Cost-effectiveness of a NewOpportunistic Screening Strategyfor Walk-in Fingertip HbA1cTesting at CommunityPharmacies in JapanDiabetes Care 2018;41:1218–1226 | https://doi.org/10.2337/dc17-1307

OBJECTIVE

A new opportunistic community-based strategy was launched in Japan in April2014 to detect lifestyle-related diseases, including diabetes, by creating SpecimenMeasurement Offices (SMOs). SMOs offer walk-in fingertip HbA1c testing. Thisarticle aimed to assess the value-for-money of HbA1c testing services at SMOs byconducting a cost-effectiveness analysis.

RESEARCH DESIGN AND METHODS

We compared two scenarios: 1) status quo, defined as HbA1c testing that is availableonly through conventional screening, and 2) HbA1c testing available at SMOs as acomplement to the status quo scenario. The model consisted of a screening modulewith a decision tree and a disease progression module with a Markov model. Wecalculated incremental cost-effectiveness ratios (i.e., cost per quality-adjusted life-years [QALYs]) over the lifetime analytic horizon as the primary end point of thecost-effectiveness analysis. In this model, we assumed the participant cohort to bepeople 40–74 years of age who sought walk-in fingertip HbA1c testing at SMOs onthe premises of community pharmacies. Costs and outcomes were discounted at arate of 3%. The cost-effectiveness was analyzed from a societal perspective.

RESULTS

The incremental cost per individual for those 40–74 years of age was estimated tobe2527 U.S. dollars (USD) (252,722 Japanese yen [JPY]) for HbA1c testing at SMOscompared with the status quo. Incremental effectiveness was estimated to be0.0203 QALYs for HbA1c testing at SMOs compared with the status quo. Therefore,this cost-effectiveness analysis showed that compared with the status quo, HbA1ctesting at SMOs was more effective and had lower cost for the population studied.

CONCLUSIONS

We consider our results to be robust because most simulations were under thethreshold of USD 50,000 (JPY 5,000,000) per QALYs gained, by sensitivity analysis.These results will be useful to managers of pharmacies or other health institutionsand/or policy makers in local government.

1Department of Public Health and Epidemiology,Faculty of Pharmaceutical Sciences, Meiji Phar-maceutical University, Tokyo, Japan2Department of Health Care Policy and HealthEconomics, Faculty of Medicine, University ofTsukuba, Tsukuba, Japan3Department of Nephrology, University of Tsu-kuba, Tsukuba, Japan4Nutrigenomics Research Group, Faculty ofMedicine, University of Tsukuba, Tsukuba, Japan5Revolution of Access to Diabetes Diagnosis(RADD) Project, Tsukuba, Japan6The Cooperation Council for Specimen Mea-surement Offices, Tokyo, Japan

Corresponding author: Aiko Shono, shono@my-pharm.ac.jp.

Received 29 June 2017 and accepted 24 March2018.

This article contains Supplementary Data online athttp://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc17-1307/-/DC1.

© 2018 by the American Diabetes Association.Readers may use this article as long as the workis properly cited, the use is educational and notfor profit, and the work is not altered. More infor-mation is available at http://www.diabetesjournals.org/content/license.

Aiko Shono,1,2 Masahide Kondo,2 Shu-

ling Hoshi,2 Reiko Okubo,3 and

Naoya Yahagi4,5,6

1218 Diabetes Care Volume 41, June 2018

EPIDEM

IOLO

GY/HEA

LTHSERVICES

RESEA

RCH

Type 2 diabetes is often asymptomaticand can remain undiagnosed and un-treated formany years, until complicationsappear (1,2). Detection of prediabetes orearly-stage diabetes would provide anopportunity to intervene and delay thedevelopment of type 2 diabetes; how-ever, there are disputes about the directevidence of screening itself (3–6).There are several different approaches

to screening for type 2 diabetes, namely,universal, targeted, community-based,and opportunistic screening. These ap-proaches have been used alone and incombination (6). Nationwide universalscreening (Specific Health Checkup [SHC])includes type 2 diabetes screening inJapan (7) and the National Health ServiceHealth Check program in the U.K., a nation-wide targeted screening, which includesa diabetes check aswell (8). Opportunisticscreening is also common in many coun-tries (6). This involves testing for type 2diabetes that is conducted at the time ofscheduled clinical visits for the manage-ment of other diseases (8). Also, oppor-tunistic screening at community pharmacieshas been in place in some countries, suchas the U.S., Switzerland, the U.K., andAustralia (8,9).In Japan, undiagnosed cases of type 2

diabetes have been estimated to accountfor ;35% of the total 9.4 million casesamong adults in 2015 (10). Currently, theSHC represents one of three availableopportunities for type 2 diabetes screen-ing, in addition to conventional opportu-nistic screening in clinics and walk-infingertip HbA1c testing at a SpecimenMeasurement Office (SMO). The Japa-nese government launched the SHCand Specific Counseling Guidance in2008 as a nationwide mass screeningprogram to detect metabolic syndromeamong adults 40–74 years of age. Insur-ers are required to allow insured persons40–74 years of age to participate in theSHC program. During an SHC visit, HbA1cor fasting blood glucose level should bescreened to detect type 2 diabetes (11).However, although it is a nationwide pro-gram, the SHC coveragewas only;48.6%in 2014 (12).A new opportunistic community-based

strategy was launched in Japan to detectlifestyle-related diseases including diabe-tes, by creating the SMOs, which offerHbA1c testing, in April 2014. This legisla-tion was prompted by the success of atrial diabetes screening project, Revolution

of Access to Diabetes Diagnosis (RADD),which was implemented at 10 pharma-cies in Adachi Ward, Tokyo, as a jointprogram of the nonprofit Adachi Diabe-tes Mellitus Society and the Adachi WardPharmacist Association (13). On 28 Feb-ruary 2018, there were 1,586 SMOs inJapan, mostly established on the prem-ises of community pharmacies or drugstores (14). In Adachi Ward, the RADDevolved in 2015 into an official subsidizedprogram for HbA1c testing at SMOs, toincrease participation in diabetes screen-ing. In this locally subsidized program, only5 U.S. dollars (USD) (equivalent to 500Japanese yen [JPY]) per test is charged toresidents of Adachi Ward, compared withthe regular fee of USD 10 (JPY 1,000) atthe 10 RADD pharmacies (15).

HbA1c testing at the SMOs is mainlytargeted to people who miss or fail topresent at SHC visits. The screeningcomplements conventional health servicesusing an innovative approach. If opportu-nistic screening at SMOs becomes wide-spread in Japan, it could offer an importantresource for people who are not tested bythe SHC program. It is suggested that itcould also provide a new mode of carein community settings for diverse healthcare professionals seeking better manage-ment of people with diabetes. This articleaimed to assess the value-for-money ofHbA1c testing services at SMOs from asocietal perspective by conducting a cost-effectiveness analysis.

RESEARCH DESIGN AND METHODS

Weconducted a cost-effectiveness analysisto evaluate the introduction of HbA1c test-ing services at SMOs in Japan by adoptinga societal perspective. We compared twoscenarios: 1) status quo, defined as HbA1ctesting that is available only during SHCvisits and conventional opportunisticscreening in clinics, and 2) HbA1c testingthat is available at SMOs as a comple-ment to the status quo scenario.

In the model, the data on SMOs weretaken from the RADD Project (16). Datawere used from 1,296 people who were40–74 years of age from a total of 2,024people who provided HbA1c data to SMOsbetween October 2010 and May 2014.The total study population (n = 2,024)had a mean age of 52.1 years (age range,17–92 years), and 48% were male. Wealso carried out a literature search to findrelevant available evidence on type 2diabetes and its complications in Japan.

The PubMed database; Igaku Chuo Zasshi(JapanMedical Abstracts Society), a Japa-nese medical literature database; annualgovernment statistical reports; Ministryof Health, Labour and Welfare (MHLW)grant system; and local government reportsand tariffs were accessed using variouscombinations of relevant terms, such as“type 2 diabetes.” We then used thosesources to estimate the annual transitionprobabilities of complications as well ascosts.

Opportunistic HbA1c Screening atSMOsThe process modeled for HbA1c testing inpharmacy settings at SMOs is as follows.A customer at a pharmacy asks the phar-macist to have their HbA1c tested. Thepharmacist confirms that the customer isnot being treated for diabetes, to excludepatients who have received a diagnosis oftype 2 diabetes. After receiving an expla-nation about the HbA1c test from thepharmacist, the customer undergoes test-ing. The pharmacist gives the customertheir results together with informationabout how to interpret the results. Thecustomer is advised to visit their familydoctor if their HbA1c level is not ,6.0%(42 mmol/mol) (16).

Target PopulationIn thismodel, we assumed the participantcohort to be individuals 40–74 years ofage who sought walk-in fingertip HbA1ctesting. In addition, we assumed a hypo-thetical scenario of only one HbA1c test atan SMO for the age-groups 40–44, 45–49,50–54, 55–59, 60–64, 65–69, and 70–74years. At the initial stage of the model,every person is assumed to have no di-agnosis of or have not been treated fortype 2 diabetes. The model consisted of ascreening module with a decision tree(Fig. 1A) and a disease progression mod-ule with a Markov model (Fig. 1B).

Model AssumptionTwo scenarios were modeled as follows:1) status quo and 2) complementaryHbA1c testing at SMOs. The model wasa combination of a decision tree and aMarkov model. The decision tree includedtwo main nodes, as follows: 1) the deci-sion node, which represented the deci-sion made by pharmacy managers aboutwhether to establish an SMO that pro-vided HbA1c testing on the pharmacypremises, and 2) chance nodes, whichrepresented the possible events that could

care.diabetesjournals.org Shono and Associates 1219

occur (17). In Japan, 6% (42 mmol/mol)has been adopted by the National Healthand Nutrition Survey as the cutoff valuefor defining the risk for the developmentof type 2 diabetes. Therefore, in the currentstudy, a participant with an HbA1c testresult above this cutoff value was definedas a person at risk for diabetes. Additionaldetails are shown in the SupplementaryAppendix, and other probabilities as-signed to the chance nodes are shownin Supplementary Table 1.The following two kinds of Markov

models followed the decision tree: aMarkov positivemodel (Markov [+]model),which suggests the prognosis for a per-son with type 2 diabetes (Fig. 1B), and aMarkov negative model (Markov [2]

model, which suggests the future for aperson without type 2 diabetes (Fig. 1C).

The Markov (+) model comprised aseries of five health states that a patientcan have at a given point in time (17): 1)type 2 diabetes without complications, 2)end-stage renal disease (ESRD), 3) car-diovascular disease (CVD), 4) amputation,5) blindness, and 6) death. For type 2diabetes, each person was assumed tohave a single complication rather thanmultiple complications. Each transitionprobability defines the rate at which eachcomplication is reached with or withouttreatment as well as the probability ofdeath from no diabetes or diabetes with-out complications as well as diabeteswith a complication. We assumed that

the age-specific probability of death forthose without diabetes and those withdiabetes without complications or withblindness only were the same (18).

The Markov (2) model comprised thefollowing two states: 1) no type 2 di-abetes at a given point in time and 2)death. Annual transition probabilities(death with no type 2 diabetes) were de-termined using vital statistics (19). Theeffects of early intervention for type 2diabetes were included in the Markov (+)model. One Markov cycle was set at1 year. Individuals who did not undergoHbA1c testing at SMOs were omitted fromthis model because they would haveproduced no effects in terms of costand effectiveness in either scenario.

Figure 1—A: Model: decision tree. B: Model: Markov model with type 2 diabetes state. C: Model: Markov model with no type 2 diabetes.

1220 Cost-effectiveness of Walk-in HbA1c Screening Diabetes Care Volume 41, June 2018

UtilitiesUtility weights used to calculate quality-adjusted life-years (QALYs) were 1.0for individuals without type 2 diabetes,0.910 for individuals with early-stagetype 2 diabetes without complications(20), 0.884 for individuals with advancedtype 2 diabetes without complications(20), 0.746 for individuals with CVD(21), 0.620 for individuals with ESRD(22), 0.559 for individuals with amputa-tion (22), 0.536 for individuals with blind-ness (22), and 0 for the individual whodied. The deterioration of utility is assumedto occur at years 6 and 10 for untreatedcase patients and treated case patients,respectively (23,24). The utility weight foreach undetected case of type 2 diabeteswas assumed to be the same as the value

for each detected case of type 2 diabeteswithout complications (25).

CostsIn this study, costs were considered forHbA1c testing and treatment for type 2diabetes and complications, including thecosts of confirmatory HbA1c testing bya doctor. These cost data were mainlyestimated using the published literatureand expert opinion. Costs are expressedin USD as well as JPY. Cost was not ad-justed for the consumer price index inthis article because consumer prices havebarely changed. These costs were usedfor modeling in the decision tree andMarkov model (Table 1). The cost forHbA1c testing at SMOs is shared bypharmacies and individuals. The costs for

confirmatory testing and treatment fortype 2 diabetes and complications areshared by social insurance and patients.The direct costs for treatment for type 2diabetes were mainly estimated usingmedical fee tariffs; costs for sectorsother than health and productivity los-ses were uncounted for in this study.

The cost of each HbA1c test at SMOs,USD 12 (JPY 1,229), was estimated basedon data from the RADD Project. The costsconsist of fixed/variable capital costs andlabor cost. The fixed capital cost, USD1 (JPY 122), is calculated from the annu-alized capital cost of the testing equip-ment and an expected 732 tests, accordingto the RADD Project, in a year. The vari-able capital cost includes the price of atest kit: USD 6 (JPY 560) per participant.

Figure 1dContinued.

care.diabetesjournals.org Shono and Associates 1221

The pharmacist’s labor cost for a 15-mintest, USD 5 (JPY 547), is calculated froman average hourly wage of USD 22 (JPY2,189) for a pharmacist (26). The cost ofan HbA1c test at the SHC was estimatedto be USD 9 (JPY 918) (27).The costs of confirmatory testing and

treatment for type 2 diabetes and com-plications were estimated as follows. Forpeople with positive HbA1c test results atSMOs, a cost of USD 78 (JPY 7,800) wasset for additional testing (22). The annualcosts for outpatient treatment for type 2diabetes was set at USD 3,700 (JPY370,000) (28). The annual cost for ESRDtreatment was set at USD 48,230 (JPY4,823,000) (29). The cost in the first yearfor CVD treatment was set at USD 19,990(JPY 1,999,000), and USD 1,790 (JPY179,000) in the second year (21). The costof amputation in the first year was set atUSD 22,910 (JPY 2,291,000) (30). Theannual cost of blindness was set at USD760 (JPY 76,000) (30). Most data wereestimated from the medical care feeschedule in Japan.We took a life-long time horizon, and

theMarkov cyclewas repeated as the agestratum reached 100 years old.

DiscountingCosts and outcomes were discounted at arate of 3% (17).

Sensitivity AnalysisIn this study, we performed one-way sen-sitivity analyses for ourmodel assumptions.In the analyses, lower and upper limitswere applied as plausible values from refer-ences in case there was an absence ofinformation on values, such as 95% CI(Supplementary Table 1).We also conducted a probabilistic sen-

sitivity analysis using a set of 1,000 MonteCarlo simulations. The distributions for

variables of probabilities, utilities, and thediscount rate were assumed to have tri-angular distribution, and those for varia-bles of costs were assumed to have ag-distribution. The lower and upper limitranges in the triangular distribution wereapplied in the same way as in the one-way sensitivity analysis.

ComparisonIn this study, we compared two scenar-ios to be chosen by pharmacy managers:1) status quo, defined as HbA1c testingonly at SHC visits and conventional op-portunistic screening in clinics, and 2)HbA1c testing at SMOs as a complementto the status quo. We calculated the re-spective costs and effects of these sce-narios, and the difference between costsand effectiveness was compared. Thatis, we calculated incremental cost-effectiveness ratios (ICERs; cost perQALYs)over the lifetime analytic horizon asthe primary end point of the cost-effectiveness analysis. ICERswere definedas follows:

ICER ¼ Incremental cost

Incremental effectiveness

¼ Cost HbA1c testing at SMOs-Cost status quo

Effectiveness HbA1c testing at SMOs-Effectiveness status quo

Model outcomes were analyzed usingTreeAge Pro 2016 (TreeAge Software,Inc., Williamstown, MA).

RESULTS

Base-Case AnalysisThe total cost per individual for thoseindividuals 40–74 years of age was esti-mated to be USD 4,951 (JPY 495,053) forHbA1c testing at SMOs, compared withUSD 5,478 (JPY 547,775) for the status

quo scenario. QALYs were estimatedto be 17.996 for HbA1c testing at SMOscompared with 17.975 for the status quo.Therefore, the total gain in QALYs wasgreater, and the total cost was lower forHbA1c testing at SMOs compared withthe status quo. Thus, this cost-effectivenessanalysis showed that compared with thestatus quo, introducing HbA1c testing inSMOs is more effective and less costly,which means cost saving with improvedQALYs for the population 40–74 years ofage (Table 2). We conducted a subgroupanalysis by each age-group and assumedthe participant cohort to be 40–74 yearsof age (in 5-year increments). The largestoutcome gain in QALYs was estimated tobe among individuals 60–64 years of age.The introduction of HbA1c testing at SMOsremained dominant in all age-groupsfrom 40 to 74 years (Table 2). We alsoconducted a subgroup analysis focusingon the participant cohort as specific age-groups in those participants 40–74 yearsof age (in 5-year increments) (Supplemen-tary Table 2).

Stability of Cost-effectivenessSensitivity analysis is commonly used forhandling uncertainty. Figure 2A showsthe results of one-way sensitivity analy-sis for the top 10 variables that effectICER, depicted as tornado diagrams. Theparameters that have a potential effecton ICER could differ depending on thevariable. Each transition probability forprogression to ESRD among individualswith type 2 diabetes with and withouttreatment for type 2 diabetes affectedthe ICERs. ICER was sensitive to the utilityof ESRD and to each probability for deathbecause of ESRD among both male andfemale participants.

Figure 2B shows the results of proba-bilistic sensitivity analyses by depicting1,000 ICERs produced by Monte Carlosimulations. HbA1c testing in SMOs wasestimated to cost less and gain more in63.8% of simulations. This was also esti-mated to be cost-effective in 90.6% ofsimulations, including dominant cases,which were under the threshold ofUSD 50,000 (JPY 5,000,000) per QALYgained; 97.7% of simulations, includingdominant cases, were under the thres-hold of USD 100,000 (JPY 10,000,000)per QALY. Figure 2C shows scatter plotsof the cost-effectiveness plane; 0% ofsimulations showed negative QALYgained.

Table 1—Model assumption on costing

Annual cost per persons

Base-case value

ReferencesUSD JPY

HbA1c check in SMO 12 1,229 26 and expert opinion

HbA1c check in SHC 9 918 27

Detailed examination 78 7,800 22

Type 2 diabetes treatment 3,700 370,000 28

ESRD treatment 48,230 4,823,000 29

CVD treatmentFirst year 19,990 1,999,000 21Second year 1,790 179,000 21

Amputation treatment (first year) 22,910 2,291,000 30

Blindness treatment 760 76,000 30

1222 Cost-effectiveness of Walk-in HbA1c Screening Diabetes Care Volume 41, June 2018

CONCLUSIONS

This study assessed the cost-effectivenessof a new walk-in fingertip HbA1c testingstrategy at SMOs in Japan, such as at com-munity pharmacies. The results showedthat HbA1c testing services available atSMOs offer cost savings compared withthe status quo of HbA1c testing only atSHC visits and conventional opportunisticscreening in clinics.We focused on walk-in fingertip HbA1c

testing for the early detection of diabetesthrough HbA1c screening in a pharmacysetting, a unique point of this study. If

given the opportunity, some individualswho test positive in this setting wouldseek confirmatory testing and receive adiagnosis by a doctor, which could lead toearlier intervention. This study is differ-ent from other cost-effectiveness studiesthat have focused on type 2 diabetesscreening strategies such as universal,community-based, targeted, and oppor-tunistic screening conducted togetherwith management for other diseaseslike hypertension (6,31). Complemen-tary HbA1c testing in SMOs could benefitthe population who live far from the SHC

or checkup sites or conventional oppor-tunistic screening.

Based on the RADD Project background,we modeled walk-in fingertip HbA1c test-ing in an SMO setting. In this model, wetook the following points into account.First, the percentage of individuals whohave never undergone routine SHC che-ckups was 21.2% on average (for individ-uals 40–74 years of age). Second, theproportion of individuals with HbA1clevel $6.0 mmol/mol among all partic-ipants (age range, 40–74 years) who hadnever had a health checkup was 32.8%.

Table 2—Results of cost-effectiveness analysis

Age of targetpopulation

Cost($ and U)

Incrementalcost ($ and U)

Effectiveness(QALYs)

Incrementaleffectiveness (QALY) ICER([$ andU]/QALY)*

Age 40–74 years

Status quo $5,478 17.975

U547,775HbA1c testing at SMOs $4,951 2$527 17.996 0.0203 Cost less, gained more

U495,053 2U52,722Age 40–44 years

Status quo $5,478 17.975

U547,775HbA1c testing at SMOs $5,407 2$71 17.977 0.0021 Cost less, gained more

U540,718 2U7,057Age 45–49 years

Status quo $5,478 17.975

U547,775HbA1c testing at SMOs $5,418 2$60 17.977 0.0019 Cost less, gained more

U541,769 2U6,007Age 50–54 years

Status quo $5,478 17.975

U547,775HbA1c testing at SMOs $5,392 2$86 17.978 0.0030 Cost less, gained more

U539,164 2U8,612Age 55–59 years

Status quo $5,478 17.975

U547,775HbA1c testing at SMOs $5,393 2$85 17.979 0.0032 Cost less, gained more

U539,281 2U8,494Age 60–64 years

Status quo $5,478 17.975

U547,775HbA1c testing at SMOs $5,379 2$99 17.980 0.0041 Cost less, gained more

U537,919 2U9,856Age 65–69 years

Status quo $5,478 17.975

U547,775HbA1c testing at SMOs $5,396 2$82 17.979 0.0037 Cost less, gained more

U539,615 2U8,160Age 70–74 years

Status quo $5,478 17.975

U547,775HbA1c testing at SMOs $5,432 2$45 17.978 0.0022 Cost less, gained more

U543,239 2U4,536

JPY, U; USD, $. *ICER (cost per QALY). Status quo defined as HbA1c testing only at SHC visits and conventional opportunistic screening in clinics. HbA1ctesting services available at SMOs.

care.diabetesjournals.org Shono and Associates 1223

Therefore, ;7% (21.2% 3 32.8%) of par-ticipants newly learned that they possiblyhad type 2 diabetes in the pharmacysettings. There is an undiagnosed popula-tionwith type 2 diabetes of;3.3million inJapan (10). Thismodel is generalized for thepopulation who are unaware of walk-infingertip HbA1c testing besides a lack ofaccess to SHC checkups.We consider our results to be stable

and robust since most simulations areunder the threshold of USD 50,000 (JPY5,000,000) per QALY gained. In one-waysensitivity analysis, this model was mostsensitive to probabilities of transitionfrom type 2 diabetes to ESRD with and

without treatment for type 2 diabetes.The speed of the progress to ESRD byearly detection of type 2 diabetes andintervention for effective managementto delay to ESRD could affect the ICERresults (32). The probability for deathbecause of ESRD was also a sensitive pa-rameter to ICER. However, the crude deathrate has remained stable at between9.2% and 10.2% for the last 20 years (33).

Our results will be useful to managersof pharmacies or other health institutionsand/or to policy makers in local govern-ment. Local policies, such as the sub-sidy program in Adachi Ward, Tokyo, forHbA1c testing services provided in 10

pharmacies as SMOs since April 2015,are worth consideration. The similar sub-sidy programs are currently implementedin Mihara-city, Hiroshima, and Gyoda-city,Saitama, as well. Our results also haveimplications for policy makers in foreigncountries that use community pharmaciesfor diabetes risk assessment programs (9).

This study has several limitations thatmust be considered when interpretingthe results. First, the model in this studyis a simplified economic model. We inputfour major complications of type 2 di-abetes and considered each as a singleitem in the model. However, the conse-quences of type 2 diabetes vary in the

Figure 2—A: Sensitivity analysis: results of one-way sensitivity analysis. B: Sensitivity analysis: results of probabilistic sensitivity analyses. C: Sensitivityanalysis: scatter plots of cost-effectiveness plane.

1224 Cost-effectiveness of Walk-in HbA1c Screening Diabetes Care Volume 41, June 2018

real world, including the complicationsused in the model. Additionally, we sim-plified the definition of CVD to includecoronary heart disease and stroke in themodel. The cost was also determinedusing the same assumption (21). Second,although the effect of treatment on type 2diabetes would depend upon treatmentuptake, types, and adherence, we as-sumed a population with 100% adher-ence to treatment in the model. The ICERmay be biased as either lower or higherwhenparticipants stopmedication.Whereasthe effectiveness is overestimated, costsmight be either lower or higher. Third, itis quite difficult to identify the transitionprobability of each complication for thepopulation of individuals who did notundergo testing. Therefore, those transi-tion probabilities were estimated usingsequential/serial data, such as checkupdata. Additionally, the possibility of mul-tiple complications may not have beenexcluded from the probabilities. Fourth,the utility weight was estimated as higher,at 9 years for treated case patients and5 years for untreated case patients(23,24). This estimation of a 4-year gapmight be an underestimated outcome;however, this gap affected the outcomesslightly. Fifth, testing HbA1c does not al-ways provide a robustmethod for screen-ing for type 2 diabetes. However, weconsidered that it was a suitable methodin the pharmacy setting because of con-venience. Finally, we should consider thedifferent progression of type 2 diabetes

at a timing of SMO testing as heteroge-neity. The effect of treatment for type 2diabetes, that is, the transition proba-bility of complications, could differ de-pending on the baseline. However, thereare no data identifying the exact durationof type 2 diabetes, especially for patientswho have never undergone testing orreceived medication. The base case wasassumed using results from national sur-veys, such as the National Health andNutrition Survey, and/or those fromretrospective/prospective research sur-veys in Japan.

In summary, this study suggests thatthe introduction of HbA1c testing in SMOscan be justified as an efficient use of finitehealth resources. Therefore, introducingopportunistic HbA1c testing at SMOs thatare mostly in community pharmacy set-tings is worth consideration by pharmacymanagers and programs subsidized bylocal governments.

Acknowledgments. The authors thank theAdachi-ku Pharmaceutical Association, the Ada-chi Diabetes Mellitus Society, The Adachi-kuMedical Association, and The Cooperation Coun-cil for Specimen Measurement Offices. Theauthors also thank Akiko Nagai of the AyaseCommunity Pharmacy.Funding. This study was funded by The ResearchFoundation for Pharmaceutical Sciences. This re-search was supported by the Mitsukoshi Healthand Welfare Foundation research fund (2010).The funding source did not play a role in any

aspect of the study or in the decision to submitthe manuscript for publication.

Duality of Interest. No potential conflicts ofinterest relevant to this article were reported.Author Contributions. A.S. wrote the manu-script and researched the data. M.K. contributedto the discussion and edited themanuscript. S.-l.H.contributed to the research and discussion.R.O. contributed to the research. N.Y. contrib-uted to the discussion. A.S. is the guarantor ofthis work and, as such, had full access to allthe data in the study and takes responsibility forthe integrity of the data and the accuracy of thedata analysis.

References1. American Diabetes Association. Screening fortype 2 diabetes. Diabetes Care 2003;26(Suppl. 1):S21–S242. Gemma R, Kitahara K, Kono E, et al. Clinicalbackground factors of patients with advanceddiabetic retinopathy at the first presentation. JJapan Diab Soc 2015;58:192–197 [in Japanese]3. Nathan DM. Diabetes: advances in diagnosisand treatment. JAMA 2015;314:1052–10624. Selph S, Dana T, Blazina I, Bougatsos C, Patel H,Chou R. Screening for type 2 diabetes mellitus:a systematic review for the U.S. Preventive ServicesTask Force. Ann Intern Med 2015;162:765–7765. Narayan KM, Weber MB. Screening for hyper-glycemia: the gateway to diabetes preventionand management for all Americans. Ann InternMed 2015;162:795–7966. Echouffo-Tcheugui JB, Ali MK, Griffin SJ,Narayan KM. Screening for type 2 diabetes anddysglycemia. Epidemiol Rev 2011;33:63–877. Kohro T, Furui Y, Mitsutake N, et al. TheJapanese national health screening and interven-tion program aimed at preventing worsening of themetabolic syndrome. Int Heart J 2008;49:193–2038. NHS England. NHS Health Check implementa-tion review and action plan [article online], 2013.Available from https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/224805/NHS_Health_Check_implementation_review_and_action_plan.pdf. Accessed 7 December 2016

Figure 2—Continued.

care.diabetesjournals.org Shono and Associates 1225

9. Willis A, Rivers P, Gray LJ, Davies M, Khunti K.The effectiveness of screening for diabetes andcardiovascular disease risk factors in a communitypharmacy setting. PLoS One 2014;9:e9115710. Ministry of Health, Labour and Welfare.Heisei 24 nen kokumin kenkou eiyoutyousakekkano gaiyou [article online]. Available from http://www.mhlw.go.jp/stf/houdou/0000032074.html.Accessed 5 June 2017 [in Japanese]11. Iseki K, Asahi K, Moriyama T, et al. Risk factorprofiles based on estimated glomerular filtrationrate and dipstick proteinuria among participantsof the Specific Health Check and Guidance Systemin Japan 2008. Clin Exp Nephrol 2012;16:244–24912. Ministry of Health, Labour and Welfare. Heisei26 nendo tokutei kenkoushinsa/tokuteihokenshidouno jisshi joukyou ni tsuite [article online], 2014.Available from http://www.mhlw.go.jp/bunya/shakaihosho/iryouseido01/info03_h26.html. Ac-cessed 7 December 2016 [in Japanese]13. Maruko M. The Japan Times: simple testsfill health-check gaps [article online]. Availablefrom http://www.japantimes.co.jp/news/2013/12/30/reference/simple-tests-fill-health-check-gaps/. Accessed 7 December 201614. The Cooperation Council for Specimen Mea-surement Offices. Trends in the number of SMOsin Japan [article online]. Available from http://www.yubisaki.org/. Accessed 1 March 201815. Adachi City. Yakkyoku de tounyoubyo risukuwo souki hakken! [article online]. Available fromhttp://www.city.adachi.tokyo.jp/kokoro/fukushi-kenko/kenko/diabetes-drugstore.html. Accessed7 December 2016 [in Japanese]16. Yahagi N. Fingertip screen using POCT HbA1canalyzer at community pharmacies is effective forearly recognition of diabetes. Rinsho Byori 2016;64:564–566

17. Drummond MF, Sculpher MJ, Claxton K,Stoddart GL, Torrance GW. Methods for theEconomic Evaluation of Health Care Programmes.Oxford, Oxford University Press, 201518. Hayes AJ, Leal J, Gray AM, Holman RR, ClarkePM. UKPDS outcomesmodel 2: a new version of amodel to simulate lifetime health outcomes ofpatients with type 2 diabetes mellitus using datafrom the 30 year United Kingdom ProspectiveDiabetes Study: UKPDS 82. Diabetologia 2013;56:1925–193319. Ministry of Health, Labour and Welfare. Vitalstatistics [article online]. Available from http://www.mhlw.go.jp/english/database/db-hw/vs01.html. Accessed 23 October 201520. Sakamaki H, Ikeda S, Ikegami N, et al. Mea-surement of HRQL using EQ-5D in patients withtype 2 diabetes mellitus in Japan. Value Health2006;9:47–5321. Okubo R, Kondo M, Hoshi SL, Yamagata K.Cost-effectiveness of obstructive sleep apneascreening for patients with diabetes or chronickidney disease. Sleep Breath 2015;19:1081–109222. Hisashige A. Sippeikanri niyoru hoken sabisno keizaiteki hyoka: report of Health and LabourSciences Research Grants [article online]. Availablefrom https://mhlw-grants.niph.go.jp/. Accessed 1August 2014 [in Japanese]23. Bacon CG, Hu FB, Giovannucci E, Glasser DB,Mittleman MA, Rimm EB. Association of typeand duration of diabetes with erectile dysfunc-tion in a large cohort of men. Diabetes Care 2002;25:1458–146324. Nihon Tounyobyo Taisaku Suishin Kaigi.Nihon ni okeru Tounyobyokanja no ashi gaikanijou oyobi Tounyobyo shinkeishougai no jittai nikansuru houkoku [article online], 2008. Available

from http://www.med.or.jp/jma/diabetes/. Ac-cessed 15 December 2013 [in Japanese]25. Gillies CL, Lambert PC, Abrams KR, et al.Different strategies for screening and preventionof type 2 diabetes in adults: cost effectivenessanalysis. BMJ 2008;336:1180–118526. Nenshu-labo. Yakuzaishi no heikin nensyu[article online]. Available from http://nensyu-labo.com/. Accessed 23March 2017 [in Japanese]27. JA Kagoshima Kouseiren. Available fromhttp://www.kago-ksr.or.jp/ken/kensin/kensin01.Accessed 29 October 2015 [in Japanese]28. Fukuda H, Ikeda S, Shiroiwa T, Fukuda T. Theeffects of diagnostic definitions in claims data onhealthcare cost estimates: evidence from a large-scale panel data analysis of diabetes care in Japan.Pharmacoeconomics 2016;34:1005–101429. Ota K, Shishido K, Tsuchiya S, et al. Dai 19-kaitouseki iryohi jittai chosa houkoku. J Japan SocDialysis Phys 2016;31:90–103 [in Japanese]30. Ikeda S, Kobayashi M. Cost-effectivenessanalysis of metformin or pioglitazone combinedwith sulfonylurea in type 2 diabetes mellitus inJapan. J Japan Diab Soc 2010;53:469–475 [inJapanese]31. Einarson TR, Bereza BG, Acs A, Jensen R.Systematic literature review of the health eco-nomic implications of early detection by screen-ing populations at risk for type 2 diabetes. CurrMed Res Opin 2017;33:331–35832. Iseki K. Predictors of diabetic end-stage renaldisease in Japan. Nephrology (Carlton) 2005;10(Suppl.):S2–S633. Japanese Society for Dialysis Therapy. Anoverview of regular dialysis treatment in Japanas of Dec. 31, 2015 [article online]. Available fromhttp://www.jsdt.or.jp/. Accessed 13October 2017[in Japanese]

1226 Cost-effectiveness of Walk-in HbA1c Screening Diabetes Care Volume 41, June 2018