 4–7 and trustwort pr hesent predicte y information that is d mean critical qualit for t y measures ex hem to ma amine ke informed d d in this pa ecisions ( per. Predict Davis e, 2004). Skeptics are also d average quality measures are measures are plan, althou Health savings account. ghcomparabl it is commo e ac n practice. ross the two groups. To the extent that this assumption is violated—perhaps due to Source: Emplo yee B enefit Research Institute estimates based o n administrative claims data. HbA1c testing 87% 86% 2.90 0.66 framework Diab wa etes Car s utilized. e— Using Four stu probit, dies exami the fon llowin ed thg e mo impact del was of CD estimate HPs on d. diabetes care (Wilson et al., 2009) (Buntin et Charlson Comorbidity Index 0.13 0.13 -0.40 0.71 Datab ? Hemoglobin A Presented are marginal effects of the difference-in-differences estimators, calculated at the mean values of all other regresso ases." Journal o 1c (HbA1c f Clinical Epi ) screenin demiolo g. gy 45, no. 6 (June 1992): 613-9. rs, which pres concerne ented d t for haeach ye t higherar. T cost sharing mi he figures als ght l o se how the ad to less us resule of ts for th preventiv e diffe er care, ence-in pr -imar differ y care, a ences (Di nd ot D) re her n grese scessary ion mode health l geogra ? phi The intr c region t No oducti tes:hat on of the was not us HSA-eligible health plan had a ne ed in the propensity score matching process—the gative impact on results may re office visits fmain or ann biased. ual physic als, LDL testing EBRI’s work advances knowledge and unders 86% 83%tanding of emplo 6.50 yee benefits and their 0.33 EBRI delive 1) $1,250 r per s a stead individual, y $ stream of invaluable research and anal 2,150 per family; or ysis. Table of Contents Source: Employee Benefit Research Institute estimates based on administrative claims data. al., 2011included: indicators for male, spouse, child, five age groups, four years, HSA, and four HSA times year interaction terms (pres ) (Haviland et al., 2011) (Rowe et al., 2008). These studies examined whether the CDHP had an ented), as well impact on HRAs are typically int * Health segrat avings ac ed with count. the health plan offered by the employer. Essentially, workers with an HRA have first- Source: Employee Benefit Research Institute estimates based on administrative claims data. presente care servic d in es, eespecial quation (1) in ly amo the A ng indivi ppe importance to dua ndix ls . The of lo coef th wer socio e nation’s econo ficiene t on conomic stat thmy e i n among policy teraus (Havila ction term makers, nd e bettw the news al., 2 een th 01e HSA 1 media, and ) (Newhouse, and Year i the public. It 1n 9di 93 cators ). well-child visits, and preventive visits in the year that the plan was adopted. In the second year, office as Charlson Comorbidity Index, years of tenure, and household size. All equations were specified using probit. Standard errors Notes:EBRI publications include in-depth coverage of key issues and trends; summarie were s of research Prob Y a ßYear ?HSA dYear HSA pX e 1 Endnotes glycated hemoglobi P resent n ed (HbA are marginal 1c) testin effects o g, eye exams, f the difference-in-diflo ferw-de ences es nsity lipo timators, calc prote ulated at in t(LDL) scree he mean values oning, f all other ne regres phro sopathy screen rs, which ing, and Fendrick, A. ? * Health savings account. LDL cholester Mark et a oll screen . Health Savin ing. gs Account-Eligible High Deductible Health Plans: Updating the Definition of does this by conducting and publishing policy research, analysis, and special reports on Introductio dollar coverage u clustered at the household level. n.................................................................................................................. * Health savings account. p to a limit (the employer contribution to the HRA), then face a deductible ........................................ , after which insurance 4 What we do are Another co shown for each 2) $2, ncern is that 500 finding per year rela indivi the s and rd tive to eual, duc policy $ ed utili 42006. ,300 develo zat p er ion mi pfamily ments; ght save . timel costs in th y factsheet e s s h on ort-term, hot topics; but might regula w r e up ll re dasult in tes on legi larger l slat ong- ive and visits incr included: eased for HSA-eligible health pl indicato rs fo r male, spo use, child, five age gran enro oups, four years llees, , HSAbut were mostly unchanged for the com , and four HSA times year interaction terms (presented), as parison Notes: microalbumin testing among pati ents w emploiy th ee benef diabetes. A its issues; holding educational br gain the findings wereiefings for EBRI memb mixed. Buntin et al. ers, congressional and (2011) and Rowe et Preventio Statistical significance denoted as follows: *** p<0.01; ** p<0.05; * p<0.10 n. Center Values are proportions unless denoted otherwise. for Value-Based Insurance Design, University of Michigan, 2014. well as Charlson Comorbidity Index, years o f tenure, and househo ld size. A ll equations were specified using probit. Standard errors were Conclusions takes effect. This is in regul contra atory de st to an HSA velopment -s; eligi comp ble healt rehen h si plan, w ve refere here em nce resou ployers cou rces o ld n simply of benefit pro fer t grams he health and pla won rkforce and Background on CDHPs ........................................................................................................................................... 4 Wher 1 term costs. e Values are proportions unless denoted otherwise. Y is a measure of use health care for person i in year t, Year is an indicator variable for each year from 2007? group. By the fourth year i n the HSA-eligible health plan, office visits for annual physicals, well-child visits, Variables not used in propensity score matching model denoted by ^. federal agency staff, and the news media; and sponsoring public opinion surveys on employee See http://www.merc clustered at the ho er.co usm eho/ne ld lev w esroom/unite l. d-states-health-benefit-cost-growth-slowed-again-in-2013.html al. (2008) found no impact on diabetes care. Haviland et al. (2011) found a negative impact on diabetes care. And, Reported p-values from t-test of differences in means across groups. Almost 90 Office perc V ent of isits fo issue enrolle s r Adults ; and es selecte major survey and Childr d thee hi ns —T o gher fhe pu-ded iblic ntroductio u at ctitudes. tiblen optio of the n, HSA-e which re ligiquir ble healt ed noh worker plan was co foun ntributio d ton have , in contrast a let workers decide for themselves whether to open and fund an HSA. 2010 (vs. 2006), HSA is an indicator for whether the person was a member of the employer implementing the full- This study adds to the SReported p-values from t-test of differences in means across groups. tatistical s literat ignificancure on e deno ted as benefit issues. consu followm s: *er-dr ** p<0. EBRI’s Ed 01 iven ; ** p< he 0.05;alth * p< uc 0. ation and Re 1plans 0 (CDHse Ps) by arch F eund xami (EBRI-ERF) performs ning use of health car the char e servi itable ces , HSA-Eligible and preventiv Health e Plans visits were ..................................................................................................... down slightly relative to the comparison group. .............................. 4 Results Fronstin, Paul. "Characteristics of the Population With Consumer-Driven and High-Deductible Health Plans, 2005–2013. " Wilson et al. (2009) found no impact on HbA1c and LDL screening, but a positive impact on eye exams and EBRI meetings present and explore issues with thought leaders from all sectors. negativ to the This Issue Brief require e impad ct on th payroll examin e de es h num duo ction of ber o w adfoptin o educa ne fficarly g e tiona vi an $3 sits for an HSA-eli l, and 00 ann scientif gibl ual nual elic func yhealth ph per ysicals, worker ti pla ons of the Instit n im wel for employ pacts the l-child visits, a ute. EBRI use ee-only covera of he n -ERF d preventive alth is a tax-ex care ge in services that the lower- visits in the year empt organizatio are n used to 2 replacement meas H ure SA-eli quality gible health of health plan, planand s during the X is a vector o four years f mem after ber cha a he racteristics for alth savings account (HSA)-eli person i in year g t in ible healt cludingh : th ple an Health Savings Accounts (HSAs) ......................................................................................................................... 5 This sample in EBRI Notescludes almost 5,000 members who were not , no. 4 (Employee Benefit Research Institut contained in e e, April 2014a ). rlier analyses (Fronstin and Roebuck 2014) EBRI regularly provides congressional testimony, and briefs policymakers, member organizations, nephropathy screening. supported by contributions and grants. that the used to ded Differing In uctib meas HSA-eligi le plaure n centives . Th quality ble e h employ ealth of health plan er contri was a plabute ns. dopted Data come d to the (Figur HSA, fro e 4). m de a lar Prio positin r to geg em the a $70 ployer that 0 for doption of the HSA- employe reple aced -only all o covera eligi f its ble h ge a health ealth nd insuran $ plan, 1,30 48 0 for ce per- ? Rates of LDL testing for adults with cardiovascular disease were reduced only in the first year of the HSA- Charlson Com Cancer Scree orbidity In nings dex, —Scre demo eni graphic ngs forcharacteristics breast cancer, c (age, ervic gender, al canc and geogra er, and colorectal phic region cancer f ), an ell i bet ndicator ween of 200 wh 6 a ether nd was adopted. As with prior research examined, the findings are mixed. Annual physicals, well-child visits, and Descriptive Findings (Fronstin and Roebuck 2013) (Fro and the media nstin, Sepulveda an on employer benefits. d Roebuck 2013) (Fronstin, Sepulveda and Roebuck 2013). After deep Health Reimbursement Arrangements (HRAs) ...................................................................................................... 5 Figure 2 Figure 7 cent of options with a family cover enrolle age, re n es had of HSA-eligible health gardless fice v of w isits for an hich pl de annua du . Uctible sl physicals, e of was se health we -c leare quality cted. ll-chi ld visits, and measures is e preventiv xam eined over visits. In 2007, t five years—of he year t which he HSA- eligible health plan. However, the introduction of the HSA-eligible health plan had a negative effect on or not the m R prevent 200 ese 7 for both enrolle arcih of ve visi ten emb ts were co er mb was the es in low ein s e eth r np re in some, o olicyhold lHSA-eli lees in HR g but eibl r (vs. sp A a n e health on t all y d HS ouse/ e pla A-e ars fol n d lig a ep ib nend ld th lowin e healt ee com nt g th ), hhou pla e a pari d s noption o ehold si s son grou into oze n f th e gr p (i. (Figure e e oup for HSA-eli ., the6). num g s But ible tud be y purpos be hr of ealth cause canc individua plan. es. Ho LD ls on th er wever, L e ."Findings From the 2013 EBRI/Greenwald & Associates Consumer Engagement in Health Care Survey." EBRI EBRI issu Mean Health es press releC as are Use and es on newsworthy Spending Among H developments, Sand i A*-Plan and s among the most widely quoted Childhood Immunizations—One study (Buntin et al., 2011) examined the impact of a CDHP on childhood investigation, it was determined that some plan IDs were renumbered during one of the study years, which resulted in some EBRI Issue Briefs is a monthly periodical with in-depth evaluation of employee benefit issues Figure Differin 1 co g nta Ince ins the ntives basel ............................................................................................................................... ine (2006) sample characteristics of the individuals in the HSA-eligible health plan and t ............. he 5 Predicted Mean Quality Measures, by Year & Health Plan Type, 2006 ?2010 eligi four yea ble hrs foll ealthow th plan was e ad a opti dop on of ted, the 44 per HSA cen -eli t had of gible healt fice vi h pl sits for an an. The an nual alysis makes u physicals, wel se of a l-child visits, a matched com nd preve parison ntive grou visits. p medication monitoring for adults on select maintenance drugs not only in the first year that the new policy), screenin an gs f d th ell more e num souber for th rces ofo e years of t nHSA-eli employee gible enur be h ee with n alth efits the plan by all media. em thapl noyer. for the , , , com , andp arison grou are par p, the ameters HSA-e to be estimated. The ligible health plan e enrolle testing Issue Brie es an for ad duutilizatio f, lts with no. 39 cardio n 3 patt (Employ erns in vascular dise ee B these enefit a plans shou sRes e was lo earch I wer ld be nstitute only in examine , Dec thed se firs ember t year parately 2013 follow bec ). a inuse HRAs an g adoption of d HSAs create the HSA-elig di ible fferent Comparison-Group Members, by Type of Health Care Service, 2006 and trends, as well as critical analyses of employee benefit policies and proposals. EBRI individuals being misclassified as immunization rates. It found lowe nor rates t continuo as a result usly enro of lled in the earlier stu the CDHP. dies. Subsequent to the earlier work, these records Our The study group consisted of 18,965 active workers an Adults With Diabetes d their dependents who were continuously covered by the comparison group, both of which included 18,419 participants. Few statistically significant differences were detected Contribution to the Literature ................................................................................................................................. 6 In contrast, t of active work hers and their e percentage depen of enrolle dents es in from a secon the comparis d larg o Figure 4 n gro e employ up with er that co such a visit ntinuw ed as mostly unc to insure mohanged. st of its workf orce health plan EBRI was intr directoduced s membe , but in rs and the other constit following three years as well. uencies to the informatio n they need and undertakes new Notes is a monthly periodical pr HSA* oviding current in Comparison formation on a variety of employee benefit term is a health incentives showed a pla nn rela i negative im , th diosyncratic error ted to ough medicati use o pact o f health non monitori . the pr care edict se neg fel rvices. HRAs a d proba l in all bility o yearrfs. Use of a e spe gettin ng di a ng a can ntibiotics for a ccounts, not s cer screenin cut g avings acco e bro relative to nchitis, the unwhich ts, and t comparison is oft he e lac n gr k of oup. 2 were rejoined and statistical power was bolstered accordingly. The impact of the HSA-eligible health plan on health services employer between the duri population ng the five e -year analysis nrolled in the HS periA-eli od. Data from a secon gible heal 2006 th plan an dd th 2007 eme compar ployer wer 2008 ison e us group. ed to cr Th 2009 eate is was the i a comparison 2010 ntende group. d effect Quality Measure (Dependent Variable) Predicted Mean Quality Measures, by Year & Health Plan Type, 2006–2010 research on an ongoi topics. ng EBRIef basis. is a weekly rou Group ndup of EBRI r Group esearch and insights, as well as updates on through traditional n CHECK OUT EBRI’S WEB SITE! etwork plans without high deductibles or savings account options. This study represents one of the Current . Work "Healt .................................................................................................................. h Savings Accounts and Health Reimbursement Arrangements: Assets, Account ...................................... Balances, and Rollovers, 6 publications However, Medicatio in 2008 (t n Mo he seco nitoring an nd year i d M na th nagem e HSA-e entligi —O bne stu le health dy (Wi plan lson et a ) the num l., 2 ber o 009) e f br xeast cancer amined medicatio and cervi n mco al ca nitoring an ncer d unnec guarant essary, was fou eed portability ncreate d to be sli s the ghtly equ hi ivale gher, nt of a while use “use-it-or-los of imagi e it ng fo ” ince r low ntive to back pai use th n, also oft e money en in u th nnecessary, al e account on so health utilization and costs was re-estimated and qualitatively similar results as previously reported were found. These results are HbA1c Testing surveys, studies, litigation, legislation and regulation affecting employee benefit plans, while In 2 This secon 008, thde s ple an sponsor cond year o (w f enrollme ith 82,154 nt in con tOffice Visits for Adults and Children thinuously e e HSA-elig niro ble lle he (N=18,419) d workers alth plan a , nd the (N=18,419) de percentage of enrollees pendents) maintained PPO covera with office visits for ge that as a result ? The HSA-eligi of matching. EBRI ble health plan reduced av maintains an Average age w d analy as 31 zin esboth the mo oidance grost ups, an of bo comp d just th antibiotics rehenover one sive databa for adults with acute bronchitis and -halfse of bot of 401(k) h groups were -type prog male. rams in the The probit equation was estimated for each of the 11 health-use quality measures. Standard errors were clustered at longest o 2006– bserv 2013." ation perio EBRI Issue ds re Brief, porteno. d with 39 a 5 (E ful mployee l-replace Be me ne nt, HSA fit Research -eligib In le healt stitute, Janua h plan, an ryd is o 2014n). e of the few studies Data and Methods ................................................................................................................................................. 7 managem increase care servic screenind gs rebounded, over ees. nt dif In additio time. ferenBreast c ces narro n, si between nce HRAs wing th ancer scre CDH e gap are P a enifunded by n n such that d g w non-C as low D ther em HP er i ployers, wo e wa enrolle n the first year, s no difference es. Th rkers may e fin but din then rebo in 2008 and gs in view t thi hse acco study w unded. C 2009 in unt as ere e br rvic mixed. “free” money east cancer al canc It er screenin foun an d that and d g HSA* Group 87% 80% 79% 81% 79% Quality Measure (Dependent Variable) Measure EBRI’s Blog supplements our regular publications 2006 Mean 2007 Mean , offe2008 ring % Bias commentar 2009 p-value y on questions 2010 available upon request. world. Its computer simulation analyses on Social Security reform and retirement income adequacy annual was not H Policyholders physicals, wel SA-eligi made bleup abo throu l-child g uvisits, and hout th t 40 percent e five-year study pre ov f eac entive visits i h group. per in od Average cr. eased to The comparison em hous 46 ehold size percent, w was ploy hiler offered t e it about was mostl 3.1. T wo ey PPO nure uncha options: with nge the d f o one r the imaging services for adults diagnosed with low back pain. Both services are often considered unnecessary. the ho with a usehol control d level. group. Pre dicted probabilities were for all five years of data for the outcomes, using mean values for the CDHP a cervical ca Well-child/adult physical visit nd non Comparison Group ncer screeni -CDHP enrol ngs betwee lees wer n HSA rece e equa ived from -ell ligi y likely to ble news healt reporters use h plan 88% appro e , p nrolle oplic riate m ym es 82% akers and t edicatio , an he com d others n 80% amo p. The arn isg on grou patients with asthma a EBRI Databook 82%p. In fact, 84% on Employee by 2010 nd (the may not hav was lower ine t h Use of Services a savings m e first year,e r netality attach bounded, but ed to the then fell accou again n. t. Colorecta In contrast, l cancer screeni HSAs are savi ng ngs accou was lower i nts t n all y hat are fu ears. lly While Dependent EBRI’s website is easy to use and packe Variables........................................................................................................... d with useful information! Look for ............................... 7 are unique. a comparison with no deduc grtoup. ible throu Thus, t gho heut the HSA-e study ligible period and on health plan ha e d no with im a deducti pact onb le office that stea visits for dily incre annuased during the al physicals, wel study l-child Fronstin, employers Paul, was and M. C about the s hristopher ame in both Roeb gr ucoups k. "Bra , 1nd-Na 2 to 12 m .1 years e and Gen , but eris c Prescri tatistically ption diff Dru erent. g Use A As note fter Ado d previously, most ption of a Full-of other regresso 3 HSA Group rs. Marginal effects estimates Benefits is a sta (calcu tistic late al r d a efer t the m enc 48% e work on em eans of all r ploye 44% egreessors) of benefit prog 46% th rams and work force-related e HSA 47% -eligible healt 46% h plan’s Difference-in-Differences (relative to 2006) -0.01 -0.01 0.00 -0.05* Inpatient hospital admissions (per 1,000 equal portabl HbA1c t fourth y See lhttp://www.qualitymeasures.ah y like ee e , o sting ar o w ly t n f e o the HSA for a d, a be mana n dd fre ults -eli q with gin g uie bg ntly f le h dia ant e b ialth u de etes nrq.gov/content.aspx?i pressa d pla ed by the was n) breast nt me mostly unaf individu dications cancer fal. e dcted, LDL . screeni =47190 Any tim However, and ne g wa testi a an worker c s high nual mo ng was lo er amo hnitorin ooses to not spend the money wern ig e g of n all n pat rolle years. ients o es in the HSA n persistent -elig in ibthe le Office Visits for Adults and Children ..................................................................................................................... 7 these special features: ? Adoption of the HSA-eligible health plan was associated with a reduction in breast cancer, cervical cancer, Comparison Group 45% 44% 45% 45% 45% visits, and perio the HSA Replac d. Both -elipre ement, g P ibv le h PO optio entive visits i e Consumer-Dir althn pla pla n n n group s 2 covere 008 ected Health was issue (relativ d s loc pr . eventive a e to ted Pl20 in t an s 06 With he M e), w rvic a H idwes es hich in ful ealth Savi was also t (7 l 9 an per d had 20 ngs A c true ent), ccount." for 2 wh percen erea 009. EBRI N s t By the comparis member co o fourt tes, on- no. h -igroup nsurance for year in 3 (Employee mem the HS bemost rs A- were LDL Testing individuals) 34 38 -1.80 0.09 impact on the likelihood of each of the dependent variables are also presented. All analyses were conducted using medicatio health plan ns (a than in ngiote the com nsin-con parison verting-en grouzp. yme ( However, ACE) inhi despite t bitors, angiote he rebou nn sin II d in rece 2008, c ptor ervical canc blockers (ARB er screeni s), digoxi ng was l n, ower HSA, he http://www.qu or she is e a a litymeasures.ah ffectively saving it rq.gov/conte for fu nt.aspx?id=47167 ture use. EBRI makes information freely available to all. Difference-in-Differences (relative to 2006) -0.03*** -0.01 0.00 -0.02** and color HSA Group ectal cancer screening in year one, although screenings for breast cancer and cervical cancer Adult ChronicInpatient hospital days (per 1,000 individuals) Disease Measures .......................................................................................................................... 87% 105 77% 13176% -1.90 78% 0.07 75% 8 eligi other b le cover health ed services. plan, office vis The comparison gr its for annual oup was physicals, w const ellructed -child visits, a by matn ching on d preve pre ntive -period (200 visits were do 6) v wn sli alues of ghtly a rel rich ative set of to more evenly Background on CDHPs Benefit Resear distributed acro ch Institutss geogra e, March 2 phic 014 regions. ). Stata/MP 12.0. • EBRI’s entire library of research publicat ions starts at the main Web page. Click on EBRI Preventive visit EBRI assumes a public service responsibility to make its findings completely accessible at www.ebri.org diuret among HSA ics, an-d elanticonvulsant igible health pls an en ) was rolle lower a es an m d t ong C he com DHP par enrollees, t ison group hou in gh t 201h 0. e act Throu ualg difference hout all of the was study ye only one ars (200 percentage 7– Comparison Group Emergency department visits (per 1,000 84% 81% 78% 83% 80% rebounded in year two. By Contact EBRI year four, breast canc Publications, (2 er scr 02) 659-0670; eening was hi fax publication gher among orders to enrollee (20 s in t 2) 775-6312. he HSA- a the com covariates to t Overuse paris Meas on gro he stu ures up. d.............................................................................................................. y gro up. The analysis sample was reduced to 18,419 after matching. More in ............................... formation on this 8 There are two HSA Group types of CDHPs: health plans that include an integrated HRA, and HSA-eligible health plans. Enrollees i n HSA-eligible health plans should also be examined sepa 48% rately from 44% HRA enroll 47% ees because th 47% e em 46% ployer Issue Briefs — so that all deci and EBRI Notes sions that for our in-depth and relate to employee ben nonpartisan periodicals. efits, whether made in Cong ress or board rooms or individuals) point. 2010), colorec Difference-in-Differences (relative to 2006) tal cancer screening was lower among HSA-elig137 ible healt -0.07** h plan 139 enroll -0.06* ees than -0.40 -0.08*** in the comparison 0.67-0.08*** group. Subscriptions to EBRI Issue Briefs are included as part of EBRI membership, or as part of a Except . for "Healt geog h Care raphic r Sp een gio di nng a , whic fter h Ado was n pting ot used in a Full-Re the p lmatching proc acement, Highess due -Deducti to th ble H ee study alth Plan employ With er’s e a He xalth treme Sav ings eligible health plan than in the comparison group. In contrast, cervical cancer screening was lower among process is Orders/ Comparison Group provided in the Appendix. 45% 44% 45% 45% 45% families’ homes, are based on the highest quality, most dependable information. EBRI’s Web site posts Cancer Source: Employee Benefit Research Institute estimates based on administrative claims data. Screenings Outpatient physician's office and clinic visits ............................................................................................................. 3.5 3.5 -1.40 ................................ 0.18 8 offering the HSA-eligible health plan $199 annual sub may simply be o scription ffering a h to EBRI Not igh-de es dand ucti E ble BRI Issue health pl Brie an fswhil . Change of e not fun Address: ding the EBRI, HSA or Difference-in-Differences (relative to 2006) Mean Quality Measures for Adults—Figure 5 contains the findings -0.03*** for various qua -0.01 lity measures s 0.00 pe-0.02** cific to adults concentration Account: A • HSA-eligible h Visit EBRI’s blog, or subscribe to the EBRI in th Five e Mi -Ye edw ar St alth plan est, covariate udy." enr EBRI Issue ollees than balance Brief, was the compari no. 3 quit8 e 8goo ef (E e-lette s mployee d on . The group in year four. Throu percent r. Be nefit R bias esearch was well Institute, July below ghout all of the commonly 2013 the study ). cited Notes: all Prescription drug fills research findings, publications, and news alert 10.2s. EBRI also exte 10.3 nds its e -0.80 ducation 0.45 and public service HSA-Eligible Health Plans 1100 13th St. NW, Suite 878, Washington, DC, 20005-4051, (202) 659-0670; fax number, facilitatinOther Diab g wo etes Meas rker contri —One study ures buti (Wilson —Two ons to an me et al., HSA asures for 200 . Workers 9) als qua o are lity of exami often o care am ned L ffD eL sc red ong a reen only the dults ing for pati wit he halt dih abet pl ents wit an a es are exami nh d le cardiovascul ft on th ned i eir o nar conditions Figure wn to fin 7. d Adults With Diabetes .......................................................................................................................................... 8 Source: Employee Benefit Research Institute estimates based on administrative claims data. a Subscriptions Health savings account. with c 10 perc ertain ent t di hrseases or con eshold of toler dance itions. It (Austin, was foun 2011). d t In hat i othe ntroduction r words, the of t c ho e HSA mparison -eligigro ble he up c alth losely res plan had a emble nd t egativ he st e impact udy yearsrole , col Spending to improving Ameri orectal cancer scr (202) 775-6312 eening wa cans’ finan s lo cial ; e-mail: wer am knowle ong subs dge criptions@ebr HSA-eligible health plan enr through its a i.org w ardMe -win mbe ning rship Information: o pllees th ublic se an in the rvice cam Inquiries paign Dependent Variables Notes: The and HSA-e use of im ligib agin le healt g for lo h pla w n back did not h pain. It f aveo any im und that LDL pact (positive or n screening an egative) on th d appropriate e perce use of n imag tage of ing for adults r low- eceivin back pai g HbA n 1c an HSA • provider an EBRI’s reliable health and retirem d open an account. ent surveys are just a click away through the topic boxes at In order to be an HSA-eligibl® e health plan, the plan must have an annual deductible of at least $1,250 for employee- Presented are marginal effects of the difference-in-differences estimators, calculated at the mean values of all other regressors, which Results Fronstin, ....................................................................................................................... Paul, Martin J. Sepulveda, and M. Christopher Roebuck. "Consumer-Directed Health .......................................... Plans Reduce The Long- 8 a Inpatient hospital costsregarding EBRI membership an $213 d/or contributions $228 to EBRI-ER -0.50 F should be dir 0.63ected to EBRI on the amounCt of LDL hoosetoSave testing for and ad the ults co wi mpanio th cardiovascu n site ww lar di w.csease i hooset nosav the first year e.org after adoption of the HSA-eligible group pr Health savings account. compari ior to adoption of th son group. e HSA-eligible health plan. included: indicators for male, spouse, child, five age groups, four years, HSA, and four HSA times year interaction terms (presented), as well The current study examined a total of 11 annual measures of use of health care services related to quality measures. In testing were both until the top of the page. hi 2 gher amon 010, the fourt g CDHP enrol h year in the lees t HSA-eli han amon gible g h no ealth n-CDHP plan. enroll The volume ees. of LDL testing was lower following the only coverage and Emergency department costs $2,500 for family President Dallas coverage in Salisbur 2014. An out-o y at the $89 above f-pock address, ( et maxim 202) 659-0670; $92um may not e -0.90 e-mail: xcee salisbur d $ 0.38 6,3y@ebri.org 50 for self-o nly Presented are marginal effects of the difference-in-differences estimators, calculated at the mean values of all other regressors, which included: indicators for Term Use Of Outpatient Physician Visits And Prescription Drugs." Health Affairs 32, no. 6 (June 2013a): 1126-1134. Descriptive Findings ........................................................................................................................................... 8 health pla as Charlson Comorbidity Index, years of tenure, and household size. All equations were specified using probit. Standard errors n, but no effect was found in the three years following the introduction of the plan. The intr were oduction of the Because employers have been offering HRA-based plans longer than they have been using HSA-eligible health plans, male, spouse, child, five age groups, four years, HSA, and fou Outpatient physician's office and clinic costs r HSA times year interaction terms (presented), as well as Charlso $294 $296 -0.80 n Comorbidity Index, years of 0.45 general, th introductio n o e aflgorithms used the HSA-eligib to construct le health pla t nh in ese varia all fourb ye lesars. followed the methodology employed by the National coverage or $12,700 for family coverage with the deductible counting toward this limit. Minimum allowable deductibles clustered at the household level. Figure 2 shows mean baseline values for health-services utilization and cost measures for the HSA-eligible health plan EBRI is sup ? The HSA-eligi ported b ble health plan was n y organizations from all industries and sectors that appreciate the value of ot associated with a change in the percentage of adults receiving HSA-eli • gible Need a number? Check out the health plan had a negative effect on EBR medica I Databook on Employee Benefits. tion monitoring for adults on select m aintenance drugs (ACE Impact tenure, and household size. All equations were specified using probit. Standard errors were clustered at the household level. of the HSA-Eligible Health Plan on Use of Quality Measures ..................................................................... 10 Editorial Board: Prescription drug costs Dallas L. Salisbury, publisher; Stephen Blakely, editor. Any $576 views expressed in this p $602 ublication and th -1.50 ose o 0.16 f the authors should most prior research on CDHPs has focused on HRA-based plans. Unfortunately, however, published papers only note Committee for Quality Assurance (NCQA) to derive Healthcare Effectiveness Data and Information Set (HEDIS) quality . Statistical significance denoted as follows: *** p<0.01; ** p<0.05; * p<0.10 "Medication Utilization and Adherence in a Health Savings Account-Eligible Plan." American Journal of Managed and maximum out-of-pocket limits are indexed to inflation. Certain primary preventive services can be exempted from Statistical significance denoted as follows: *** p<0.01; ** p<0.05; * p<0.10 group an unbiased, reliable information on emplo HbA1c testing d the comparison until the four group. In 2 th year 006, individu . LDL testing was l als in yee b clude enefits. d in t ower as a resu he Vis study ha it www.e d a lt of the intr n b average ri.org/about/join/ of oduction of th 34–38 hospita for more. e HSA- lizations not be ascribed to the officers, trustees, members, or other sponsors of the Employee Benefit Research Institute, the EBRI Education and Total healthcare costs $2,407 $2,498 -1.40 0.17 inhibitors, ARBs, digoxin, diuretics, and anticonvulsants) not only in the first year that the new health plan was Study Limitations ................................................................................................................................................. 12 Current Work that they are examining “CDHPs”, without specifying whether the CDHP is an HRA-based plan, an HSA-eligible health measures. In • Instan fact, these dependent varia tly get e-mail noti?cations of the latest EB bles were selected from the la RI data, rger set surveys, publications, and m of all HEDIS measures beca eetings use they Resear Care ch Fund, 19, no. or 1 their 2 (Dec staffs. embe Nothin r 201 g her 3b ein is to be co ): e400-e40 nstr7. ued as an attem pt to aid or hinder the adoption of any pending legislation, regulation, the ded uctible and covered in full, typically those deemed to prevent the onset of disease (Fendrick et al., 2014). (per 1,000 individuals Total healthcare costs paid by plan sponsor ), 105–131 inpatient hospital days (per $1,948 1,000 individuals), $2,001 137–139 emer -0.90 genc0.40 y department visits eligible health plan in all years. introduced, but the following three years as well. Study Limitations plan, o Conclusions or interpretative r both. ............................................................................................................................... Also, studies t rule, or as legal, hat have accounting, e actuarial, o xamined CDHPs r other such prof may have o essional advice. nly beww en w. exami ebri.o n rg in g a health plan w .......................... ith a high 13 This wereIssue Brief calculable from builds on admi th niis earli strative claims er work on C dataD only. I HPs an nd use of he cluded were a th lth care services related to e following measures: quality measures. The and sem Source: Employee Benefit Research Institute estimates based on administrative claims data. inars by clicking on the “Notify Me” or “RSS” buttons at the top of our home page. (per 1,000 individuals), about 3.5 outpatient physician’s office and clinic visits per person, and just over 10 prescription deductible; in other words, Haviland, AmeNotes: lia M. et al. "H aow Do health Consu plan th mat er-Dir was not ected associ Health ated wit Plans A h a ffect savings-acco Vulnerable Po unt opt pulatio ion. ns?" While t Forum for he earlyH studies ealth prior stu Unlike prior r dies e examined search onearly heal experi th careence wi qualityth CDHPs. O measures an fte study er the intro (Havil duand ction of et al., a C 2011) used data DHP that has foc from as recentl used mostly on y HRA as Appendix ............................................................................................................................................................ 15 Measures a drug fills per ddr person (30-day EBR essing avoi I Issue Brief da is r nce o adjusted). egister f service ed in the U. s that are oft S. Patent and Tre adem n overused, s ark Office. ISu SN: 0887 ch as t ? h13 e 7X/90 use of a 0887 ntibiotic ?137X/90 $ . prescriptions 50+.50 for adults * Health savings account. There’s lots more! on CDHPs Economics and Policy that were not spe 14, cific as no. 2 to (201 whe 1): ther th 1-12.ey e xamined an HRA-based plan or HSA-eligible health plan most Office Visits for Adults and Children 200 plans, th 7, but mos is studty uses used data data o from 20 n a full 03 -re –20 plac 06 e. T ment, HSA his Issue Brief -eligible ex healt amine h s data plan. It is from also a 2006b –2 le to fol 010. low over 18,000 enrollees References .......................................................................................................................................................... 16 Values are proportions unless denoted otherwise. with acute bronchitis, and imaging for adults with low back pain, were also examined. In both cases, the services are th likely used data on HRA-based plans, that is not n1100 13 ecessarily the Street case. NW · Suite 878 Not kno wing what type of CDHP was examined over five years, uses a com Reported p-values from t-test of differences in means across groups. parison grou Visit EBRI on-line today: p, and is the longest study of www.ebri.org consumer-dire cted health plans compared to the Mean overall health spending, including both the employer and member shares, were $213–$228 for inpatient services, ? Well-child/adult physical visit. often Endnotes unn ...................................................................................................................... ecessary according to clinical practice guidelines, and thus are used as quality indicators ...................................... by NCQA and 17 © 2014, Employee Benefit Research Institute Washington, DC ?Education and Re 20005 search Fund. All rights reserved. makes it difficult to draw conclusions about prior work, for the reasons outlined earlier. previous studies discussed. $89–92 for emergency department, $294–$296 for outpatient physician’s office and clinic visits, and $576–602 for (202) 659-0670 www.ebri.org www.choosetosave.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 A monthl ebri.org ebri.org Issue Brief • September 2014 • No. 404 ebri.org Issue Brief • September 2014 • No. 404 ebri.org Issue Brief • September 2014 • No. 404 Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri Issue Bri y research report f e e e e e e e e e e e e efffffffffffff • September • September • September • September • September • September • September • September • September • September • September • September • September rom the EBRI Education and Re 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 2014 • No. 404 search Fund © 2 014 Em ployee Benefit Research Institute 11 14 9 16 12 15 17 13 10 8 6 3 2 7 4 5