- Email: [email protected]
Potentially inappropriate medication use in elderly Japanese patients
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Potentially Inappropriate Medication Use in Elderly Japanese Patients Manabu Akazawa, PhD, MPH1; Hirohisa Imai, MD, PhD2; Ataru Igarashi, PhD3; and Kiichiro Tsutani, MD, PhD3 1Drug
Management and Policy, Faculty of Pharmacy, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Ishikawa, Japan; 2Department of Epidemiology, National Institute of Public Health, Saitama, Japan; and 3Department of Drug Policy and Management, Graduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo, Japan
ABSTRACT Background: Modified Beers criteria for elderly Japanese patients were developed in 2008 by consensus among 9 experts to reflect regional clinical practice and available medications in Japan. Since then, many physicians and pharmacists have expressed interest in obtaining more information about the criteria and alternative drug choices. Objective: This study examined the incidence, health care utilization, and costs associated with potentially inappropriate medications (PIMs) in elderly patients based on the modified Beers criteria. Methods: A retrospective, observational cohort study was conducted using health insurance claims data in Japan. The study population included elderly patients aged ≥65 years who had at least 2 pharmacy claims in separate months over a 1-year period (April 2006 through March 2007). Use of the PIMs was identified using the modified criteria, and 1-year incidence rates were calculated for the total study population and for subgroups stratified by age and sex. A logistic regression model was used to examine demographic and clinical characteristics associated with PIMs. Health care utilization rates and costs were also analyzed and compared between patients with and without PIMs using generalized linear models. All models included dummy variables indicating age category, female sex, hospitalization, polypharmacy, index month, and number of Elixhauser comorbidities to adjust for potential confounders. Results: Among 6628 elderly patients, 71.2% (4721/6628) were female and 62.9% (4167/6628) were aged 65 to 74 years; 43.6% (2889/6628) were prescribed at least one PIM. The most commonly used PIMs were histamine-2 blockers (20.5% [1356/6628]), benzodiazepines (11.4% [756/6628]), and anticholinergics and antihistamines (7.9% [526/6628]). No significant differences in incidence rates were observed based on age or sex. Inpatient service use, polypharmacy, and comorbidities of peptic ulcer, depression, and cardiac arrhythmias were significant predictors of PIM use while controlling for other factors. PIM users had significantly higher hospitalization risk (1.68-fold), more outpatient visit days (1.18-fold), and higher medical costs (33% increase) than did nonusers. Conclusions: In a group of elderly Japanese patients, 43.6% used at least one PIM over a 1-year period in this study. PIM use was associated with greater health care utilization rates and costs. (Am J Geriatr Pharmacother. 2010;8:146–160) © 2010 Excerpta Medica Inc. Key words: Beers criteria, potentially inappropriate medication, incidence, utilization, cost, elderly, drug utilization review. 5IJTTUVEZXBTSFQPSUFEJOQBSUBTBQPTUFSQSFTFOUBUJPOBUUIFUI"OOVBM*OUFSOBUJPOBM.FFUJOHPGUIF*OUFSOBUJPOBM4PDJFUZGPS1IBSNBDPFDPOPNJDT BOE0VUDPNFT3FTFBSDI .BZo 0SMBOEP 'MPSJEB Accepted for publication February 23, 2010. ª&YDFSQUB.FEJDB*OD"MMSJHIUTSFTFSWFE
146
April 2010
7PMVNFt/VNCFS
EPJKBNKPQIBSN TFFGSPOUNBUUFS
M. Akazawa et al.
INTRODUCTION To adjust for increasing medical expenditures among the elderly in Japan, the government has initiated health care system reforms, including the development of a long-term care insurance for the elderly population.1 Elderly individuals are a heterogeneous population, and many of them have several chronic illnesses that require multiple medications, making them vulnerable to adverse drug events that contribute to elevated medical costs but are preventable.2–6 Appropriate pharmaceutical care for the elderly has been determined based on clinical trials conducted with younger adults. However, because of changes in pharmacokinetics and pharmacodynamics associated with aging, greater attention is needed in pharmacy management for patients aged >65 years.7 The Beers criteria are a consensus-based list of potentially inappropriate medications (PIMs) for individuals aged ≥65 years, first developed for nursing home residents in 1991 and updated twice in 1997 and 2003 for the general elderly population based on new clinical evidence.8–10 Medications or medication classes were selected for this list if potential risks outweighed potential benefits and if an effective alternative medication was available. The Beers criteria were widely used as tools for drug utilization reviews and outcome measures related to PIM use in geriatric health care in the United States, Europe, and Asian countries. These studies reported prevalence rates of PIM use ranging from 6% to 41%, depending on the study population and country.11–15 In some studies, the use of PIMs was associated with adverse outcomes, including increased health care utilization rates and costs16,17; in other settings, however, the results were inconclusive.18–20 These regional variations were explained by differences in medication availability, drug policy, and clinical practice. Some medications were not available or were rarely used in clinical settings. Thus, country-specific modifications of the criteria have been recommended to allow precise assessments of medication use and consequences. In addition to the Beers criteria, several other screening tools to evaluate PIMs have been published. The McLeod criteria were developed in Canada in 1997 using the Delphi consensus approach and included 38 medications explicitly contraindicated or likely to cause drug–drug or drug–disease interactions.21 This list was subsequently updated as the Improved Prescribing in the Elderly Tool (IPET) in 2000 and included 14 categories of contraindicated medications or possible drug–disease interactions.22 According to an
The American Journal of Geriatric Pharmacotherapy
observational study conducted in Ireland, IPET was a convenient tool for users but was not effective at identifying PIM use compared with the Beers criteria because it omitted too many well-recognized classes of drugs.23 Recently, a new tool, the Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions (STOPP), was introduced and evaluated in Europe.24,25 STOPP was developed by a Delphi consensus method and contained 65 medications or medication classes. The tool was found to be more sensitive than the Beers criteria in identifying elderly patients with PIMs because several medications listed in the Beers criteria are either not prescribed or not available in most European countries.14 In 2008, Imai et al26 developed modified Beers criteria for the elderly Japanese population. Using the methods employed to develop the Beers criteria and their updates, 9 expert panel members selected 47 medications or medication classes that should be generally avoided for all elderly patients. Medications listed in the modified criteria and differences from the 2003 Beers criteria are shown in the appendix. (Because the original report of the Japanese Beers criteria was written in Japanese, interested readers may request a provisional translation of the method section from the authors.) Since publication of the Japanese Beers criteria, many physicians and pharmacists have expressed an interest in obtaining more information about the criteria and alternative drug choices. The objective of this study was to conduct a drug utilization review based on the modified Beers criteria in Japan. In this article, we focused on medications or medication classes that should generally be avoided independent of diseases and conditions. The incidence of PIM use (at least one PIM over a 1-year period) by elderly patients was determined using health insurance claims data. Health care utilization and costs associated with PIM use were also estimated. Our findings will be used to increase awareness of drug-related problems relevant to PIM use and to improve geriatric pharmaceutical care provided by physicians and pharmacists for the aging Japanese population.
METHODS Study Design and Data Source This was a retrospective, observational cohort study using medical and pharmacy claims between April 2006 and September 2007. Claims were obtained from the database of Japan Medical Data Center (JMDC) Ltd. (Tokyo, Japan), which includes data from 5 corporate health insurance societies with 330,000 beneficiaries
147
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
(insured persons and their family members). The JMDC database consists of monthly claims from medical institutions and pharmacies and includes information about beneficiaries (age and sex), institutions (number of beds and specialties), diagnoses (International Statistical Classification of Diseases and Related Health Problems, 10th Revision [ICD-10] codes27), inpatients (admission date and length of stay), outpatients (number of visits), medications dispensed (generic or brand name, dose and administration, dispensing date), and service points. The claims data were obtained after removing all personal-information identifiers to ensure confidentiality. The study was approved by the institutional review board at the National Institute of Public Health (Saitama, Japan) and was in compliance with the Japanese Ethical Guidelines for Epidemiological Research updated in November 2007.28
Study Population Elderly patients aged ≥65 years who had at least 2 pharmacy claims in separate months over a 1-year period (April 2006 through March 2007) or during fiscal year (FY) 2006 were selected as our study participants. PIM users were defined as patients who were taking at least one medication on the modified list during this time period, and nonusers were defined as patients who were not. The index month was assigned as the first month that any of the PIMs were initiated for PIM users, or as the first month of any pharmacy claim for nonusers. Baseline characteristics, including demographics and health care utilization patterns, were measured from claims in the index month, and outcomes were measured from claims in the 6 months after the index month (eg, if the index month was May 2006, the outcomes were measured using claims between June 2006 and November 2006). Therefore, the total study period was April 2006 to September 2007.
Measures The main study outcomes were the rates of elderly patients who used at least one PIM over a 1-year period, as well as health care utilization rates and costs associated with PIM use. Medications or medication classes listed in the modified criteria were identified, and the likelihood of using any of those medications was examined from 1 year of pharmacy claims. For health care utilization, the likelihood of having an inpatient event, days of inpatient stay, and number of outpatient visits during 6 months after the index month were evaluated. Total medical costs included all service points taken from both medical and pharmacy claims and expressed
148
in US dollars ($1 ≈ ¥100 in June 2009). All costs were expressed in 2006 values. Explanatory variables included age category (70–74, 75–79, 80–84, 85–89, and ≥90 years), female sex, comorbidities, health care utilization, and prescribers’ institutions and their specialties in the index month. Comorbidity risk adjustments were rarely carried out in research conducted in Japan. In this study, comorbid conditions (n = 30) were identified from ICD-10 diagnosis codes using the Elixhauser classification system.29 This system was developed to predict health care expenditures and mortality; in addition, Quan et al30 established the cross-national validity of using ICD-10 codes. Health care utilization patterns were evaluated in the index month by inpatient events, number of medical institutions used, and number of unique prescription medications. Prescribers’ institutions were categorized by bed number (clinic, if 0–19 beds; hospital, if ≥20 beds, according to the Japanese Medical Care Act31), and prescribers were identified by specialty (general medicine, psychiatrist/neurologist, or other). Many prescribers belonging to hospitals reported their specialty as unknown or miscellaneous, so these were categorized as other. To adjust for seasonality of outcome measures, we included dummy variables indicating the index month.
Analysis Patients’ demographic and utilization characteristics in the index month, and outcome measures in the 6 months after the index month, were analyzed descriptively. The number of elderly patients with PIM use according to the modified Beers criteria was counted, and 1-year incidence rates for the total study population and for subgroups stratified by age (65– 74 years, ≥75 years) and sex were calculated. In addition, 1-year incidence rates were calculated for the 10 most frequently used PIMs. A logistic regression was used to examine demographic and clinical factors associated with PIMs, and the results were reported as odds ratios (ORs) and 95% CIs. Generalized linear models (GLMs) were then used to predict outcomes associated with PIM use. To predict total medical costs, likelihood of a hospital admission, and outpatient days, we used a log-link function with a γ distribution, a logit-link function with a binomial distribution, and a log-link function with a negative binomial distribution, respectively.32 GLM coefficient estimates were exponentiated to represent the ratio of expected costs, OR, and incidence rate ratio in PIM users compared with nonusers. All models in-
M. Akazawa et al.
cluded dummy variables indicating age category, female sex, hospitalization, polypharmacy (ie, having ≥5 unique prescriptions simultaneously), index month, and number of Elixhauser comorbidities to adjust for potential confounders. Statistical significance was expressed at the 5% level (P < 0.05). All statistical analyses were performed using SAS version 9.1.3 (SAS Institute Inc., Cary, North Carolina).
RESULTS A total of 7259 elderly patients aged ≥65 years were identified in the JMDC database during FY2006. Of these, 6628 patients (91.3%) had at least 2 pharmacy claims in separate months, and they were selected as the study population. According to the Japanese modification of the Beers criteria, 2889 patients (43.6%) had at least one PIM during FY2006. Demographic and clinical characteristics observed in the index month are summarized in Table I. In the total group, 71.2% (4721/6628) of the patients were female and 62.9% (4167/6628) were aged 65 to 74 years. No significant difference was observed in sex distribution between PIM users and nonusers. However, the number of Elixhauser comorbidities and health care utilization patterns differed between the groups. Patients who had at least one PIM were significantly more likely to have multiple comorbidities (eg, ≥4 comorbidities: 20.2% vs 7.3%), to be hospitalized (mean, 9% vs 2%), to receive services from many institutions (1.58 vs 1.27), and to receive a large number of unique medications in the index month (9.78 vs 4.75) (all, P < 0.001). PIM users were also more likely than nonusers to receive prescriptions from doctors at hospitals or from practitioners whose specialties were general medicine or psychiatrist/neurologist (all, P < 0.001). Because of the definition of the index month, nonusers tended to have earlier index months than did PIM users. Table II lists the 10 most frequently used PIMs among elderly Japanese patients. The most common PIMs were histamine-2 (H2) blockers, for which 20.5% (1356/6628) of the elderly received at least one prescription in FY2006, followed by benzodiazepines (11.4% [756/6628]) and anticholinergics and antihistamines (7.9% [526/6628]). Medications added to the modified criteria, including famotidine (12.0%), ranitidine (2.6%), sulpiride (2.5%), flunitrazepam (2.0%), and ultra– long-acting benzodiazepines (1.8%), were also prescribed frequently. No elderly patients were prescribed amphetamines, vesnarinone, isoxsuprine, or methyltestosterone, even though they were included in the modified Beers criteria.
The American Journal of Geriatric Pharmacotherapy
One-year incidence rates based on the 2003 Beers criteria10 and the Japanese modified criteria26 are described and compared in the figure. In the total study population, 2889 of the 6628 elderly patients (43.6%) were prescribed at least one PIM based on the Japanese modified Beers criteria. Excluding the medications newly added to the criteria (ie, according to the 2003 Beers criteria), 28.5% (1890/6628) of elderly patients used at least one PIM. No significant differences were observed in the incidence rates based on age or sex. The results of logistic regression analysis to model the likelihood of PIM use are presented in Table III. Factors positively associated with PIM prescriptions at a significance level of 5% included the following: hospital admission (OR = 3.35); polypharmacy (OR = 5.69); prescriptions from a hospital (OR = 1.19), general medicine practitioner (OR = 1.46), or psychiatrist/ neurologist (OR = 2.33); and comorbid conditions including peptic ulcer disease without bleeding (OR = 4.18), depression (OR = 3.69), cardiac arrhythmias (OR = 1.93), other neurologic disorders (Parkinson’s disease, multiple sclerosis, and epilepsy; OR = 1.88), and congestive heart failure (OR = 1.46). Age categories and sex had no association with PIM use. Health care utilization and costs observed in the 6 months after the index month are summarized descriptively in Table IV. The mean total costs in 6 months, including medical and pharmacy services, were $2695 for PIM users and $1277 for nonusers. These cost differences were explained by health care utilization patterns. Individuals with PIM use in the index month were more likely to have had an inpatient event (8% vs 3%; P < 0.001), had longer inpatient stays (21.85 vs 13.83 days; P = 0.004), and made more outpatient visits (16.08 vs 11.19 days; P < 0.001). Some individuals (1.6% [105/6628]) had no medical costs in the 6 months after the index month, but the proportions of these individuals were relatively small (<5%) and similar between the 2 groups (1.8% [51/ 2889] for PIM users vs 1.4% [54/3739] for nonusers). Therefore, they were excluded from further analyses using GLM regressions. The results of GLM regression analyses are summarized in Table V. All models included dummy variables indicating PIM use, age category, female sex, hospitalization, polypharmacy, index month, and the number of Elixhauser comorbidities as potential confounders for predicting outcomes in the 6 months after the index month. PIM use was associated with 33% higher total medical costs, 1.68-fold higher hospitalization risk, and 1.18-fold more outpatient visit days
149
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Table I. Demographic and clinical characteristics in the index month among elderly Japanese patients who were potentially inappropriate medication (PIM) users versus nonusers.* Data are number (%), unless otherwise indicated. $IBSBDUFSJTUJD
"MM1BUJFOUT /
1*.6TFST O
/POVTFST O
P †
"HFDBUFHPSZ Z o o o o o ȑ
4FY GFNBMF
†
&MJYIBVTFSDPNPSCJEJUJFT‡ ȑ
)FBMUITFSWJDFVUJMJ[BUJPO NFBO 4%
)PTQJUBMJ[BUJPOSJTL /PPGJOTUJUVUJPOTVTFE /PPGVOJRVFESVHT
§ § §
1SFTDSJCFSTJOTUJUVUJPOIPTQJUBM NFBO 4%
§
1SFTDSJCFSTTQFDJBMUZ NFBO 4%
(FOFSBMNFEJDJOF 1TZDIJBUSJTUOFVSPMPHJTU
§ §
*OEFYNPOUI< "QSJM .BZ +VOF +VMZ "VHVTU 4FQUFNCFS 0DUPCFS /PWFNCFS %FDFNCFS +BOVBSZ 'FCSVBSZ .BSDI
†
†
* $PMVNOTNBZOPUUPUBMEVFUPSPVOEJOH † χ2 Test. ‡ Comorbidities were identified from diagnoses on claims using definitions originally developed by Elixhauser et al29 and modified for International Statistical Classification of Diseases and Related Health Problems, 10th Revision diagnosis codes developed by Quan et al.30 § t5FTU < Index monthXBTEFGJOFEBTUIFGJSTUNPOUIXIFOBOZPGUIF1*.TXFSFQSFTDSJCFEJOGJTDBMZFBS ': "QSJMUISPVHI.BSDI BNPOH1*.VTFST PSUIFGJSTUNPOUIXJUIBOZQSFTDSJQUJPOJO':BNPOHOPOVTFST
150
M. Akazawa et al.
The American Journal of Geriatric Pharmacotherapy
Table II. The 10 most common potentially inappropriate medications used among elderly Japanese patients. /PPG1BUJFOUT /
:FBS*ODJEFODF ':
"OZQPUFOUJBMMZJOBQQSPQSJBUFNFEJDBUJPO
HCMPDLFST $JNFUJEJOF 3BOJUJEJOF† 'BNPUJEJOF† 3PYBUJEJOF† /J[BUJEJOF† -BGVUJEJOF†
#FO[PEJB[FQJOFT 'MVOJUSB[FQBN† 4IPSUBDUJOHCFO[PEJB[FQJOFT -POHBDUJOHCFO[PEJB[FQJOFT 6MUSBoMPOHBDUJOHCFO[PEJB[FQJOFT†
"OUJDIPMJOFSHJDTBOEBOUJIJTUBNJOFT $IMPSQIFOJSBNJOF )ZESPYZ[JOF
%PYB[PTJO
5JDMPQJEJOF
4VMQJSJEF†
(BTUSPJOUFTUJOBMBOUJTQBTNPEJDT
1FOUB[PDJOF
4IPSUBDUJOHEJQZSJEBNPMF
*OEPNFUIBDJO
.FEJDBUJPOBOE$MBTT
': GJTDBMZFBS ) IJTUBNJOF * 'PS)CMPDLFSTBOECFO[PEJB[FQJOFT TPNFQBUJFOUTVTFENVMUJQMFNFEJDBUJPOT JOUIFTBNFESVHDMBTTFT'PS BOUJDIPMJOFSHJDT POMZNFEJDBUJPOTXJUINBOZVTFST XFSFMJTUFE † .FEJDBUJPOPSDMBTTVOJRVFGPSUIFNPEJGJFEDSJUFSJBJO+BQBO
while controlling for all other explanatory variables. All observed differences were statistically significant at the 5% level.
DISCUSSION Using the country-specific modification of the Beers criteria, we found that 43.6% of elderly Japanese patients were at risk of PIM use. In this study, PIM use, along with other potential problems such as inpatient care or polypharmacy, was positively associated with hospitalizations, outpatient visits, and medical costs. Thus,
a drug utilization review to identify potentially inappropriate pharmaceutical care among the elderly would be one possible approach to reduce adverse outcomes and control growing medical expenditures. Direct comparisons to other studies are difficult because the modified criteria excluded medications not approved or rarely used in Japan and added some regional medications to the criteria; other discrepancies include different countries, time periods, and settings. High incidence rates of PIM use may have resulted from the common use of H2 blockers and benzodiaze-
151
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
2003 Beers criteria Japanese modification
One-Year Incidence of PIM Use (%)
50
40
30
20
10
0 Total population (N = 6628)
65—74 y (n = 4167)
≥75 y (n = 2461)
Male (n = 1907)
Female (n = 4721)
Figure. One-year incidence (fiscal year 2006) of potentially inappropriate medication (PIM) use among elderly Japanese patients based on the 2003 updated Beers criteria10 and the Japanese modification of the Beers criteria26 for the total study population and subpopulations stratified by age and sex.
Table III. Factors associated with potentially inappropriate medication use among elderly Japanese patients: Logistic regression analysis. 7BSJBCMF
0EET3BUJP
8BME$*
"HFDBUFHPSZ Z o o o o o ȑ
3FGFSFODF
o o o o o
4FY GFNBMF
o
6UJMJ[BUJPOJOJOEFYNPOUI )PTQJUBMJ[BUJPOSJTL 1PMZQIBSNBDZ ȑVOJRVFQSFTDSJQUJPOT
† †
o o
1SFTDSJCFSTJOTUJUVUJPOIPTQJUBM
†
o
1SFTDSJCFSTTQFDJBMUZ (FOFSBMNFEJDJOF 1TZDIJBUSJTUOFVSPMPHJTU
† †
o o
&MJYIBVTFSDPNPSCJEJUJFT 1FQUJDVMDFSEJTFBTFXJUIPVUCMFFEJOH %FQSFTTJPO
† †
o o DPOUJOVFE
152
M. Akazawa et al.
The American Journal of Geriatric Pharmacotherapy
Table III (continued). 7BSJBCMF &MJYIBVTFSDPNPSCJEJUJFT DPOUJOVFE
$BSEJBDBSSIZUINJBT 0UIFSOFVSPMPHJDEJTPSEFST‡ 1TZDIPTFT 8FJHIUMPTT $POHFTUJWFIFBSUGBJMVSF 3IFVNBUPJEBSUISJUJTDPMMBHFOWBTDVMBSEJTFBTFT -JWFSEJTFBTF 4PMJEUVNPSXJUIPVUNFUBTUBTJT 'MVJEBOEFMFDUSPMZUFEJTPSEFST )ZQFSUFOTJPO 7BMWVMBSEJTFBTF 1FSJQIFSBMWBTDVMBSEJTPSEFST #MPPEMPTTPSEFGJDJFODZBOFNJB§ )ZQPUIZSPJEJTN %JBCFUFT .FUBTUBUJDDBODFS $PBHVMPQBUIZ $ISPOJDQVMNPOBSZEJTFBTF 0CFTJUZ 3FOBMGBJMVSF "MDPIPMPSESVHBCVTF§ -ZNQIPNB 1BSBMZTJT 1VMNPOBSZDJSDVMBUJPOEJTPSEFST
0EET3BUJP
8BME$*
† † † † †
o o o o o o o o o o o o o o o o o o o o o o o o
* "MMWBSJBCMFTXFSFJEFOUJGJFEVTJOHDMBJNTEBUBJOUIFJOEFYNPOUI †4UBUJTUJDBMMZTJHOJGJDBOUBUP 8BMEχ2 test. ‡ Other neurologic disorders included Parkinson’s disease, multiple sclerosis, and epilepsy. § Patients with blood loss were classified with deficiency anemia and those with drug abuse were classified with alcohol abuse because of small sample sizes (<10). No patients had a diagnosis of HIV/AIDS.
pines; >10% of elderly patients were prescribed each of these medication classes in Japan. Moreover, there are some medications not listed in the 2003 Beers criteria that may have potential harms similar to the medications on the Japanese modified list. The modified criteria included additional H2 blockers, such as famotidine and ranitidine, as well as flunitrazepam and sulpiride, which are prescribed frequently for elderly patients, as shown in this study. Ulcer prevalence differs by country; research shows that duodenal ulcers predominate in Western countries, whereas gastric ulcers are more frequent in Asia, especially in Japan.33 H2 blockers are often used as preventive medications and are prescribed with NSAIDs and ste-
roids to avoid potential adverse effects.34 Therefore, prescription rates of H2 blockers (including newly added famotidine and ranitidine) are extremely high in Japan, and special attention is needed to avoid potential adverse effects, especially in the elderly population. Flunitrazepam was added to the modified list because it has an extremely long half-life in elderly patients (~20 hours),35 producing prolonged sedation and increasing the risk of falls and fracture.36 Sulpiride was added because it may cause extrapyramidal symptoms,37 and better tolerated alternatives are available to treat mild depression. Studies conducted in Europe and Taiwan have indicated potential limitations of drug utilization reviews using the Beers criteria.12,15 Some commonly used medi-
153
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Table IV. Outcome measures during 6 months after the index month among elderly Japanese patients who were potentially inappropriate medication (PIM) users versus nonusers. Data are presented as mean (SD), except as indicated.* .FBTVSF
"MM1BUJFOUT /
1*.6TFST O
/POVTFST O
P
5PUBMNFEJDBMDPTUT 64
†
†
*OQBUJFOUEBZT JGBOZ
†
0VUQBUJFOUEBZT
†
‡
*OQBUJFOUFWFOU
/PNFEJDBMDPTUTJONPOUIT OP
* $PTUEBUBBSFFYQSFTTFEJOWBMVFT6UJMJ[BUJPOEBUBXFSFDPMMFDUFEGSPN.BZUISPVHI4FQUFNCFS † t5FTU 1*.VTFSTWFSTVTOPOVTFST ‡ χ2 Test, PIM users versus nonusers.
Table V. Results of generalized linear models (GLMs) for predicting health care utilization rates and costs over 6 months associated with potentially inappropriate medication (PIM) use (vs nonuse) among elderly Japanese patients.* 0VUDPNF
.FUIPE
1BSBNFUFS
&YQPOFOUJBUFE $PFGGJDJFOU†
5PUBMNFEJDBMDPTUT
-PHMJOLGVODUJPOXJUIγ distribution
Ratio of expected costs
1.33
Hospitalization risk
Logit-link function with binomial distribution
Odds ratio
1.68
Outpatient days
Log-link function with negative binomial distribution
Incidence rate ratio
1.18
"MM(-.SFHSFTTJPOBOBMZTFTXFSFDPOEVDUFEJOQBUJFOUTXIPIBEBOZNFEJDBMDPTUTJOUIFNPOUITBGUFSUIFJOEFYNPOUI O "MMNPEFMTJODMVEFE EVNNZWBSJBCMFTJOEJDBUJOH1*.VTF BHFDBUFHPSZ GFNBMFTFY IPTQJUBMJ[BUJPO QPMZQIBSNBDZ JOEFYNPOUI BOEOVNCFSPG&MJYIBVTFSDPNPSCJEJUJFTUPBEKVTU GPSQPUFOUJBMDPOGPVOEFST †&BDI P
cations, especially those developed outside of the United States, were not listed (eg, flunitrazepam and etofylline in Europe), and the PIM use identified using the criteria can underestimate the potential problems. Because available medications and disease characteristics may differ across countries, drug utilization reviews using country-specific criteria may be useful for increasing prescriber awareness of PIMs and optimizing geriatric pharmaceutical care. Thus, modified criteria that take into account specific clinical environments and drug formularies may be a more comprehensive and useful tool. This study also indicated that the availability of alternative medications may influence medication choices. In a study conducted in Japanese long-
154
term care facilities, ticlopidine was the most frequently prescribed PIM, with 6.3% (105/1669) of elderly patients being prescribed this medication in 2002.13 However, the present study found that only 2.7% of elderly patients used ticlopidine in 2006. Of course, the numbers may not be comparable in different clinical settings (ie, outpatient vs long-term care). Another possible explanation for this difference, however, could be that a better tolerated alternative, clopidogrel,38 became available in Japan in January 2006, about 8 years after being marketed in the United States. A similar problem related to inaccessibility was reported in Europe, where ticlopidine was recommended for use in the elderly because clopidogrel was more expensive.12
M. Akazawa et al.
Using country-specific criteria, our study quantified potentially negative outcomes associated with PIM use in the elderly by adopting cohort study designs often used in outcome studies. Fick et al16 reported that Medicare managed-care patients who were prescribed one of the PIMs from the Beers criteria had significantly higher costs (mean, $4472 vs $2065) and greater utilization of inpatient services (0.54 vs 0.19 days), outpatient services (1.79 vs 0.98 days), and emergency department services (0.34 vs 0.15 days) (all, P < 0.001). Another study using the 2000–2001 Medical Expenditure Panel Survey estimated that the incremental health care cost associated with PIM use was $749 per patient per year in the United States.17 Moreover, 2 outcome studies related to PIM use among the elderly in Taiwan have been reported.15,39 Lin et al15 conducted an observational cohort study to examine PIM use and its association with adverse outcomes using computerized claims data from a large medical center in Taiwan. They reported that, using the 2003 Beers criteria, the 1-month prevalence of PIM use was 23.7% (1359/5741) in elderly patients receiving ambulatory care, and the incidence of adverse outcomes (emergency visits, hospitalization, and death) was higher in PIM users than in PIM nonusers (25.1% vs 17.5% in 6 months; P < 0.001). Chen et al39 used a nationwide computerized claims database from the National Health Insurance system and reported that 14.7% (193,897/1,318,943) of elderly patients who visited the emergency department with prescriptions used at least one PIM annually. This medication use was associated with more ambulatory care visits (mean, 45.79 vs 33.47 visits per patient), emergency department visits (2.26 vs 1.52), and hospital admissions (2.09 vs 1.87) (all, P < 0.001). Both research groups15,39 emphasized the importance of drug utilization reviews or surveillance systems based on computerized databases, as well as updates of the Beers criteria that account for local medical practices. There were several limitations associated with this study. First, our study population was selected from beneficiaries covered by the employees’ health insurance system.40 Because most beneficiaries are working adults or their family members, the proportion of elderly patients aged ≥65 years is very low (7259 of 330,000, or 2%), and these elderly patients tend to be younger and healthier than the general elderly population. Unlike previous studies, ours did not find an association of age or sex with PIM use, and this unique finding might be due to our selection of the study population.12,13,15 Second, this study could not examine the cause–effect relationship between PIM use and
The American Journal of Geriatric Pharmacotherapy
the study outcomes (eg, health care utilization or cost). The insurance claims included a monthly summary of health care services provided by health care providers; thus, no date information was available for most services. Therefore, we had to define the index month in which the first prescription of the PIM occurred rather than an index date, and we measured outcomes observed in the following 6 months by assuming that the study population would be at risk of exposure during the follow-up period. Measurement errors in both exposure and outcome status could occur if the patient stopped taking a medication just after the index month or if an adverse drug outcome occurred during the index month but could not be observed thereafter. Finally, we could not record the reasons for prescribing specific medications to the elderly patients because unlike medical charts, the claims data did not provide medical information. Bias is also possible due to confounding by indication and unobserved confounders. Because initiation of PIM use should be associated with both medical conditions in the index month and utilization rates and costs during the following months, multivariable regression methods could not adjust for all potential confounding effects, and the observed estimates should be biased upward. In 2007, the Japanese government set a priority for a full transition to online medical insurance claims and the establishment of a nationwide database including the claims data.41 The Japanese Society for Pharmacoepidemiology expects to use the national database for postmarketing surveillance and future studies of pharmacoepidemiology and pharmacoeconomics.42 The present study would be one example of an active drug utilization review using the claims data in Japan. For future studies, we plan to develop pharmaceutical care intervention programs provided by clinical pharmacists to determine whether controlling PIM use can optimize utilization of health care services for the elderly in both primary care settings and long-term care facilities.
CONCLUSIONS This study found that 43.6% of elderly Japanese patients used at least one PIM over a 1-year period. PIM use was associated with increased health care utilization rates and costs.
ACKNOWLEDGMENTS This study was conducted using a research grant in fiscal year 2008 from the Institute for Health Economics and Policy, Tokyo, Japan. The authors have indicated
155
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
that they have no other conflicts of interest regarding the content of this article. The authors thank Christopher M. Blanchette (Lovelace Respiratory Research Institute, Albuquerque, New Mexico) for assistance in editing this manuscript.
REFERENCES 1. Ministry of Health, Labour and Welfare. Annual Health, Labour and Welfare Report 2007-2008. http://www. mhlw.go.jp/english/wp/wp-hw2/. Accessed October 16, 2009. 2. Bates DW, Spell N, Cullen DJ, et al, for the Adverse Drug Events Prevention Study Group. The costs of adverse drug events in hospitalized patients. JAMA. 1997; 277:307–311. 3. Mita Y, Akishita M, Tanaka K, et al. Improvement of inappropriate prescribing and adverse drug withdrawal events after admission to long-term care facilities. Geriatr Gerontol Int. 2004;4:146–150. 4. Simonson W, Feinberg JL. Medication-related problems in the elderly: Defining the issues and identifying solutions. Drugs Aging. 2005;22:559–569. 5. Hajjar ER, Cafiero AC, Hanlon JT. Polypharmacy in elderly patients. Am J Geriatr Pharmacother. 2007;5:345–351. 6. Zhang M, Holman CD, Price SD, et al. Comorbidity and repeat admission to hospital for adverse drug reactions in older adults: Retrospective cohort study. BMJ. 2009;338: a2752. 7. Avorn J, Shrank WH. Adverse drug reactions in elderly people: A substantial cause of preventable illness. BMJ. 2008;336:956–957. 8. Beers MH, Ouslander JG, Rollingher I, et al. Explicit criteria for determining inappropriate medication use in nursing home residents. UCLA Division of Geriatric Medicine. Arch Intern Med. 1991;151:1825–1832. 9. Beers MH. Explicit criteria for determining potentially inappropriate medication use by the elderly. An update. Arch Intern Med. 1997;157:1531–1536. 10. Fick DM, Cooper JW, Wade WE, et al. Updating the Beers criteria for potentially inappropriate medication use in older adults: Results of a US consensus panel of experts [published correction appears in Arch Intern Med. 2004;164:298]. Arch Intern Med. 2003;163:2716– 2724. 11. Aparasu RR, Mort JR. Inappropriate prescribing for the elderly: Beers criteria-based review. Ann Pharmacother. 2000;34:338–346. 12. Fialová D, Topinková E, Gambassi G, et al, for the AdHOC Project Research Group. Potentially inappropriate medication use among elderly home care patients in Europe. JAMA. 2005;293:1348–1358.
156
13. Niwata S, Yamada Y, Ikegami N. Prevalence of inappropriate medication using Beers criteria in Japanese longterm care facilities. BMC Geriatr. 2006;6:1. 14. Gallagher P, Barry P, O’Mahony D. Inappropriate prescribing in the elderly. J Clin Pharm Ther. 2007;32:113– 121. 15. Lin HY, Liao CC, Cheng SH, et al. Association of potentially inappropriate medication use with adverse outcomes in ambulatory elderly patients with chronic diseases: Experience in a Taiwanese medical setting. Drugs Aging. 2008;25:49–59. 16. Fick DM, Waller JL, Maclean JR, et al. Potentially inappropriate medication use in a Medicare managed care population: Association with higher costs and utilization. J Manag Care Pharm. 2001;7:407–413. 17. Fu AZ, Jiang JZ, Reeves JH, et al. Potentially inappropriate medication use and healthcare expenditures in the US community-dwelling elderly. Med Care. 2007;45:472– 476. 18. Page RL II, Ruscin JM. The risk of adverse drug events and hospital-related morbidity and mortality among older adults with potentially inappropriate medication use. Am J Geriatr Pharmacother. 2006;4:297–305. 19. Budnitz DS, Shehab N, Kegler SR, Richards CL. Medication use leading to emergency department visits for adverse drug events in older adults. Ann Intern Med. 2007;147:755–765. 20. Jano E, Aparasu RR. Healthcare outcomes associated with Beers’ criteria: A systematic review. Ann Pharmacother. 2007;41:438–447. 21. McLeod PJ, Huang AR, Tamblyn RM, Gayton DC. Defining inappropriate practices in prescribing for elderly people: A national consensus panel. CMAJ. 1997;156: 385–391. 22. Naugler CT, Brymer C, Stolee P, Arcese ZA. Development and validation of an improving prescribing in the elderly tool. Can J Clin Pharmacol. 2000;7:103–107. 23. Barry PJ, O’Keefe N, O’Connor KA, O’Mahony D. Inappropriate prescribing in the elderly: A comparison of the Beers criteria and the improved prescribing in the elderly tool (IPET) in acutely ill elderly hospitalized patients. J Clin Pharm Ther. 2006;31:617–626. 24. Gallagher P, O’Mahony D. STOPP (Screening Tool of Older Persons’ potentially inappropriate Prescriptions): Application to acutely ill elderly patients and comparison with Beers’ criteria. Age Ageing. 2008;37:673–679. 25. Ryan C, O’Mahony D, Kennedy J, et al. Potentially inappropriate prescribing in an Irish elderly population in primary care. Br J Clin Pharmacol. 2009;68:936–947. 26. Imai H, Beers MH, Fick DM, et al. Development of Japanese Beers Criteria for inappropriate medication use
M. Akazawa et al.
27.
28.
29.
30.
31.
32. 33. 34.
in elderly patients [in Japanese]. Japan Med Assoc J. 2008;137:84–91. World Health Organization. International Statistical Classification of Diseases and Related Health Problems, 10th Revision. Version for 2007. http://apps.who.int/ classifications/apps/icd/icd10online/. Accessed March 7, 2010. Ministry of Health, Labour and Welfare. Ethical guidelines for epidemiological research [in Japanese]. http:// www.niph.go.jp/wadai/ekigakurinri/rinrishishin_ 071101.htm#1. Accessed March 7, 2010. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36:8–27. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD10 administrative data. Med Care. 2005;43:1130–1139. Medical Care Act (Act No. 205 of 1948) [in Japanese]. http://www.houko.com/00/01/S23/205.HTM#s4. Accessed March 7, 2010. SAS/STAT 9.1 User’s Guide: The GLM Procedure (Book Excerpt). Cary, NC: SAS Institute Inc; 2008. Yuan Y, Padol IT, Hunt RH. Peptic ulcer disease today. Nat Clin Pract Gastroenterol Hepatol. 2006;3:80–89. Lazzaroni M, Bianchi Porro G. Prophylaxis and treatment of non-steroidal anti-inflammatory drug-induced upper gastrointestinal side-effects. Dig Liver Dis. 2001; 33(Suppl):S44–S58.
The American Journal of Geriatric Pharmacotherapy
35. Mattila MA, Larni HM. Flunitrazepam: A review of its pharmacological properties and therapeutic use. Drugs. 1980;20:353–374. 36. Wang PS, Bohn RL, Glynn RJ, et al. Hazardous benzodiazepine regimens in the elderly: Effects of half-life, dosage, and duration on risk of hip fracture. Am J Psychiatry. 2001;158:892–898. 37. Mauri MC, Bravin S, Bitetto A, et al. A risk-benefit assessment of sulpiride in the treatment of schizophrenia. Drug Saf. 1996;14:288–298. 38. Bennett CL, Connors JM, Carwile JM, et al. Thrombotic thrombocytopenic purpura associated with clopidogrel. N Engl J Med. 2000;342:1773–1777. 39. Chen YC, Hwang SJ, Lai HY, et al. Potentially inappropriate medication for emergency department visits by elderly patients in Taiwan. Pharmacoepidemiol Drug Saf. 2009;18:53–61. 40. Social Insurance Agency. Overview of the Social Insurance Systems. http://www.sia.go.jp/e/index.html. Accessed October 16, 2009. 41. IT Strategic Headquarters. Priority Policy Program 2007. http://www.kantei.go.jp/foreign/policy/it/index_ e.html. Accessed October 16, 2009. 42. Japanese Society for Pharmacoepidemiology. Provisional translation of the report of the working group on “Future Postmarketing Surveillance in Japan” (published in 2003). http://www.jspe.jp/mt-static/FileUpload/files/ pms_en.pdf. Accessed March 7, 2010.
(continued on next page)
Address correspondence to: Manabu Akazawa, PhD, MPH, Drug Management and Policy, Faculty of Pharmacy, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kakuma-machi, Kanazawa-City, Ishikawa, 920-1192, Japan. E-mail: [email protected]
157
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Appendix. Medications or medication classes listed in the 2003 Beers criteria and the modified criteria for Japanese elderly patients. (Medications excluded from the 2003 criteria are underlined, and those newly added to the Japanese modification are expressed in italics.) .PEJGJFE$SJUFSJBGPS +BQBOFTF&MEFSMZ1BUJFOUT
#FFST$SJUFSJB "OBMHFTJDBOUJJOGMBNNBUPSZ *OEPNFUIBDJO ,FUPSPMBD .FQFSJEJOF -POHUFSNVTFPGGVMMEPTF MPOHFSIBMGMJGF DZDMPPYZHFOBTF OPOTFMFDUJWF/4"*%TOBQSPYFO PYBQSP[JO BOEQJSPYJDBN 1FOUB[PDJOF 1SPQPYZQIFOFBOEDPNCJOBUJPOQSPEVDUT "OUJBOFNJD 'FSSPVTTVMGBUF NHE
*OEPNFUIBDJO
-POHFSIBMGMJGFDZDMPPYZHFOBTFOPOTFMFDUJWF/4"*%T OBQSPYFO PYBQSP[JO QJSPYJDBN
1FOUB[PDJOF
'FSSPVTTVMGBUF NHE
"OUJBSSIZUINJD "NJPEBSPOF
"NJPEBSPOF Digitoxin %JHPYJO TIPVMEOPUFYDFFENHEFYDFQUXIFOUSFBUJOH %JHPYJO NHE BUSJBMBSSIZUINJBT
%JTPQZSBNJEF %JTPQZSBNJEF Pilsicainide Verapamil Vesnarinone "OUJCBDUFSJBM /JUSPGVSBOUPJO "OUJDIPMJOFSHJD "OUJDIPMJOFSHJDTBOEBOUJIJTUBNJOFTEJQIFOIZESBNJOF DIMPSQIFOJSBNJOF IZESPYZ[JOF DZQSPIFQUBEJOF QSPNFUIB[JOF USJQFMFOOBNJOF BOEEFYDIMPSQIFOJSBNJOF (BTUSPJOUFTUJOBMBOUJTQBTNPEJDESVHTEJDZDMPNJOF IZPTDZBNJOF QSPQBOUIFMJOF CFMMBEPOOBBMLBMPJET BOE DMJEJOJVNDIMPSEJB[FQPYJEF "OUJDMPUUJOH 4IPSUBDUJOHEJQZSJEBNPMF EPOPUDPOTJEFSMPOHBDUJOH EJQZSJEBNPMF XIJDIIBTCFUUFSQSPQFSUJFTUIBOUIFTIPSU BDUJOHJOPMEFSBEVMUT FYDFQUGPSQBUJFOUTXJUIBSUJGJDJBM IFBSUWBMWFT 5JDMPQJEJOF "OUJEFQSFTTBOU "NJUSJQUZMJOF DIMPSEJB[FQPYJEFBNJUSJQUZMJOF BOE QFSQIFOB[JOFBNJUSJQUZMJOF %PYFQJO %BJMZGMVPYFUJOF
"OUJDIPMJOFSHJDTBOEBOUJIJTUBNJOFTEJQIFOIZESBNJOF DIMPSQIFOJSBNJOF IZESPYZ[JOF DZQSPIFQUBEJOF BOE QSPNFUIB[JOF (BTUSPJOUFTUJOBMBOUJTQBTNPEJDT EJDZDMPNJOF QSPQBOUIFMJOF UJNFQJEJVN NFUIZMTDPQPMBNJOF NFUIZMPDUBUSPQJOF
4IPSUBDUJOHEJQZSJEBNPMF
5JDMPQJEJOF "NJUSJQUZMJOF
Milnacipran DPOUJOVFE
158
M. Akazawa et al.
The American Journal of Geriatric Pharmacotherapy
Appendix (continued). .PEJGJFE$SJUFSJBGPS +BQBOFTF&MEFSMZ1BUJFOUT
#FFST$SJUFSJB "OUJFNFUJD 5SJNFUIPCFO[BNJEF "OUJIZQFSUFOTJWF $MPOJEJOF %PYB[PTJO (VBOBESFM (VBOFUIJEJOF .FUIZMEPQBBOENFUIZMEPQBIZESPDIMPSPUIJB[JEF 3FTFSQJOFBUEPTFT NHE 4IPSUBDUJOHOJGFEJQJOF
$MPOJEJOF %PYB[PTJO .FUIZMEPQB Prazosin Propranolol 3FTFSQJOF NHE 4IPSUBDUJOHOJGFEJQJOF
"OUJQTZDIPUJD "MMCBSCJUVSBUFT FYDFQUQIFOPCBSCJUBM FYDFQUXIFOVTFEUP DPOUSPMTFJ[VSFT
Amantadine #BSCJUVSBUFTFYDFQUGPSTFJ[VSFDPOUSPM Dihydroergotoxine Gabapentin Olanzapine
.FTPSJEB[JOF Selegiline Sulpiride 5IJPSJEB[JOF 0SQIFOBESJOF %JVSFUJD &UIBDSZOJDBDJE &SHPUNFTZMPJE &SHPUNFTZMPJETBOEDZDMBOEFMBUF HBOUBHPOJTU $JNFUJEJOF
$JNFUJEJOF H2 blockers: ranitidine, famotidine, roxatidine, BOE nizatidine
)ZQPHMZDFNJD $IMPSQSPQBNJEF
$IMPSQSPQBNJEF
)PSNPOBM %FTJDDBUFEUIZSPJE &TUSPHFOTPOMZ PSBM .FUIZMUFTUPTUFSPOF
5IZSPJE &TUSPHFO .FUIZMUFTUPTUFSPOF
-BYBUJWF -POHUFSNVTFPGTUJNVMBOUMBYBUJWFTCJTBDPEZM DBTDBSB TBHSBEB BOEDBTUPSPJMFYDFQUJOUIFQSFTFODFPGPQJBUF BOBMHFTJDVTF .JOFSBMPJM
4UJNVMBOUMBYBUJWFTCJTBDPEZM DBTDBSBTBHSBEB BOEcastor oil
DPOUJOVFE
159
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Appendix (continued). .PEJGJFE$SJUFSJBGPS +BQBOFTF&MEFSMZ1BUJFOUT
#FFST$SJUFSJB .VTDMFSFMBYBOU .VTDMFSFMBYBOUTBOEBOUJTQBTNPEJDTNFUIPDBSCBNPM DBSJTPQSPEPM DIMPS[PYB[POF NFUBYBMPOF DZDMPCFO[BQSJOF BOEPYZCVUZOJO EPOPUDPOTJEFSFYUFOEFESFMFBTF PYZCVUZOJO 4FEBUJWF 'MVSB[FQBN .FQSPCBNBUF 4IPSUBDUJOHCFO[PEJB[FQJOFT EPTFTHSFBUFSUIBO MPSB[FQBNNHE BMQSB[PMBNNHE USJB[PMBNNHE PYB[FQBNNHE UFNB[FQBNNHE -POHBDUJOHCFO[PEJB[FQJOFTDIMPSEJB[FQPYJEF DIMPSEJB[FQPYJEFBNJUSJQUZMJOF DMJEJOJVN DIMPSEJB[FQPYJEF EJB[FQBN RVB[FQBN DMPSB[FQBUF BOE IBMB[FQBN
'MVSB[FQBN Flunitrazepam 4IPSUBDUJOHCFO[PEJB[FQJOFT EPTFTHSFBUFSUIBO MPSB[FQBNNHE BMQSB[PMBNNHE USJB[PMBNNHE BOE etizolam 3 mg/d -POHBDUJOHCFO[PEJB[FQJOFTDIMPSEJB[FQPYJEF EJB[FQBN RVB[FQBN BOEDMPSB[FQBUF
Ultra–long-acting benzodiazepines: loflazepate, flutoprazepam, mexazolam, haloxazolam, BOE cloxazolam 4UJNVMBOU "NQIFUBNJOFTBOEBOPSFYJDBHFOUT
"NQIFUBNJOFT FYDMVEJOHNFUIZMQIFOJEBUF
7BTPEJMBUPS $ZDMBOEFMBUF *TPYTVQSJOF
*TPYTVQSJOF
160
M. Akazawa et al.
Potentially Inappropriate Medication Use in Elderly Japanese Patients Manabu Akazawa, PhD, MPH1; Hirohisa Imai, MD, PhD2; Ataru Igarashi, PhD3; and Kiichiro Tsutani, MD, PhD3 1Drug
Management and Policy, Faculty of Pharmacy, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Ishikawa, Japan; 2Department of Epidemiology, National Institute of Public Health, Saitama, Japan; and 3Department of Drug Policy and Management, Graduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo, Japan
ABSTRACT Background: Modified Beers criteria for elderly Japanese patients were developed in 2008 by consensus among 9 experts to reflect regional clinical practice and available medications in Japan. Since then, many physicians and pharmacists have expressed interest in obtaining more information about the criteria and alternative drug choices. Objective: This study examined the incidence, health care utilization, and costs associated with potentially inappropriate medications (PIMs) in elderly patients based on the modified Beers criteria. Methods: A retrospective, observational cohort study was conducted using health insurance claims data in Japan. The study population included elderly patients aged ≥65 years who had at least 2 pharmacy claims in separate months over a 1-year period (April 2006 through March 2007). Use of the PIMs was identified using the modified criteria, and 1-year incidence rates were calculated for the total study population and for subgroups stratified by age and sex. A logistic regression model was used to examine demographic and clinical characteristics associated with PIMs. Health care utilization rates and costs were also analyzed and compared between patients with and without PIMs using generalized linear models. All models included dummy variables indicating age category, female sex, hospitalization, polypharmacy, index month, and number of Elixhauser comorbidities to adjust for potential confounders. Results: Among 6628 elderly patients, 71.2% (4721/6628) were female and 62.9% (4167/6628) were aged 65 to 74 years; 43.6% (2889/6628) were prescribed at least one PIM. The most commonly used PIMs were histamine-2 blockers (20.5% [1356/6628]), benzodiazepines (11.4% [756/6628]), and anticholinergics and antihistamines (7.9% [526/6628]). No significant differences in incidence rates were observed based on age or sex. Inpatient service use, polypharmacy, and comorbidities of peptic ulcer, depression, and cardiac arrhythmias were significant predictors of PIM use while controlling for other factors. PIM users had significantly higher hospitalization risk (1.68-fold), more outpatient visit days (1.18-fold), and higher medical costs (33% increase) than did nonusers. Conclusions: In a group of elderly Japanese patients, 43.6% used at least one PIM over a 1-year period in this study. PIM use was associated with greater health care utilization rates and costs. (Am J Geriatr Pharmacother. 2010;8:146–160) © 2010 Excerpta Medica Inc. Key words: Beers criteria, potentially inappropriate medication, incidence, utilization, cost, elderly, drug utilization review. 5IJTTUVEZXBTSFQPSUFEJOQBSUBTBQPTUFSQSFTFOUBUJPOBUUIFUI"OOVBM*OUFSOBUJPOBM.FFUJOHPGUIF*OUFSOBUJPOBM4PDJFUZGPS1IBSNBDPFDPOPNJDT BOE0VUDPNFT3FTFBSDI .BZo 0SMBOEP 'MPSJEB Accepted for publication February 23, 2010. ª&YDFSQUB.FEJDB*OD"MMSJHIUTSFTFSWFE
146
April 2010
7PMVNFt/VNCFS
EPJKBNKPQIBSN TFFGSPOUNBUUFS
M. Akazawa et al.
INTRODUCTION To adjust for increasing medical expenditures among the elderly in Japan, the government has initiated health care system reforms, including the development of a long-term care insurance for the elderly population.1 Elderly individuals are a heterogeneous population, and many of them have several chronic illnesses that require multiple medications, making them vulnerable to adverse drug events that contribute to elevated medical costs but are preventable.2–6 Appropriate pharmaceutical care for the elderly has been determined based on clinical trials conducted with younger adults. However, because of changes in pharmacokinetics and pharmacodynamics associated with aging, greater attention is needed in pharmacy management for patients aged >65 years.7 The Beers criteria are a consensus-based list of potentially inappropriate medications (PIMs) for individuals aged ≥65 years, first developed for nursing home residents in 1991 and updated twice in 1997 and 2003 for the general elderly population based on new clinical evidence.8–10 Medications or medication classes were selected for this list if potential risks outweighed potential benefits and if an effective alternative medication was available. The Beers criteria were widely used as tools for drug utilization reviews and outcome measures related to PIM use in geriatric health care in the United States, Europe, and Asian countries. These studies reported prevalence rates of PIM use ranging from 6% to 41%, depending on the study population and country.11–15 In some studies, the use of PIMs was associated with adverse outcomes, including increased health care utilization rates and costs16,17; in other settings, however, the results were inconclusive.18–20 These regional variations were explained by differences in medication availability, drug policy, and clinical practice. Some medications were not available or were rarely used in clinical settings. Thus, country-specific modifications of the criteria have been recommended to allow precise assessments of medication use and consequences. In addition to the Beers criteria, several other screening tools to evaluate PIMs have been published. The McLeod criteria were developed in Canada in 1997 using the Delphi consensus approach and included 38 medications explicitly contraindicated or likely to cause drug–drug or drug–disease interactions.21 This list was subsequently updated as the Improved Prescribing in the Elderly Tool (IPET) in 2000 and included 14 categories of contraindicated medications or possible drug–disease interactions.22 According to an
The American Journal of Geriatric Pharmacotherapy
observational study conducted in Ireland, IPET was a convenient tool for users but was not effective at identifying PIM use compared with the Beers criteria because it omitted too many well-recognized classes of drugs.23 Recently, a new tool, the Screening Tool of Older Persons’ Potentially Inappropriate Prescriptions (STOPP), was introduced and evaluated in Europe.24,25 STOPP was developed by a Delphi consensus method and contained 65 medications or medication classes. The tool was found to be more sensitive than the Beers criteria in identifying elderly patients with PIMs because several medications listed in the Beers criteria are either not prescribed or not available in most European countries.14 In 2008, Imai et al26 developed modified Beers criteria for the elderly Japanese population. Using the methods employed to develop the Beers criteria and their updates, 9 expert panel members selected 47 medications or medication classes that should be generally avoided for all elderly patients. Medications listed in the modified criteria and differences from the 2003 Beers criteria are shown in the appendix. (Because the original report of the Japanese Beers criteria was written in Japanese, interested readers may request a provisional translation of the method section from the authors.) Since publication of the Japanese Beers criteria, many physicians and pharmacists have expressed an interest in obtaining more information about the criteria and alternative drug choices. The objective of this study was to conduct a drug utilization review based on the modified Beers criteria in Japan. In this article, we focused on medications or medication classes that should generally be avoided independent of diseases and conditions. The incidence of PIM use (at least one PIM over a 1-year period) by elderly patients was determined using health insurance claims data. Health care utilization and costs associated with PIM use were also estimated. Our findings will be used to increase awareness of drug-related problems relevant to PIM use and to improve geriatric pharmaceutical care provided by physicians and pharmacists for the aging Japanese population.
METHODS Study Design and Data Source This was a retrospective, observational cohort study using medical and pharmacy claims between April 2006 and September 2007. Claims were obtained from the database of Japan Medical Data Center (JMDC) Ltd. (Tokyo, Japan), which includes data from 5 corporate health insurance societies with 330,000 beneficiaries
147
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
(insured persons and their family members). The JMDC database consists of monthly claims from medical institutions and pharmacies and includes information about beneficiaries (age and sex), institutions (number of beds and specialties), diagnoses (International Statistical Classification of Diseases and Related Health Problems, 10th Revision [ICD-10] codes27), inpatients (admission date and length of stay), outpatients (number of visits), medications dispensed (generic or brand name, dose and administration, dispensing date), and service points. The claims data were obtained after removing all personal-information identifiers to ensure confidentiality. The study was approved by the institutional review board at the National Institute of Public Health (Saitama, Japan) and was in compliance with the Japanese Ethical Guidelines for Epidemiological Research updated in November 2007.28
Study Population Elderly patients aged ≥65 years who had at least 2 pharmacy claims in separate months over a 1-year period (April 2006 through March 2007) or during fiscal year (FY) 2006 were selected as our study participants. PIM users were defined as patients who were taking at least one medication on the modified list during this time period, and nonusers were defined as patients who were not. The index month was assigned as the first month that any of the PIMs were initiated for PIM users, or as the first month of any pharmacy claim for nonusers. Baseline characteristics, including demographics and health care utilization patterns, were measured from claims in the index month, and outcomes were measured from claims in the 6 months after the index month (eg, if the index month was May 2006, the outcomes were measured using claims between June 2006 and November 2006). Therefore, the total study period was April 2006 to September 2007.
Measures The main study outcomes were the rates of elderly patients who used at least one PIM over a 1-year period, as well as health care utilization rates and costs associated with PIM use. Medications or medication classes listed in the modified criteria were identified, and the likelihood of using any of those medications was examined from 1 year of pharmacy claims. For health care utilization, the likelihood of having an inpatient event, days of inpatient stay, and number of outpatient visits during 6 months after the index month were evaluated. Total medical costs included all service points taken from both medical and pharmacy claims and expressed
148
in US dollars ($1 ≈ ¥100 in June 2009). All costs were expressed in 2006 values. Explanatory variables included age category (70–74, 75–79, 80–84, 85–89, and ≥90 years), female sex, comorbidities, health care utilization, and prescribers’ institutions and their specialties in the index month. Comorbidity risk adjustments were rarely carried out in research conducted in Japan. In this study, comorbid conditions (n = 30) were identified from ICD-10 diagnosis codes using the Elixhauser classification system.29 This system was developed to predict health care expenditures and mortality; in addition, Quan et al30 established the cross-national validity of using ICD-10 codes. Health care utilization patterns were evaluated in the index month by inpatient events, number of medical institutions used, and number of unique prescription medications. Prescribers’ institutions were categorized by bed number (clinic, if 0–19 beds; hospital, if ≥20 beds, according to the Japanese Medical Care Act31), and prescribers were identified by specialty (general medicine, psychiatrist/neurologist, or other). Many prescribers belonging to hospitals reported their specialty as unknown or miscellaneous, so these were categorized as other. To adjust for seasonality of outcome measures, we included dummy variables indicating the index month.
Analysis Patients’ demographic and utilization characteristics in the index month, and outcome measures in the 6 months after the index month, were analyzed descriptively. The number of elderly patients with PIM use according to the modified Beers criteria was counted, and 1-year incidence rates for the total study population and for subgroups stratified by age (65– 74 years, ≥75 years) and sex were calculated. In addition, 1-year incidence rates were calculated for the 10 most frequently used PIMs. A logistic regression was used to examine demographic and clinical factors associated with PIMs, and the results were reported as odds ratios (ORs) and 95% CIs. Generalized linear models (GLMs) were then used to predict outcomes associated with PIM use. To predict total medical costs, likelihood of a hospital admission, and outpatient days, we used a log-link function with a γ distribution, a logit-link function with a binomial distribution, and a log-link function with a negative binomial distribution, respectively.32 GLM coefficient estimates were exponentiated to represent the ratio of expected costs, OR, and incidence rate ratio in PIM users compared with nonusers. All models in-
M. Akazawa et al.
cluded dummy variables indicating age category, female sex, hospitalization, polypharmacy (ie, having ≥5 unique prescriptions simultaneously), index month, and number of Elixhauser comorbidities to adjust for potential confounders. Statistical significance was expressed at the 5% level (P < 0.05). All statistical analyses were performed using SAS version 9.1.3 (SAS Institute Inc., Cary, North Carolina).
RESULTS A total of 7259 elderly patients aged ≥65 years were identified in the JMDC database during FY2006. Of these, 6628 patients (91.3%) had at least 2 pharmacy claims in separate months, and they were selected as the study population. According to the Japanese modification of the Beers criteria, 2889 patients (43.6%) had at least one PIM during FY2006. Demographic and clinical characteristics observed in the index month are summarized in Table I. In the total group, 71.2% (4721/6628) of the patients were female and 62.9% (4167/6628) were aged 65 to 74 years. No significant difference was observed in sex distribution between PIM users and nonusers. However, the number of Elixhauser comorbidities and health care utilization patterns differed between the groups. Patients who had at least one PIM were significantly more likely to have multiple comorbidities (eg, ≥4 comorbidities: 20.2% vs 7.3%), to be hospitalized (mean, 9% vs 2%), to receive services from many institutions (1.58 vs 1.27), and to receive a large number of unique medications in the index month (9.78 vs 4.75) (all, P < 0.001). PIM users were also more likely than nonusers to receive prescriptions from doctors at hospitals or from practitioners whose specialties were general medicine or psychiatrist/neurologist (all, P < 0.001). Because of the definition of the index month, nonusers tended to have earlier index months than did PIM users. Table II lists the 10 most frequently used PIMs among elderly Japanese patients. The most common PIMs were histamine-2 (H2) blockers, for which 20.5% (1356/6628) of the elderly received at least one prescription in FY2006, followed by benzodiazepines (11.4% [756/6628]) and anticholinergics and antihistamines (7.9% [526/6628]). Medications added to the modified criteria, including famotidine (12.0%), ranitidine (2.6%), sulpiride (2.5%), flunitrazepam (2.0%), and ultra– long-acting benzodiazepines (1.8%), were also prescribed frequently. No elderly patients were prescribed amphetamines, vesnarinone, isoxsuprine, or methyltestosterone, even though they were included in the modified Beers criteria.
The American Journal of Geriatric Pharmacotherapy
One-year incidence rates based on the 2003 Beers criteria10 and the Japanese modified criteria26 are described and compared in the figure. In the total study population, 2889 of the 6628 elderly patients (43.6%) were prescribed at least one PIM based on the Japanese modified Beers criteria. Excluding the medications newly added to the criteria (ie, according to the 2003 Beers criteria), 28.5% (1890/6628) of elderly patients used at least one PIM. No significant differences were observed in the incidence rates based on age or sex. The results of logistic regression analysis to model the likelihood of PIM use are presented in Table III. Factors positively associated with PIM prescriptions at a significance level of 5% included the following: hospital admission (OR = 3.35); polypharmacy (OR = 5.69); prescriptions from a hospital (OR = 1.19), general medicine practitioner (OR = 1.46), or psychiatrist/ neurologist (OR = 2.33); and comorbid conditions including peptic ulcer disease without bleeding (OR = 4.18), depression (OR = 3.69), cardiac arrhythmias (OR = 1.93), other neurologic disorders (Parkinson’s disease, multiple sclerosis, and epilepsy; OR = 1.88), and congestive heart failure (OR = 1.46). Age categories and sex had no association with PIM use. Health care utilization and costs observed in the 6 months after the index month are summarized descriptively in Table IV. The mean total costs in 6 months, including medical and pharmacy services, were $2695 for PIM users and $1277 for nonusers. These cost differences were explained by health care utilization patterns. Individuals with PIM use in the index month were more likely to have had an inpatient event (8% vs 3%; P < 0.001), had longer inpatient stays (21.85 vs 13.83 days; P = 0.004), and made more outpatient visits (16.08 vs 11.19 days; P < 0.001). Some individuals (1.6% [105/6628]) had no medical costs in the 6 months after the index month, but the proportions of these individuals were relatively small (<5%) and similar between the 2 groups (1.8% [51/ 2889] for PIM users vs 1.4% [54/3739] for nonusers). Therefore, they were excluded from further analyses using GLM regressions. The results of GLM regression analyses are summarized in Table V. All models included dummy variables indicating PIM use, age category, female sex, hospitalization, polypharmacy, index month, and the number of Elixhauser comorbidities as potential confounders for predicting outcomes in the 6 months after the index month. PIM use was associated with 33% higher total medical costs, 1.68-fold higher hospitalization risk, and 1.18-fold more outpatient visit days
149
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Table I. Demographic and clinical characteristics in the index month among elderly Japanese patients who were potentially inappropriate medication (PIM) users versus nonusers.* Data are number (%), unless otherwise indicated. $IBSBDUFSJTUJD
"MM1BUJFOUT /
1*.6TFST O
/POVTFST O
P †
"HFDBUFHPSZ Z o o o o o ȑ
4FY GFNBMF
†
&MJYIBVTFSDPNPSCJEJUJFT‡ ȑ
)FBMUITFSWJDFVUJMJ[BUJPO NFBO 4%
)PTQJUBMJ[BUJPOSJTL /PPGJOTUJUVUJPOTVTFE /PPGVOJRVFESVHT
§ § §
1SFTDSJCFSTJOTUJUVUJPOIPTQJUBM NFBO 4%
§
1SFTDSJCFSTTQFDJBMUZ NFBO 4%
(FOFSBMNFEJDJOF 1TZDIJBUSJTUOFVSPMPHJTU
§ §
*OEFYNPOUI< "QSJM .BZ +VOF +VMZ "VHVTU 4FQUFNCFS 0DUPCFS /PWFNCFS %FDFNCFS +BOVBSZ 'FCSVBSZ .BSDI
†
†
* $PMVNOTNBZOPUUPUBMEVFUPSPVOEJOH † χ2 Test. ‡ Comorbidities were identified from diagnoses on claims using definitions originally developed by Elixhauser et al29 and modified for International Statistical Classification of Diseases and Related Health Problems, 10th Revision diagnosis codes developed by Quan et al.30 § t5FTU < Index monthXBTEFGJOFEBTUIFGJSTUNPOUIXIFOBOZPGUIF1*.TXFSFQSFTDSJCFEJOGJTDBMZFBS ': "QSJMUISPVHI.BSDI BNPOH1*.VTFST PSUIFGJSTUNPOUIXJUIBOZQSFTDSJQUJPOJO':BNPOHOPOVTFST
150
M. Akazawa et al.
The American Journal of Geriatric Pharmacotherapy
Table II. The 10 most common potentially inappropriate medications used among elderly Japanese patients. /PPG1BUJFOUT /
:FBS*ODJEFODF ':
"OZQPUFOUJBMMZJOBQQSPQSJBUFNFEJDBUJPO
HCMPDLFST $JNFUJEJOF 3BOJUJEJOF† 'BNPUJEJOF† 3PYBUJEJOF† /J[BUJEJOF† -BGVUJEJOF†
#FO[PEJB[FQJOFT 'MVOJUSB[FQBN† 4IPSUBDUJOHCFO[PEJB[FQJOFT -POHBDUJOHCFO[PEJB[FQJOFT 6MUSBoMPOHBDUJOHCFO[PEJB[FQJOFT†
"OUJDIPMJOFSHJDTBOEBOUJIJTUBNJOFT $IMPSQIFOJSBNJOF )ZESPYZ[JOF
%PYB[PTJO
5JDMPQJEJOF
4VMQJSJEF†
(BTUSPJOUFTUJOBMBOUJTQBTNPEJDT
1FOUB[PDJOF
4IPSUBDUJOHEJQZSJEBNPMF
*OEPNFUIBDJO
.FEJDBUJPOBOE$MBTT
': GJTDBMZFBS ) IJTUBNJOF * 'PS)CMPDLFSTBOECFO[PEJB[FQJOFT TPNFQBUJFOUTVTFENVMUJQMFNFEJDBUJPOT JOUIFTBNFESVHDMBTTFT'PS BOUJDIPMJOFSHJDT POMZNFEJDBUJPOTXJUINBOZVTFST XFSFMJTUFE † .FEJDBUJPOPSDMBTTVOJRVFGPSUIFNPEJGJFEDSJUFSJBJO+BQBO
while controlling for all other explanatory variables. All observed differences were statistically significant at the 5% level.
DISCUSSION Using the country-specific modification of the Beers criteria, we found that 43.6% of elderly Japanese patients were at risk of PIM use. In this study, PIM use, along with other potential problems such as inpatient care or polypharmacy, was positively associated with hospitalizations, outpatient visits, and medical costs. Thus,
a drug utilization review to identify potentially inappropriate pharmaceutical care among the elderly would be one possible approach to reduce adverse outcomes and control growing medical expenditures. Direct comparisons to other studies are difficult because the modified criteria excluded medications not approved or rarely used in Japan and added some regional medications to the criteria; other discrepancies include different countries, time periods, and settings. High incidence rates of PIM use may have resulted from the common use of H2 blockers and benzodiaze-
151
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
2003 Beers criteria Japanese modification
One-Year Incidence of PIM Use (%)
50
40
30
20
10
0 Total population (N = 6628)
65—74 y (n = 4167)
≥75 y (n = 2461)
Male (n = 1907)
Female (n = 4721)
Figure. One-year incidence (fiscal year 2006) of potentially inappropriate medication (PIM) use among elderly Japanese patients based on the 2003 updated Beers criteria10 and the Japanese modification of the Beers criteria26 for the total study population and subpopulations stratified by age and sex.
Table III. Factors associated with potentially inappropriate medication use among elderly Japanese patients: Logistic regression analysis. 7BSJBCMF
0EET3BUJP
8BME$*
"HFDBUFHPSZ Z o o o o o ȑ
3FGFSFODF
o o o o o
4FY GFNBMF
o
6UJMJ[BUJPOJOJOEFYNPOUI )PTQJUBMJ[BUJPOSJTL 1PMZQIBSNBDZ ȑVOJRVFQSFTDSJQUJPOT
† †
o o
1SFTDSJCFSTJOTUJUVUJPOIPTQJUBM
†
o
1SFTDSJCFSTTQFDJBMUZ (FOFSBMNFEJDJOF 1TZDIJBUSJTUOFVSPMPHJTU
† †
o o
&MJYIBVTFSDPNPSCJEJUJFT 1FQUJDVMDFSEJTFBTFXJUIPVUCMFFEJOH %FQSFTTJPO
† †
o o DPOUJOVFE
152
M. Akazawa et al.
The American Journal of Geriatric Pharmacotherapy
Table III (continued). 7BSJBCMF &MJYIBVTFSDPNPSCJEJUJFT DPOUJOVFE
$BSEJBDBSSIZUINJBT 0UIFSOFVSPMPHJDEJTPSEFST‡ 1TZDIPTFT 8FJHIUMPTT $POHFTUJWFIFBSUGBJMVSF 3IFVNBUPJEBSUISJUJTDPMMBHFOWBTDVMBSEJTFBTFT -JWFSEJTFBTF 4PMJEUVNPSXJUIPVUNFUBTUBTJT 'MVJEBOEFMFDUSPMZUFEJTPSEFST )ZQFSUFOTJPO 7BMWVMBSEJTFBTF 1FSJQIFSBMWBTDVMBSEJTPSEFST #MPPEMPTTPSEFGJDJFODZBOFNJB§ )ZQPUIZSPJEJTN %JBCFUFT .FUBTUBUJDDBODFS $PBHVMPQBUIZ $ISPOJDQVMNPOBSZEJTFBTF 0CFTJUZ 3FOBMGBJMVSF "MDPIPMPSESVHBCVTF§ -ZNQIPNB 1BSBMZTJT 1VMNPOBSZDJSDVMBUJPOEJTPSEFST
0EET3BUJP
8BME$*
† † † † †
o o o o o o o o o o o o o o o o o o o o o o o o
* "MMWBSJBCMFTXFSFJEFOUJGJFEVTJOHDMBJNTEBUBJOUIFJOEFYNPOUI †4UBUJTUJDBMMZTJHOJGJDBOUBUP 8BMEχ2 test. ‡ Other neurologic disorders included Parkinson’s disease, multiple sclerosis, and epilepsy. § Patients with blood loss were classified with deficiency anemia and those with drug abuse were classified with alcohol abuse because of small sample sizes (<10). No patients had a diagnosis of HIV/AIDS.
pines; >10% of elderly patients were prescribed each of these medication classes in Japan. Moreover, there are some medications not listed in the 2003 Beers criteria that may have potential harms similar to the medications on the Japanese modified list. The modified criteria included additional H2 blockers, such as famotidine and ranitidine, as well as flunitrazepam and sulpiride, which are prescribed frequently for elderly patients, as shown in this study. Ulcer prevalence differs by country; research shows that duodenal ulcers predominate in Western countries, whereas gastric ulcers are more frequent in Asia, especially in Japan.33 H2 blockers are often used as preventive medications and are prescribed with NSAIDs and ste-
roids to avoid potential adverse effects.34 Therefore, prescription rates of H2 blockers (including newly added famotidine and ranitidine) are extremely high in Japan, and special attention is needed to avoid potential adverse effects, especially in the elderly population. Flunitrazepam was added to the modified list because it has an extremely long half-life in elderly patients (~20 hours),35 producing prolonged sedation and increasing the risk of falls and fracture.36 Sulpiride was added because it may cause extrapyramidal symptoms,37 and better tolerated alternatives are available to treat mild depression. Studies conducted in Europe and Taiwan have indicated potential limitations of drug utilization reviews using the Beers criteria.12,15 Some commonly used medi-
153
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Table IV. Outcome measures during 6 months after the index month among elderly Japanese patients who were potentially inappropriate medication (PIM) users versus nonusers. Data are presented as mean (SD), except as indicated.* .FBTVSF
"MM1BUJFOUT /
1*.6TFST O
/POVTFST O
P
5PUBMNFEJDBMDPTUT 64
†
†
*OQBUJFOUEBZT JGBOZ
†
0VUQBUJFOUEBZT
†
‡
*OQBUJFOUFWFOU
/PNFEJDBMDPTUTJONPOUIT OP
* $PTUEBUBBSFFYQSFTTFEJOWBMVFT6UJMJ[BUJPOEBUBXFSFDPMMFDUFEGSPN.BZUISPVHI4FQUFNCFS † t5FTU 1*.VTFSTWFSTVTOPOVTFST ‡ χ2 Test, PIM users versus nonusers.
Table V. Results of generalized linear models (GLMs) for predicting health care utilization rates and costs over 6 months associated with potentially inappropriate medication (PIM) use (vs nonuse) among elderly Japanese patients.* 0VUDPNF
.FUIPE
1BSBNFUFS
&YQPOFOUJBUFE $PFGGJDJFOU†
5PUBMNFEJDBMDPTUT
-PHMJOLGVODUJPOXJUIγ distribution
Ratio of expected costs
1.33
Hospitalization risk
Logit-link function with binomial distribution
Odds ratio
1.68
Outpatient days
Log-link function with negative binomial distribution
Incidence rate ratio
1.18
"MM(-.SFHSFTTJPOBOBMZTFTXFSFDPOEVDUFEJOQBUJFOUTXIPIBEBOZNFEJDBMDPTUTJOUIFNPOUITBGUFSUIFJOEFYNPOUI O "MMNPEFMTJODMVEFE EVNNZWBSJBCMFTJOEJDBUJOH1*.VTF BHFDBUFHPSZ GFNBMFTFY IPTQJUBMJ[BUJPO QPMZQIBSNBDZ JOEFYNPOUI BOEOVNCFSPG&MJYIBVTFSDPNPSCJEJUJFTUPBEKVTU GPSQPUFOUJBMDPOGPVOEFST †&BDI P
cations, especially those developed outside of the United States, were not listed (eg, flunitrazepam and etofylline in Europe), and the PIM use identified using the criteria can underestimate the potential problems. Because available medications and disease characteristics may differ across countries, drug utilization reviews using country-specific criteria may be useful for increasing prescriber awareness of PIMs and optimizing geriatric pharmaceutical care. Thus, modified criteria that take into account specific clinical environments and drug formularies may be a more comprehensive and useful tool. This study also indicated that the availability of alternative medications may influence medication choices. In a study conducted in Japanese long-
154
term care facilities, ticlopidine was the most frequently prescribed PIM, with 6.3% (105/1669) of elderly patients being prescribed this medication in 2002.13 However, the present study found that only 2.7% of elderly patients used ticlopidine in 2006. Of course, the numbers may not be comparable in different clinical settings (ie, outpatient vs long-term care). Another possible explanation for this difference, however, could be that a better tolerated alternative, clopidogrel,38 became available in Japan in January 2006, about 8 years after being marketed in the United States. A similar problem related to inaccessibility was reported in Europe, where ticlopidine was recommended for use in the elderly because clopidogrel was more expensive.12
M. Akazawa et al.
Using country-specific criteria, our study quantified potentially negative outcomes associated with PIM use in the elderly by adopting cohort study designs often used in outcome studies. Fick et al16 reported that Medicare managed-care patients who were prescribed one of the PIMs from the Beers criteria had significantly higher costs (mean, $4472 vs $2065) and greater utilization of inpatient services (0.54 vs 0.19 days), outpatient services (1.79 vs 0.98 days), and emergency department services (0.34 vs 0.15 days) (all, P < 0.001). Another study using the 2000–2001 Medical Expenditure Panel Survey estimated that the incremental health care cost associated with PIM use was $749 per patient per year in the United States.17 Moreover, 2 outcome studies related to PIM use among the elderly in Taiwan have been reported.15,39 Lin et al15 conducted an observational cohort study to examine PIM use and its association with adverse outcomes using computerized claims data from a large medical center in Taiwan. They reported that, using the 2003 Beers criteria, the 1-month prevalence of PIM use was 23.7% (1359/5741) in elderly patients receiving ambulatory care, and the incidence of adverse outcomes (emergency visits, hospitalization, and death) was higher in PIM users than in PIM nonusers (25.1% vs 17.5% in 6 months; P < 0.001). Chen et al39 used a nationwide computerized claims database from the National Health Insurance system and reported that 14.7% (193,897/1,318,943) of elderly patients who visited the emergency department with prescriptions used at least one PIM annually. This medication use was associated with more ambulatory care visits (mean, 45.79 vs 33.47 visits per patient), emergency department visits (2.26 vs 1.52), and hospital admissions (2.09 vs 1.87) (all, P < 0.001). Both research groups15,39 emphasized the importance of drug utilization reviews or surveillance systems based on computerized databases, as well as updates of the Beers criteria that account for local medical practices. There were several limitations associated with this study. First, our study population was selected from beneficiaries covered by the employees’ health insurance system.40 Because most beneficiaries are working adults or their family members, the proportion of elderly patients aged ≥65 years is very low (7259 of 330,000, or 2%), and these elderly patients tend to be younger and healthier than the general elderly population. Unlike previous studies, ours did not find an association of age or sex with PIM use, and this unique finding might be due to our selection of the study population.12,13,15 Second, this study could not examine the cause–effect relationship between PIM use and
The American Journal of Geriatric Pharmacotherapy
the study outcomes (eg, health care utilization or cost). The insurance claims included a monthly summary of health care services provided by health care providers; thus, no date information was available for most services. Therefore, we had to define the index month in which the first prescription of the PIM occurred rather than an index date, and we measured outcomes observed in the following 6 months by assuming that the study population would be at risk of exposure during the follow-up period. Measurement errors in both exposure and outcome status could occur if the patient stopped taking a medication just after the index month or if an adverse drug outcome occurred during the index month but could not be observed thereafter. Finally, we could not record the reasons for prescribing specific medications to the elderly patients because unlike medical charts, the claims data did not provide medical information. Bias is also possible due to confounding by indication and unobserved confounders. Because initiation of PIM use should be associated with both medical conditions in the index month and utilization rates and costs during the following months, multivariable regression methods could not adjust for all potential confounding effects, and the observed estimates should be biased upward. In 2007, the Japanese government set a priority for a full transition to online medical insurance claims and the establishment of a nationwide database including the claims data.41 The Japanese Society for Pharmacoepidemiology expects to use the national database for postmarketing surveillance and future studies of pharmacoepidemiology and pharmacoeconomics.42 The present study would be one example of an active drug utilization review using the claims data in Japan. For future studies, we plan to develop pharmaceutical care intervention programs provided by clinical pharmacists to determine whether controlling PIM use can optimize utilization of health care services for the elderly in both primary care settings and long-term care facilities.
CONCLUSIONS This study found that 43.6% of elderly Japanese patients used at least one PIM over a 1-year period. PIM use was associated with increased health care utilization rates and costs.
ACKNOWLEDGMENTS This study was conducted using a research grant in fiscal year 2008 from the Institute for Health Economics and Policy, Tokyo, Japan. The authors have indicated
155
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
that they have no other conflicts of interest regarding the content of this article. The authors thank Christopher M. Blanchette (Lovelace Respiratory Research Institute, Albuquerque, New Mexico) for assistance in editing this manuscript.
REFERENCES 1. Ministry of Health, Labour and Welfare. Annual Health, Labour and Welfare Report 2007-2008. http://www. mhlw.go.jp/english/wp/wp-hw2/. Accessed October 16, 2009. 2. Bates DW, Spell N, Cullen DJ, et al, for the Adverse Drug Events Prevention Study Group. The costs of adverse drug events in hospitalized patients. JAMA. 1997; 277:307–311. 3. Mita Y, Akishita M, Tanaka K, et al. Improvement of inappropriate prescribing and adverse drug withdrawal events after admission to long-term care facilities. Geriatr Gerontol Int. 2004;4:146–150. 4. Simonson W, Feinberg JL. Medication-related problems in the elderly: Defining the issues and identifying solutions. Drugs Aging. 2005;22:559–569. 5. Hajjar ER, Cafiero AC, Hanlon JT. Polypharmacy in elderly patients. Am J Geriatr Pharmacother. 2007;5:345–351. 6. Zhang M, Holman CD, Price SD, et al. Comorbidity and repeat admission to hospital for adverse drug reactions in older adults: Retrospective cohort study. BMJ. 2009;338: a2752. 7. Avorn J, Shrank WH. Adverse drug reactions in elderly people: A substantial cause of preventable illness. BMJ. 2008;336:956–957. 8. Beers MH, Ouslander JG, Rollingher I, et al. Explicit criteria for determining inappropriate medication use in nursing home residents. UCLA Division of Geriatric Medicine. Arch Intern Med. 1991;151:1825–1832. 9. Beers MH. Explicit criteria for determining potentially inappropriate medication use by the elderly. An update. Arch Intern Med. 1997;157:1531–1536. 10. Fick DM, Cooper JW, Wade WE, et al. Updating the Beers criteria for potentially inappropriate medication use in older adults: Results of a US consensus panel of experts [published correction appears in Arch Intern Med. 2004;164:298]. Arch Intern Med. 2003;163:2716– 2724. 11. Aparasu RR, Mort JR. Inappropriate prescribing for the elderly: Beers criteria-based review. Ann Pharmacother. 2000;34:338–346. 12. Fialová D, Topinková E, Gambassi G, et al, for the AdHOC Project Research Group. Potentially inappropriate medication use among elderly home care patients in Europe. JAMA. 2005;293:1348–1358.
156
13. Niwata S, Yamada Y, Ikegami N. Prevalence of inappropriate medication using Beers criteria in Japanese longterm care facilities. BMC Geriatr. 2006;6:1. 14. Gallagher P, Barry P, O’Mahony D. Inappropriate prescribing in the elderly. J Clin Pharm Ther. 2007;32:113– 121. 15. Lin HY, Liao CC, Cheng SH, et al. Association of potentially inappropriate medication use with adverse outcomes in ambulatory elderly patients with chronic diseases: Experience in a Taiwanese medical setting. Drugs Aging. 2008;25:49–59. 16. Fick DM, Waller JL, Maclean JR, et al. Potentially inappropriate medication use in a Medicare managed care population: Association with higher costs and utilization. J Manag Care Pharm. 2001;7:407–413. 17. Fu AZ, Jiang JZ, Reeves JH, et al. Potentially inappropriate medication use and healthcare expenditures in the US community-dwelling elderly. Med Care. 2007;45:472– 476. 18. Page RL II, Ruscin JM. The risk of adverse drug events and hospital-related morbidity and mortality among older adults with potentially inappropriate medication use. Am J Geriatr Pharmacother. 2006;4:297–305. 19. Budnitz DS, Shehab N, Kegler SR, Richards CL. Medication use leading to emergency department visits for adverse drug events in older adults. Ann Intern Med. 2007;147:755–765. 20. Jano E, Aparasu RR. Healthcare outcomes associated with Beers’ criteria: A systematic review. Ann Pharmacother. 2007;41:438–447. 21. McLeod PJ, Huang AR, Tamblyn RM, Gayton DC. Defining inappropriate practices in prescribing for elderly people: A national consensus panel. CMAJ. 1997;156: 385–391. 22. Naugler CT, Brymer C, Stolee P, Arcese ZA. Development and validation of an improving prescribing in the elderly tool. Can J Clin Pharmacol. 2000;7:103–107. 23. Barry PJ, O’Keefe N, O’Connor KA, O’Mahony D. Inappropriate prescribing in the elderly: A comparison of the Beers criteria and the improved prescribing in the elderly tool (IPET) in acutely ill elderly hospitalized patients. J Clin Pharm Ther. 2006;31:617–626. 24. Gallagher P, O’Mahony D. STOPP (Screening Tool of Older Persons’ potentially inappropriate Prescriptions): Application to acutely ill elderly patients and comparison with Beers’ criteria. Age Ageing. 2008;37:673–679. 25. Ryan C, O’Mahony D, Kennedy J, et al. Potentially inappropriate prescribing in an Irish elderly population in primary care. Br J Clin Pharmacol. 2009;68:936–947. 26. Imai H, Beers MH, Fick DM, et al. Development of Japanese Beers Criteria for inappropriate medication use
M. Akazawa et al.
27.
28.
29.
30.
31.
32. 33. 34.
in elderly patients [in Japanese]. Japan Med Assoc J. 2008;137:84–91. World Health Organization. International Statistical Classification of Diseases and Related Health Problems, 10th Revision. Version for 2007. http://apps.who.int/ classifications/apps/icd/icd10online/. Accessed March 7, 2010. Ministry of Health, Labour and Welfare. Ethical guidelines for epidemiological research [in Japanese]. http:// www.niph.go.jp/wadai/ekigakurinri/rinrishishin_ 071101.htm#1. Accessed March 7, 2010. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36:8–27. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD10 administrative data. Med Care. 2005;43:1130–1139. Medical Care Act (Act No. 205 of 1948) [in Japanese]. http://www.houko.com/00/01/S23/205.HTM#s4. Accessed March 7, 2010. SAS/STAT 9.1 User’s Guide: The GLM Procedure (Book Excerpt). Cary, NC: SAS Institute Inc; 2008. Yuan Y, Padol IT, Hunt RH. Peptic ulcer disease today. Nat Clin Pract Gastroenterol Hepatol. 2006;3:80–89. Lazzaroni M, Bianchi Porro G. Prophylaxis and treatment of non-steroidal anti-inflammatory drug-induced upper gastrointestinal side-effects. Dig Liver Dis. 2001; 33(Suppl):S44–S58.
The American Journal of Geriatric Pharmacotherapy
35. Mattila MA, Larni HM. Flunitrazepam: A review of its pharmacological properties and therapeutic use. Drugs. 1980;20:353–374. 36. Wang PS, Bohn RL, Glynn RJ, et al. Hazardous benzodiazepine regimens in the elderly: Effects of half-life, dosage, and duration on risk of hip fracture. Am J Psychiatry. 2001;158:892–898. 37. Mauri MC, Bravin S, Bitetto A, et al. A risk-benefit assessment of sulpiride in the treatment of schizophrenia. Drug Saf. 1996;14:288–298. 38. Bennett CL, Connors JM, Carwile JM, et al. Thrombotic thrombocytopenic purpura associated with clopidogrel. N Engl J Med. 2000;342:1773–1777. 39. Chen YC, Hwang SJ, Lai HY, et al. Potentially inappropriate medication for emergency department visits by elderly patients in Taiwan. Pharmacoepidemiol Drug Saf. 2009;18:53–61. 40. Social Insurance Agency. Overview of the Social Insurance Systems. http://www.sia.go.jp/e/index.html. Accessed October 16, 2009. 41. IT Strategic Headquarters. Priority Policy Program 2007. http://www.kantei.go.jp/foreign/policy/it/index_ e.html. Accessed October 16, 2009. 42. Japanese Society for Pharmacoepidemiology. Provisional translation of the report of the working group on “Future Postmarketing Surveillance in Japan” (published in 2003). http://www.jspe.jp/mt-static/FileUpload/files/ pms_en.pdf. Accessed March 7, 2010.
(continued on next page)
Address correspondence to: Manabu Akazawa, PhD, MPH, Drug Management and Policy, Faculty of Pharmacy, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kakuma-machi, Kanazawa-City, Ishikawa, 920-1192, Japan. E-mail: [email protected]
157
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Appendix. Medications or medication classes listed in the 2003 Beers criteria and the modified criteria for Japanese elderly patients. (Medications excluded from the 2003 criteria are underlined, and those newly added to the Japanese modification are expressed in italics.) .PEJGJFE$SJUFSJBGPS +BQBOFTF&MEFSMZ1BUJFOUT
#FFST$SJUFSJB "OBMHFTJDBOUJJOGMBNNBUPSZ *OEPNFUIBDJO ,FUPSPMBD .FQFSJEJOF -POHUFSNVTFPGGVMMEPTF MPOHFSIBMGMJGF DZDMPPYZHFOBTF OPOTFMFDUJWF/4"*%TOBQSPYFO PYBQSP[JO BOEQJSPYJDBN 1FOUB[PDJOF 1SPQPYZQIFOFBOEDPNCJOBUJPOQSPEVDUT "OUJBOFNJD 'FSSPVTTVMGBUF NHE
*OEPNFUIBDJO
-POHFSIBMGMJGFDZDMPPYZHFOBTFOPOTFMFDUJWF/4"*%T OBQSPYFO PYBQSP[JO QJSPYJDBN
1FOUB[PDJOF
'FSSPVTTVMGBUF NHE
"OUJBSSIZUINJD "NJPEBSPOF
"NJPEBSPOF Digitoxin %JHPYJO TIPVMEOPUFYDFFENHEFYDFQUXIFOUSFBUJOH %JHPYJO NHE BUSJBMBSSIZUINJBT
%JTPQZSBNJEF %JTPQZSBNJEF Pilsicainide Verapamil Vesnarinone "OUJCBDUFSJBM /JUSPGVSBOUPJO "OUJDIPMJOFSHJD "OUJDIPMJOFSHJDTBOEBOUJIJTUBNJOFTEJQIFOIZESBNJOF DIMPSQIFOJSBNJOF IZESPYZ[JOF DZQSPIFQUBEJOF QSPNFUIB[JOF USJQFMFOOBNJOF BOEEFYDIMPSQIFOJSBNJOF (BTUSPJOUFTUJOBMBOUJTQBTNPEJDESVHTEJDZDMPNJOF IZPTDZBNJOF QSPQBOUIFMJOF CFMMBEPOOBBMLBMPJET BOE DMJEJOJVNDIMPSEJB[FQPYJEF "OUJDMPUUJOH 4IPSUBDUJOHEJQZSJEBNPMF EPOPUDPOTJEFSMPOHBDUJOH EJQZSJEBNPMF XIJDIIBTCFUUFSQSPQFSUJFTUIBOUIFTIPSU BDUJOHJOPMEFSBEVMUT FYDFQUGPSQBUJFOUTXJUIBSUJGJDJBM IFBSUWBMWFT 5JDMPQJEJOF "OUJEFQSFTTBOU "NJUSJQUZMJOF DIMPSEJB[FQPYJEFBNJUSJQUZMJOF BOE QFSQIFOB[JOFBNJUSJQUZMJOF %PYFQJO %BJMZGMVPYFUJOF
"OUJDIPMJOFSHJDTBOEBOUJIJTUBNJOFTEJQIFOIZESBNJOF DIMPSQIFOJSBNJOF IZESPYZ[JOF DZQSPIFQUBEJOF BOE QSPNFUIB[JOF (BTUSPJOUFTUJOBMBOUJTQBTNPEJDT EJDZDMPNJOF QSPQBOUIFMJOF UJNFQJEJVN NFUIZMTDPQPMBNJOF NFUIZMPDUBUSPQJOF
4IPSUBDUJOHEJQZSJEBNPMF
5JDMPQJEJOF "NJUSJQUZMJOF
Milnacipran DPOUJOVFE
158
M. Akazawa et al.
The American Journal of Geriatric Pharmacotherapy
Appendix (continued). .PEJGJFE$SJUFSJBGPS +BQBOFTF&MEFSMZ1BUJFOUT
#FFST$SJUFSJB "OUJFNFUJD 5SJNFUIPCFO[BNJEF "OUJIZQFSUFOTJWF $MPOJEJOF %PYB[PTJO (VBOBESFM (VBOFUIJEJOF .FUIZMEPQBBOENFUIZMEPQBIZESPDIMPSPUIJB[JEF 3FTFSQJOFBUEPTFT NHE 4IPSUBDUJOHOJGFEJQJOF
$MPOJEJOF %PYB[PTJO .FUIZMEPQB Prazosin Propranolol 3FTFSQJOF NHE 4IPSUBDUJOHOJGFEJQJOF
"OUJQTZDIPUJD "MMCBSCJUVSBUFT FYDFQUQIFOPCBSCJUBM FYDFQUXIFOVTFEUP DPOUSPMTFJ[VSFT
Amantadine #BSCJUVSBUFTFYDFQUGPSTFJ[VSFDPOUSPM Dihydroergotoxine Gabapentin Olanzapine
.FTPSJEB[JOF Selegiline Sulpiride 5IJPSJEB[JOF 0SQIFOBESJOF %JVSFUJD &UIBDSZOJDBDJE &SHPUNFTZMPJE &SHPUNFTZMPJETBOEDZDMBOEFMBUF HBOUBHPOJTU $JNFUJEJOF
$JNFUJEJOF H2 blockers: ranitidine, famotidine, roxatidine, BOE nizatidine
)ZQPHMZDFNJD $IMPSQSPQBNJEF
$IMPSQSPQBNJEF
)PSNPOBM %FTJDDBUFEUIZSPJE &TUSPHFOTPOMZ PSBM .FUIZMUFTUPTUFSPOF
5IZSPJE &TUSPHFO .FUIZMUFTUPTUFSPOF
-BYBUJWF -POHUFSNVTFPGTUJNVMBOUMBYBUJWFTCJTBDPEZM DBTDBSB TBHSBEB BOEDBTUPSPJMFYDFQUJOUIFQSFTFODFPGPQJBUF BOBMHFTJDVTF .JOFSBMPJM
4UJNVMBOUMBYBUJWFTCJTBDPEZM DBTDBSBTBHSBEB BOEcastor oil
DPOUJOVFE
159
The American Journal of Geriatric Pharmacotherapy
M. Akazawa et al.
Appendix (continued). .PEJGJFE$SJUFSJBGPS +BQBOFTF&MEFSMZ1BUJFOUT
#FFST$SJUFSJB .VTDMFSFMBYBOU .VTDMFSFMBYBOUTBOEBOUJTQBTNPEJDTNFUIPDBSCBNPM DBSJTPQSPEPM DIMPS[PYB[POF NFUBYBMPOF DZDMPCFO[BQSJOF BOEPYZCVUZOJO EPOPUDPOTJEFSFYUFOEFESFMFBTF PYZCVUZOJO 4FEBUJWF 'MVSB[FQBN .FQSPCBNBUF 4IPSUBDUJOHCFO[PEJB[FQJOFT EPTFTHSFBUFSUIBO MPSB[FQBNNHE BMQSB[PMBNNHE USJB[PMBNNHE PYB[FQBNNHE UFNB[FQBNNHE -POHBDUJOHCFO[PEJB[FQJOFTDIMPSEJB[FQPYJEF DIMPSEJB[FQPYJEFBNJUSJQUZMJOF DMJEJOJVN DIMPSEJB[FQPYJEF EJB[FQBN RVB[FQBN DMPSB[FQBUF BOE IBMB[FQBN
'MVSB[FQBN Flunitrazepam 4IPSUBDUJOHCFO[PEJB[FQJOFT EPTFTHSFBUFSUIBO MPSB[FQBNNHE BMQSB[PMBNNHE USJB[PMBNNHE BOE etizolam 3 mg/d -POHBDUJOHCFO[PEJB[FQJOFTDIMPSEJB[FQPYJEF EJB[FQBN RVB[FQBN BOEDMPSB[FQBUF
Ultra–long-acting benzodiazepines: loflazepate, flutoprazepam, mexazolam, haloxazolam, BOE cloxazolam 4UJNVMBOU "NQIFUBNJOFTBOEBOPSFYJDBHFOUT
"NQIFUBNJOFT FYDMVEJOHNFUIZMQIFOJEBUF
7BTPEJMBUPS $ZDMBOEFMBUF *TPYTVQSJOF
*TPYTVQSJOF
160