Effect of the Women’s Health Initiative on Osteoporosis Therapy andExpenditure in Medicaid
Jacob A Udell,1,2 Michael A Fischer,1 M Alan Brookhart,1 Daniel H Solomon,1,3 and Niteesh K Choudhry1
ABSTRACT: Decreasing HRT use among postmenopausal women may have a reciprocal impact on otherosteoporosis therapy. Time series analysis of prescribing trends for millions of Medicaid beneficiaries revealeda 57% decline in HRT without augmenting the pace of bisphosphonate use. Prescribing changes dramaticallyincreased Medicaid spending on osteoporosis therapy over the last decade and requires further evaluation ofcost effectiveness.
Introduction: Hormone replacement therapy (HRT) has been commonly prescribed to postmenopausalwomen, but its use is decreasing because adverse cardiac outcomes were reported by the Womens HealthInitiative (WHI) in July 2002. The reciprocal impact of the WHI on other osteoporosis medications use andexpenditure is unknown.Materials and Methods: We conducted a time series analysis on prescription data from 50 state Medicaidprograms between 1995 and 2004. Five medication categories were used: HRT, bisphosphonates, calcium,calcitonin, and raloxifene.Results: HRT was increasing before publication of the WHI, reaching 5 million prescriptions per year bymid-2002 (136 prescriptions per 1000 beneficiaries). Bisphosphonate prescribing rose in parallel until mid-2002. WHI publication was associated with a rapid reduction in HRT use, declining 57% by mid-2004 to anaverage of 59 prescriptions per 1000 beneficiaries (p � 0.01). WHI publication did not augment bisphospho-nates’ nearly linear rate of rise (p � 0.43) as their prescribing pace continued, whereas HRT declined.Medicaid spending on osteoporosis therapy also changed dramatically during the last decade, as yearlyexpenditure increased 664% from $1465 to $9742 per 1000 beneficiaries. Over this period, a significant shiftfrom daily to weekly bisphosphonates also occurred.Conclusions: A dramatic decline in HRT and continued rise in bisphosphonate prescribing has occurred sincethe publication of the WHI. During this time, there have also been substantial increases in osteoporosismedication spending within Medicaid. Determining whether these trends are clinically appropriate and costeffective for osteoporosis therapy will have important implications for the development of future drug reim-bursement programs, especially for elderly patients.J Bone Miner Res 2006;21:765–771. Published online on January 30, 2006; doi: 10.1359/JBMR.060119
Key words: estrogen, bisphosphonate, osteoporosis, Medicaid, women’s health
INTRODUCTION
OSTEOPOROSIS IS ASSOCIATED with significant morbidityand mortality and imposes a substantial economic
burden on the health care system.(1–3) Unfortunately, os-teoporosis therapy is often underused in high-risk pa-tients.(4,5)
Until the publication of the Women’s Health Initiative(WHI) in July 2002, hormone replacement therapy (HRT)
was a common pharmacologic approach to osteoporosistreatment in the United States.(6–10) The WHI showed thatHRT use increased the risk of breast cancer, cardiovasculardisease, and dementia, despite its benefit in osteoporosisprevention.(11) Accordingly, substantial reductions in HRTuse have been documented.(12–19)
Little is known about the reciprocal impact of the WHIand subsequent revisions to HRT recommendations(20,21)
on the use of other drug treatments for osteoporosis. Weanalyzed data from state Medicaid programs to evaluatethe effect of the WHI on trends in osteoporosis drug useand expenditure. We hypothesized that reductions in HRTuse should be accompanied by accelerated increases in the
Dr Solomon receives salary support from Merck and Proctor andGamble for unrelated activities. All other authors state that theyhave no conflicts of interest.
1Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and HarvardMedical School, Boston, Massachusetts, USA; 2Department of Medicine, Vancouver Hospital and Health Sciences Centre and Universityof British Columbia, Vancouver, British Columbia, Canada; 3Division of Rheumatology, Department of Medicine, Brigham and Women’sHospital and Harvard Medical School, Boston, Massachusetts, USA.
use of other medications, most notably bisphosphonates.Because there is a marked cost difference between HRTand bisphosphonates,(22) these changes in practice patternsmay have profound economic implications.
MATERIALS AND METHODS
Sources of data
Our analysis used aggregate, state-level data on quarterlydrug use by Medicaid recipients. Medicaid, an Americanhealth insurance entitlement program that is jointly fundedby state and federal governments,(23) provides health careservices and prescription drug coverage for >16 millionwomen in the United States. In 2004, this represented ∼9%of American women(24) and one third of all low-incomewomen.(25) To qualify for Medicaid, a woman must earn<200% of the federal poverty level (approximately $15,000)and be either pregnant, a mother of a child <18 years of age,�65 years of age, or have a physical impairment.(24) Eightypercent of out-patient Medicaid prescription expendituresare for drugs consumed by enrollees �65 years of age andyounger persons with disabilities,(26) and overall, 71% ofMedicaid beneficiaries �65 years of age are women.(24)
Our data were provided by the Center for Medicare andMedicaid Services (CMS)(27) and included the total numberof prescriptions filled, total units of medication dispensed,and total Medicaid reimbursement for each medication.Data were available for 49 states and the District of Co-lumbia. No data were available for Arizona. The CMS data,grouped by National Drug Code (NDC), were linked to theNational Drug Data File to incorporate detailed informa-tion about drug strength, dose type (e.g., extended releasecapsule, tablet), and therapeutic class.(22) We also obtaineddata on the number and demographic characteristics ofMedicaid recipients.(28,29) No data about individual patientswere used in this analysis. This study was approved by theinstitutional review board of Brigham and Women’s Hos-pital.
Osteoporosis medication categories
We grouped osteoporosis medications into five catego-ries: HRT, bisphosphonates, calcium, calcitonin, and theselective estrogen receptor modifier (SERM) raloxifene.HRT includes oral, transdermal, and injectable forms ofestrogen alone and estrogen-progesterone combinationtherapies. Bisphosphonates include oral, intravenous, andinjectable forms of etidronate, alendronate, risedronate,and pamidronate. Both daily and weekly oral dosing ver-sions of alendronate and risedronate were included. Thecalcium category includes calcium supplements and calciumcarbonate. Synthetic forms of vitamin D, PTH, and proges-terone alone were excluded from our analysis. The numberof prescriptions was reported quarterly by category fromJanuary 1995 through July 2004. The beginning of the studyperiod was chosen to coincide with the first availability of adaily oral bisphosphonate.
Data analysis
Using the categories described above, we calculated thetotal number of prescriptions filled, units of medication dis-
TA
BL
E1.
YE
AR
LY
OST
EO
PO
RO
SIS
PR
ESC
RIP
TIO
NS,
ME
DIC
AID
1995
–200
4
Yea
rB
enef
.(m
illio
ns)
Per
cent
age
�65
year
sT
otal
OP
drug
spr
escr
ibed
Tot
alO
Pdr
ugs
per
1000
bene
f.
Pre
scri
ptio
ns
HR
TB
isph
osph
onat
esR
alox
ifen
eC
alci
toni
nC
alci
um
No.
Per
cent
ofto
tal
No.
Per
cent
ofto
tal
No.
Per
cent
ofto
tal
No.
Per
cent
ofto
tal
No.
Per
cent
ofto
tal
1995
36.3
11.6
2,86
0,29
279
2,31
8,78
881
.142
,819
1.5
00
31,2
811.
146
7,40
416
.319
9942
.910
.46,
189,
978
144
3,57
6,85
457
.877
8,78
412
.623
7,44
63.
862
3,66
810
.197
3,22
615
.720
0044
.510
.37,
598,
880
171
4,18
8,38
155
.11,
031,
437
13.6
431,
588
5.7
794,
636
10.5
1,15
2,83
815
.220
0148
.49.
99,
630,
628
199
4,93
9,23
851
.31,
726,
695
17.9
591,
294
6.1
871,
550
9.0
1,50
1,85
115
.620
0251
.49.
510
,703
,137
208
4,81
6,66
945
.02,
575,
696
24.1
713,
236
6.7
856,
190
8.0
1,74
1,34
616
.320
0353
.39.
410
,613
,029
199
3,68
3,62
534
.73,
393,
176
32.0
798,
374
7.5
804,
611
7.6
1,93
3,24
318
.220
04*
54.6
9.3
10,5
90,6
4419
43,
194,
512
30.2
3,84
7,03
636
.378
6,54
67.
472
1,14
86.
82,
041,
402
19.3
*20
04in
clud
esan
nual
ized
figu
res
base
don
Janu
ary
thro
ugh
July
2004
.Ben
efic
iary
data
are
from
the
Cen
ter
for
Med
icar
ean
dM
edic
aid
Serv
ices
;pre
scri
ptio
nus
efr
omst
ate
drug
use
data
.Bal
tim
ore,
MD
:Cen
ters
for
Med
icar
ean
dM
edic
aid
Serv
ices
,200
4.(2
7–29)
OP
,ost
eopo
rosi
s;H
RT
,hor
mon
ere
plac
emen
tth
erap
y;N
o.,v
olum
eof
pres
crip
tion
s;B
enef
.,nu
mbe
rof
Med
icai
dbe
nefi
ciar
ies
per
year
.
UDELL ET AL.766
pensed, and dollar expenditure by Medicaid nationwide ineach calendar quarter for each category of medication. Toaccount for changes over time in the number of Medicaidbeneficiaries, we present medication use and spending per1000 beneficiaries. We calculated the quarterly and annualrates of change in each of these measures of prescribing.For the bisphosphonates, we performed similar calculationsto describe the total use of daily or weekly formulations ofthese agents.
We modeled the trends in use of HRT and bisphospho-nates using autoregressive integrated moving average(ARIMA) models, incorporating a ramp function to esti-mate the effect of the publication of the WHI results andreporting significance at the p < 0.05 level. Models withboth total prescriptions and total Medicaid spending as thedependent variable were developed. We controlled forchanges in the age distribution of the Medicaid populationby including the proportion of Medicaid beneficiaries whowere >65 years of age as an additional independent variablein the ARIMA models. Data on the total number of ben-eficiaries and number of elderly beneficiaries were avail-able on an annual basis. We used linear interpolation togenerate the quarter specific proportion of elderly benefi-ciaries. Statistical analyses were conducted using STATA,version 8.0 (Stata Corp., College Station, TX, USA) andSAS software (version 8.0; SAS Institute, Cary, NC, USA).
RESULTS
The number of Medicaid beneficiaries included in ouranalysis increased from 36.3 million in 1995 to 54.6 millionin 2004 (Table 1). During this time, the proportion of totalfemale Medicaid beneficiaries was ∼60% (63% of duallyeligible Medicaid/Medicare beneficiaries were female),whereas the percentage of beneficiaries �65 years of agedecreased from 11.6% to 9.3% between 1995 and 2004.
Trends in drug use
During the study period, the annual number of prescrip-tions for osteoporosis medications that were filled increased
from ∼2.9 to 10.6 million. This represents a rise from 79prescriptions per 1000 beneficiaries in 1995 to 194 prescrip-tions per 1000 beneficiaries in 2004 (246%). Notably, totalosteoporosis drug use peaked in 2002 (208 prescriptions per1000 beneficiaries) and declined subsequently by 4.6% in2003 (199 prescriptions per 1000 beneficiaries) and a further2.6% in 2004 (194 prescriptions per 1000 beneficiaries).
Considerable differences in prescribing among osteopo-rosis medication categories were observed (Table 1). Be-fore the publication of the WHI, prescriptions dispensed forHRT had increased steadily, peaking at ∼5 million per year(136 prescriptions per 1000 beneficiaries) in 2002. After therelease of the WHI in July 2002, HRT prescribing declinedrapidly, representing a 57% reduction by mid-2004 to 59prescriptions written per 1000 beneficiaries (Fig. 1; p �
0.01). Prescribing rates for bisphosphonates rose approxi-mately in parallel with HRT prescriptions after the intro-duction of a daily oral formulation in 1995. Bisphosphonateprescribing rates have continued to steadily increase with-out any change in their rate of rise after publication of theWHI (p � 0.43).
Changes in the types of bisphosphonates prescribed dur-ing the study period were also observed. By 2001, bothalendronate and risedronate were made available in once-weekly oral formulations, and both versions rapidly rose inmarket share from 19% in 2001 to 88% by mid-2004. Re-ciprocally, the use of once-daily alendronate and risedro-nate sharply declined from ∼74% to 8% of market share.Etidronate and pamidronate were minimally prescribed be-tween 1991 and 2004. The rise in use of once-weekly oralbisphosphonates did not significantly change as a result ofthe WHI publication.
Slower rates of growth were observed for calcium, ral-oxifene, and calcitonin between 1995 and 2004 (Fig. 1). Cal-cium prescribing peaked at 31 prescriptions per 1000 ben-eficiaries in 2001, but the rate of growth of calciumprescribing has slowed to 3% in 2004 (37 prescriptions per1000 beneficiaries). Calcitonin prescribing peaked in 2001at 18 prescriptions per 1000 beneficiaries and has subse-
FIG. 1. Quarterly volume of Medicaid pre-scriptions for osteoporosis therapy by drugclass, January 1995 to July 2004. �, HRT; ◊,bisphosphonates; X, calcium; �, calcitonin;�, raloxifene.
EFFECT OF THE WHI ON OSTEOPOROSIS TREATMENT 767
quently declined 27%. Raloxifene prescribing seems tohave peaked in 2003 (15 prescriptions per 1000 beneficia-ries).
Trends in drug expenditure
The annual expenditure on osteoporosis medications in-creased from approximately $53 to $532 million, represent-ing a 664% rise from $1465 to $9742 per 1000 beneficiaries.Trends in the cost of osteoporosis therapy are presented inTable 2. Steady increases in expenditure for each class ofmedications were observed before the publication of theWHI (Fig. 2). The publication of the WHI in July 2002 wasassociated with a rapid decline in HRT expenditure (p �0.0006) but had no effect on the rise in bisphosphonateexpenditures (p � 0.21). Bisphosphonate spending in-creased from $2540 per 1000 beneficiaries per year in 2001when once-weekly oral bisphosphonates were first madeavailable to $5400 per 1000 beneficiaries per year in 2004.
DISCUSSION
In the United States Medicaid program, we observedsubstantial reductions in HRT prescribing after the publi-cation of the WHI. These results are consistent with previ-ously published finding(12–19) and show a persistent declinein HRT use through 2004. We expected that decreased hor-mone therapy after the report of the WHI would have beenaccompanied by a reciprocal acceleration in the rate ofbisphosphonate prescribing, presumably because a propor-tion of those patients on HRT were benefiting from osteo-porosis protection. Although prescriptions for bisphospho-nates, raloxifene, and calcium did rise after the WHI, ourresults indicate that the publication of the WHI did notaugment their rate of increase. In addition, overall rates ofosteoporosis drug use may have declined slightly in the past2 years.
Nevertheless, we did observe an increase in the overallrate of bisphosphonate prescribing within our observationperiod and a significant shift since 2000 from once-dailybisphosphonates toward once-weekly formulations, pre-sumably based on the assumption that less frequent admin-istration of an ongoing medication may improve patientcompliance.(30) This overall increase in medication use, ifappropriate, is likely cost-saving to the health care systembecause of avoided fractures and their complications.(31–34)
However, whether the rate of bisphosphonate prescribingeffectively compensates for decreases in HRT use is un-clear.
The striking rise in drug expenditures that we observedhighlights the concern that certain bisphosphonate formu-lations can be cost-prohibitive drug choices when absoluterisk reduction and compliance are considered.(35,36) Thesetrends likely represent the natural result of removing HRTas a choice in the pharmacological armamentarium for os-teoporosis treatment and provide insight into the conse-quences of eliminating a member of a therapeutic class onrates of use and expenditures for other drugs within theclass. For example, the recent withdrawal of rofecoxib andvaldecoxib could lead to reductions in expenditures of os-
TA
BL
E2.
YE
AR
LY
OST
EO
PO
RO
SIS
DR
UG
EX
PE
ND
ITU
RE
,ME
DIC
AID
1995
–200
4
Yea
rB
enef
.(m
illio
ns)
Per
cent
age
�65
year
s
Tot
alex
pend
iture
for
OP
drug
s($
mill
ions
)
Ave
rage
OP
drug
cost
per
1000
bene
f.($
)
Dru
gex
pend
iture
HR
TB
isph
osph
onat
esR
alox
ifen
eC
alci
toni
nC
alci
um
Dol
lars
(mill
ions
)P
erce
ntof
tota
lD
olla
rs(m
illio
ns)
Per
cent
ofto
tal
Dol
lars
(mill
ions
)P
erce
ntof
tota
lD
olla
rs(m
illio
ns)
Per
cent
ofto
tal
Dol
lars
(mill
ions
)P
erce
ntof
tota
l
1995
36.3
11.6
$53.
18$1
,465
$44.
4783
.6$2
.93
5.5
$00
$2.3
24.
4$3
.46
6.5
1999
42.9
10.4
$191
.10
$4,4
54$8
4.49
44.2
$52.
6627
.6$1
5.00
7.8
$31.
6716
.6$7
.28
3.8
2000
44.5
10.3
$261
.92
$5,8
85$1
09.3
641
.8$7
3.10
27.9
$28.
0710
.7$4
2.74
14.6
$8.6
53.
320
0148
.49.
9$3
60.0
5$7
,439
$138
.06
38.3
$123
.05
34.2
$39.
6511
.0$4
2.74
16.3
$10.
572.
920
0251
.49.
5$4
39.4
3$8
,549
$142
.12
32.3
$185
.44
42.2
$50.
0611
.4$4
8.72
13.5
$11.
302.
620
0353
.39.
4$4
93.9
0$9
,266
$118
.02
23.9
$252
.60
51.1
$59.
9812
.1$5
0.51
11.5
$11.
402.
320
04*
54.6
9.3
$531
.93
$9,7
42$1
06.6
220
.0$2
96.3
355
.7$6
2.29
11.7
$55.
2110
.4$1
1.49
2.2
*20
04in
clud
esan
nual
ized
figu
res
base
don
Janu
ary
thro
ugh
July
2004
.Ben
efic
iary
data
are
from
the
Cen
ter
for
Med
icar
ean
dM
edic
aid
Serv
ices
;pre
scri
ptio
nex
pend
itur
esfr
omst
ate
drug
use
data
.Bal
tim
ore,
MD
:C
ente
rsfo
rM
edic
are
and
Med
icai
dSe
rvic
es,2
004.
(27–29)
OP
,ost
eopo
rosi
s;H
RT
,hor
mon
ere
plac
emen
tth
erap
y;B
enef
.,nu
mbe
rof
Med
icai
dbe
nefi
ciar
ies
per
year
.
UDELL ET AL.768
teoarthritis therapy through the greater use of acetamino-phen and nonselective nonsteroidal anti-inflammatorydrugs, assuming that there are no changes in rates of use ofgastroprotective agents.
Osteoporosis therapy expenditure trends may strain thesustainability of drug insurance programs for the elderly, inparticular those that are publicly funded. The increasingnumber of eligible Medicaid recipients(29) and the imple-mentation of the new Medicare Part D drug benefit pro-gram(37) may add to the challenge. Therefore, cost-containment strategies that include rational policies forosteoporosis drug use, such as the recently announced de-cision to use preferred drug lists within the new Medicaredrug benefit program,(37) should be further considered.
Reported trends in HRT use within the largest U.S.health insurer(23) are consistent with findings among inter-national public health insurance programs. Within a year ofthe WHI publication, significant HRT declines of 30–60%were shown within Canada,(12) New Zealand,(14) Austra-lia,(16) Chile,(17) Hong Kong,(18) and Israel.(19) Two of thesestudies were large provincial cohorts(12,18) that analyzedHRT use over time and as a function of age, although nei-ther compared trends in HRT use in the context of otherosteoporosis medication. Analyses of nationally represen-tative U.S. databases detailing retail prescribing and physi-cian office visits for HRT and osteoporosis treatment foundthat HRT use for osteoporosis treatment began to declineafter bisphosphonates were introduced in 1994(38) and thatHRT use overall declined 38% in the first year after pub-lication of the WHI.(15) Qualitatively, our data suggest thattotal HRT use decelerated at a slower pace and later datewithin Medicaid; however, overall trends are consistentwith earlier analyses.
There may have been other historical events that influ-enced the use of osteoporosis therapy, such as broadeningof osteoporosis screening. For example, the Bone MassMeasurement Act of July 1, 1998,(39) mandated reimburse-ment coverage for BMD screening within Medicare, thefederal health insurance program for those �65 years ofage. In contrast, we studied patients enrolled in Medicaid,
which is regulated on a state-by-state basis. As a conse-quence, there is some variation across the program withrespect to BMD testing regulations,(40) and only 14 stateshave a policy mandating private insurance coverage screen-ing for osteoporosis similar to the Bone Mass MeasurementAct.(40)
The use of Medicaid data to describe osteoporosistherapy patterns has certain limitations. No data about in-dividual patients were used in this analysis; hence, our datadescribe aggregate out-patient medication use only. Pre-scription use cannot be specifically linked to individualclinical characteristics, particularly age or sex, and thus wecannot present stratum specific use rates nor can we disen-tangle whether there were changes in the proportion ofpatients that were receiving both HRT and bisphospho-nates before compared with after the publication of theWHI. Therefore, whereas our results are suggestive, we arelimited in our ability to comment on whether the observedtrends are clinically appropriate. In addition, we cannot ad-just for variability in individual consumption or distinguishbetween medication use for other indications, particularlyestrogen use for menopausal symptoms. However, ∼38% ofpostmenopausal women in the United States used HRT tomanage symptoms of menopause by the end of the 1990s,and the potential preventive effects of HRT on osteoporo-sis and cardiovascular outcomes were increasingly a consid-eration for its continuation.(10)
Our study was limited to Medicaid beneficiaries, which islargely composed of low-income women,(24–26,28,29) onefifth of whom are �65 years of age. Low-income women arereported to use HRT less frequently for menopausal symp-tom relief than other American women.(10,41,42) As a result,their use of HRT for osteoporosis therapy may differslightly from those of other elderly American women.
Finally, the apparent lag in calcium use as an adjunct toother appropriate osteoporosis prescribing may be an arti-fact of our inability to evaluate nonprescription calciumuse. That said, if Medicaid calcium coverage policies haveremained constant over the study period, presumably this
FIG. 2. Quarterly Medicaid spending onosteoporosis therapy by drug class, January1995 to July 2004. �, HRT; ◊, bisphospho-nates; X, calcium; �, calcitonin; �, raloxif-ene.
EFFECT OF THE WHI ON OSTEOPOROSIS TREATMENT 769
under-reporting bias should have remained constant aswell,(38) and therefore, it is possible that calcium is beingunderused.
In summary, because publication of the WHI made clearthat HRT was not ideal pharmacotherapy for osteoporosis,a dramatic decline in its use has occurred within Medicaid.In contrast, the WHI did not augment the steady rate of risein bisphosphonate prescribing. Nevertheless, substantial in-creases in spending on osteoporosis medication have alsooccurred. This trend has had an unintended, and perhapsunavoidable, impact on medication spending by state Med-icaid programs as more costly forms of osteoporosis medi-cation are used. Whether this trend in drug use compen-sates postmenopausal women at risk for osteoporoticfracture is unclear. Future research and policy planningshould consider trends in drug use and total expenditureswhen developing osteoporosis therapy strategies for pa-tients and drug-reimbursement programs, especially for el-derly patients.
ACKNOWLEDGMENTS
Dr Choudhry had full access to all of the data in the studyand takes responsibility for the integrity of the data and theaccuracy of the data analysis. Dr Udell was supported by aCanada-United States Fulbright Fellowship. Dr Solomonwas supported by National Institutes of Health GrantK23AR48616. The Fulbright Program and National Insti-tutes of Health had no involvement in the design and con-duct of the study; collection, management, analysis, andinterpretation of the data; and preparation, review, or ap-proval of the manuscript.
REFERENCES
1. National Osteoporosis Foundation 2003 America’s BoneHealth: The State of Osteoporosis and Low Bone Mass in OurNation. National Osteoporosis Foundation, Washington, DC,USA.
2. Tosteson AN, Hammond CS 2002 Quality-of-life assessment inosteoporosis: Health-status and preference-based measures.Pharmacoeconomics 20:289–303.
3. Cheung AM, Feig DS, Kapral M, Diaz-Granados N, Dodin S2004 Prevention of osteoporosis and osteoporotic fractures inpostmenopausal women: Recommendation statement from theCanadian Task Force on Preventive Health Care. Can MedAssoc J 170:1665–1667.
4. Solomon DH, Finkelstein JS, Katz JN, Mogun H, Avorn J 2003Underuse of osteoporosis medications in elderly patients withfractures. Am J Med 115:398–400.
5. Andrade SE, Majumdar SR, Chan KA, Buist DS, Go AS,Goodman M, Smith DH, Platt R, Gurwitz JH 2003 Low fre-quency of treatment of osteoporosis among postmenopausalwomen following a fracture. Arch Intern Med 163:2052–2057.
6. Nelson HD, Rizzo J, Harris E, Cauley J, Ensrud K, Bauer DC,Orwoll E; for the Study of Osteoporotic Fractures ResearchGroup. 2002 Osteoporosis and fractures in postmenopausalwomen using estrogen. Arch Intern Med 162:2278–2284.
7. Genant HK, Cooper C, Poor G, Reid I, Ehrlich G, Kanis J,Nordin BE, Barrett-Connor E, Black D, Bonjour JP, Dawson-Hughes B, Delmas PD, Dequeker J, Ragi Eis S, Gennari C,Johnell O, Johnston CC Jr, Lau EM, Liberman UA, LindsayR, Martin TJ, Masri B, Mautalen CA, Meunier PJ, Miller PD,Mithal A, Moril H, Papapoulos S, Woolf A, Yu W, Khaltaer N
1999 Interim report and recommendations of the WorldHealth Organization Task-Force for Osteoporosis. OsteoporosInt 10:259–264.
8. NIH Consensus Development Panel on Osteoporosis Preven-tion 2001 Diagnosis, and Treatment. Osteoporosis prevention,diagnosis, and therapy. JAMA 285:785–795.
9. Lee E, Zuckerman IH, Weiss SR 2002 Patterns of pharmaco-therapy and counseling for osteoporosis management in visitsto US ambulatory care physicians by women. Arch Intern Med162:2362–2366.
10. Keating NL, Cleary PD, Rossi AS, Zaslavsky AM, Ayanian JZ1999 Use of hormone replacement therapy by postmenopausalwomen in the United States. Ann Intern Med 130:545–553.
11. Rossouw JE, Anderson GL, Prentice RL, LaCroix AZ,Kooperberg C, Stefanick ML, Jackson RD, Beresford SA,Howard BV, Johnson KC, Kotchen JM, Ockene J; for theWriting Group for the Women’s Health Initiative Investiga-tors. 2002 Risks and benefits of estrogen plus progestin inhealthy postmenopausal women: Principal results from theWomen’s Health Initiative randomized controlled trial. JAMA288:321–333.
12. Austin PC, Mamdani MM, Tu K, Jaakkimainen L 2003 Pre-scriptions for estrogen replacement therapy in Ontario beforeand after publication of the Women’s Heath Initiative study.JAMA 289:3241–3242.
13. Haas JS, Kaplan CP, Gerstenberger EP, Kerlikowske K 2004Changes in the use of postmenopausal hormone therapy afterthe publication of clinical trial results. Ann Intern Med140:184–188.
14. Lawton B, Rose S, McLeod D, Dowell A 2003 Changes in useof hormone replacement therapy after the report from theWomen’s Health Initiative: Cross sectional survey of users.BMJ 327:845–846.
15. Hersh AL, Stefanick ML, Stafford RS 2004 National use ofpostmenopausal hormone replacement therapy: Annual trendsand response to recent evidence. JAMA 291:47–53.
16. MacLennan AH, Taylor AW, Wilson DH 2004 Hormonetherapy use after the Women’s Health Initiative. Climacteric7:138–142.
17. Blumel JE, Castelo-Branco C, Chedraui PA, Binfa L, DowlaniB, Gomez MS, Sarra S 2004 Patients’ and clinicians’ attitudesafter the Women’s Health Initiative Study. Menopause 11:57–61.
18. Leung KY, Ling M, Tang GWK 2005 Use of hormone replace-ment therapy in the Hong Kong public health sector after theWomen’s Health Initiative trial. Maturitas 52:277–285.
19. Keinan-Boker L, Dictiar R, Garty N, Green MS 2005 Preva-lence and correlates of hormonal therapy among Israeliwomen in the post-WHI era. Maturitas 52:364–373.
20. Wathen CN, Feig DS, Feightner JW, Abramson BL, CheungAM 2004 Hormone replacement therapy for the primary pre-vention of chronic diseases: Recommendation statement fromthe Canadian Task Force on Preventive Health Care. Can MedAssoc J 170:1535–1539.
21. US Preventive Services Task Force 2002 Hormone replace-ment therapy for primary prevention of chronic conditions:Recommendations and rationale. Available at http://www.ahrq.gov/clinic/3rduspstf/hrt/hrtrr.htm. Accessed Decem-ber 1, 2004.
22. Drug Topics Red Book 2004 Thomson PDR, Montvale, NJ,USA.
23. Perkins J 2004 “Medicaid,” in Poverty Law Manual for theNew Lawyer. National Center on Poverty Law, Chicago, IL,USA.
24. Anonymous 2004 Medicaid’s role for women. Available athttp://www.kff.org/womenshealth/upload/Medicaid-s-Role-for-Women.pdf. Accessed December 1, 2004.
25. National Family Planning & Reproductive Health Association2004 Medicaid: A vital source of health care for low incomewomen and a critical source of support for family planningservices. Available at http://www.nfprha.org/pac/factsheets/medicaid.asp. Accessed December 1, 2005.
26. Bruen B, Ghosh A 2004 Medicaid prescription drug spending
UDELL ET AL.770
and use. Kaiser Commission on Medicaid and the uninsured.Available at http://www.kff.org/medicaid/upload/Medicaid-Prescription-Drug-Spending-and-Use.pdf. Accessed Decem-ber 1, 2004.
27. Centers for Medicare and Medicaid Services 2004 State DrugUtilization Data. Centers for Medicare and Medicaid Services,Baltimore, MD, USA.
28. Centers for Medicare and Medicaid Services 2004 MedicaidStatistical Information System Statistical Reports for FederalFiscal Years 1999–2004. Centers for Medicare and MedicaidServices, Baltimore, MD, USA.
29. Centers for Medicare and Medicaid Services 2004 Medicaidenrollment and beneficiaries, selected fiscal years. Available athttp://www.cms.hhs.gov/researchers/pubs/datacompendium/2003/03pg34.pdf. Accessed December 1, 2004.
30. de Klerk E 2001 Patient compliance with enteric-coatedweekly fluoxetine during continuation treatment of major de-pressive disorder. J Clin Psychiatry 62:S43–S47.
31. Gabriel SE, Tosteson AN, Leibson CL, Crowson CS, PondGR, Hammond CS, Melton LJ III 2002 Direct medical costsattributable to osteoporotic fractures. Osteoporos Int 13:323–330.
32. Kirk D, Fish DA 2004 Medical management of osteoporosis.Am J Manag Care 10:445–455.
33. Johnell O, Jonsson B, Jonsson L, Black D 2003 Cost effective-ness of alendronate (Fosamax) for the treatment of osteopo-rosis and prevention of fractures. Pharmacoeconomics 21:305–314.
35. Curtiss FR 2003 ERT, HRT, raloxifene, calcitonin, or bisphos-phonates for osteoporosis. J Manag Care Pharm 9:178–180.
36. Curtiss FR 2004 Reducing the risk of fractures—the optionsnarrow for cost-effective therapeutic alternatives. J ManagCare Pharm 10:168–179.
37. Pear R 2005 New Medicare Rules on Drugs Balance Accessagainst Costs. New York Times. A5, January 22, 2005.
38. Stafford RS, Drieling RL, Hersh AL 2004 National trends inosteoporosis visits and osteoporosis treatment, 1988–2003.Arch Intern Med 164:1525–1530.
39. Federal Register 1998 63 FR 34320-34328. Medicare program:Medicare coverage of and payment for bone mass measure-ments. Available at http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?IPaddress�&dbname�1998_register&docid�98-16783-filed. Accessed December 1, 2005.
40. The Henry J. Kaiser Family Foundation State Health FactsWeb Site 2004 50 state comparisons: State mandated benefits:Osteoporosis screening. Available at http://statehealthfacts.org/cgi-bin/healthfacts.cgi?action�compare&category�Women%27s+Health&subcategory�Mandated+Benefits+for+Women&topic�Osteoporosis+Screening. Accessed De-cember 1, 2005.
41. Stafford RS, Saglam D, Causino M, Blumenthal D 1998 Thedeclining impact of race and insurance status on hormone re-placement therapy. Menopause 5:140–144.
42. Friedman-Koss D, Crespo CJ, Bellantoni MF, Anderson RE2002 The relationship of race/ethnicity and social class to hor-mone replacement therapy: Results from the Third NationalHealth and Nutrition Examination Survey 1988–1994. Meno-pause 9:264–272.
Address reprint requests to:Niteesh K Choudhry, MD, PhD
Division of Pharmacoepidemiology andPharmacoeconomics
Brigham and Women’s HospitalHarvard Medical School
1620 Tremont Street, Suite 3030Boston, MA 02120, USA
