Cost-benefit analysis of screening for diabetic retinopathy among Chinese type 2 diabetes: decision from different perspectives of a population-based study

doi:10.4236/health.2009.14050

Paper Menu >>

Journal Menu >>

Vol.1, No.4, 304-312 (2009)

doi:10.4236/health.2009.14050

HEALTH

Cost-benefit analysis of screening for diabetic

retinopathy among Chinese type 2 diabetes: decision

from different perspectives of a p opulation-based study

in Kinmen, Taiwan

Ching-Ching Peng1, Pesus Chou2, Tao-Hsin Tung1,3, Wu-Hsiung Chien1, Fenq-Lih Lee4, Shih-Jen

Chen4, Shih-Tzer Tsai5, Chi-Ming Liu1,2, Jorn-Hon Liu1,6*

1Cheng Hsin General Hospital, Taipei, Taiwan, China; ch2876@chgh.org.tw

2Community Medicine Research Center & Institute of Public Health, National Yang-Ming University, Taipei, Taiwan, China

3Faculty of Public Health, School of Medicine, Fu-Jen Catholic University, Taipei, Taiwan, China

4Department of O p hthalmology, Ve terans General Hospital, Taipei, Taiwan, China

5Department of Medicine, Veterans General Hospital, Taipei, Taiwan, China

6Faculty of Medicine, School of Medicine, National Ya ng-Ming University, Taipei, Taiwan, China

*Corresponding author

Received 3 September 2009; revised 14 September 2009; accepted 15 September 2009.

ABSTRACT

The purpose of this study is to explore whether

it is worthwhile to launch a routine diabetic

retinopathy (DR) screening for blindness pre-

vention among Chinese type 2 diabetes from

different perspective based on the popula-

tion-based study in Kinmen, Taiwan. A total of

971 community dwelling adults previously di-

agnosed with type 2 diabetes in 1991-1993 un-

derwent DR screening in 1999-2002 by a panel

of ophthalmologists using on-site indirect oph-

thalmoscopy and 45-degree color fundus retinal

photographs. The cost-benefit analysis is used

to evaluate the DR screening. In terms of bene-

fit-cost ratio, the different screening programs

for DR could save New Taiwan Dollars (NTD)

from 14.38 to 36.83 in discounted costs for each

dollar incurred in different screening years from

the societal viewpoint for Taiwan and save NTD

from 0.81 to 1.80 in different screening years

from health care payer’s perspective. The av-

erage estimate of willingness-to-pay to translate

into benefit yields NTD from 937.8 to 4,689 be-

nefits per case due to DR screening in different

screening years during 10-year follow-up. The

net present value of the DR screening w ere NTD

from -167,318 to -307,251.2 in different screening

years. In conclusion, it is worthwhile to initial a

routine DR screening of Chinese type 2 d iabetes

for blindness prevention from the societal per-

spective but not from consumer decision based

on the willingness-to-pay perspective.

Keywords: Type 2 Diabetes; Diabetic Retinopathy;

Cost-Benefit Analysis; Population-Based Study

1. INTRODUCTION

In developed countries, diabetic retinopathy (DR) is a

major microvascular disease and is associated with in-

creased visual impairment in type 2 diabetics [1]. In

Taiwan, previous community-based studies showed that

the prevalence and annual incidence density of DR were

an estimated 15% to 45% and 6.62x10-2/year (95%CI:

5.36x10-2/year—8.06x10-2/year), respectively [2,3]. Since

the average time from development of no DR to blind-

ness is approximately 26.5 years in persons with type 2

diabetes, assessing the progression of DR by screening is

a worthwhile preventive measure [4].

Whether screening for DR is worthwhile is also con-

tingent on whether subjects are willing to pay the eye

screening program that would decrease the risk of

blindness. Benefit due to the reduction of severe com-

plication in cost-benefit analysis is often measured by

the human capital approach for which the value of eye

screening program is measured by its effect on the pa-

tient’s sight-time earning. According to welfare eco-

nomic theory, the benefit to an individual of a service or

an intervention is defined as that individual’s maximum

willingness to pay (WTP) for the service or intervention

[5]. WTP is a contingent valuation and involves using a

hypothetical survey to directly ask individuals the

maximum amount they are willing to pay for the com-

modity in question [6,7]. Due to the medical resources

are limited, the government is in a dilemma about

C. C. Peng et al. / HEALTH 1 (2009) 304-312

305

whether it is necessary to popularize the eye screening

programs among type 2 diabetes into a nationwide pro-

gram. Our previous study have showed that degree of

DR was the independent factor affecting WTP values in

DR screening among community-dwelling adults with

type 2 diabetes [8]. The purpose of this study is further

to conduct a cost-benefit analysis to determine whether it

is worthwhile to launch eye screening program for

blindness prevention among population-based adults

diagnosed with type 2 diabetes in Kinmen, Taiwan.

2. METHODS

2.1. Study Design and Subjects

The pro cedu r es fo r co st - b en ef it analysis of DR screening

among community dwelling patients with type 2 diabe-

tes in 1999-2003 showed in Figure 1. Data used in this

study were derived from a population-based screening

for type 2 diabetes targeted to subjects aged 30 years or

more in Kinmen, Taiwan, between January 1991 and

December 1993. Details of the study design and execu-

tion have been described in full elsewhere [9]. Identifi-

cation for type 2 diabetes was based on the WHO 1999

definition [10], that is, su bjects with FPG  126mg/dl or

2-hr postload glucose concentration ≥ 200mg/dl were

defined as having type 2 diabetes. Subjects with a his-

tory of type 2 diabetes and who received medication

were defined as known cases. A total of 1,123 cases of

type 2 diabetes aged 30 and over were found based on

the population survey carried out by the Yang-Ming

Crusade, which was organized from the medical students

of the National Yang-Ming University, Taipei, Taiwan.

The screened diabetic subjects then were referred to the

region hospital for further treatment and follow-up rou-

tinely. Of 1,123 subjects with type 2 diabetes, 152 emi-

grated or died between 1994 and 1998. After excluding

these subjects, the remaining 971 had the uptake of fun-

dus check-up annually since 1999. A panel of popula-

tion-based follow-up screening of DR was then con-

ducted annually from 1999 to 2002. These 971 partici-

pants were invited to receive eye screening by invitation

letter or calls. Based on the eye screening results, differ-

ent treatment strategies were used, that is, routine fol-

low-up for patients with mild or moderate DR and laser

photocoagulation for patients with severe DR. In addi-

tion, informed consent was obtained from all participan ts

before the investigation began. Access to personal re-

cords was approved by the hospital human subjects re-

view board at Cheng-Hsin General Hospital, Taipei,

Taiwan.

We initiated a DR screening program after six years of

mass screening because Kinmen is an offshore island

from Taiwan lacking medical resources. Here, DR

screening requires mobilizing manpower and equipment,

coordinating between clinical personnel and field work,

and transporting to the island. By 1999, a team for DR

screening was successfully organized, including 4 well-

trained senior ophthalmologists from Veterans General

Hospital, Taipei, 4 clinical nurses, and 20 medical stu-

dents from the Yang-Ming Crusade.

2.2. Screening and Diagnosis for Diabetic

Retinopathy

The diagnosis of DR was based on the on-site direct and

indirect ophthalmoscopic examination and the sin-

gle-field fundus photographs that analyzed later. On-site

screening was conducted by two ophthalmologists using

direct and indirect ophthalmoscopy after pupil dilatation

with topical 0.5% mydriacyl. The graders wrote down

the diagnosis and marked the findings in the record.

Then the subj ects were taken on e 45-d egree color fu ndus

photography with Polaroid 600 films (Polaroid, Nether-

land) centered at macula in each eye using Topcon fun-

dus camera (TRC-50VT, Tokyo, Japan). The single-field

photographs were then printed out and filed. Grading of

the photographs was done by two well-trained senior

ophthalmologists beginning no later than one month

after the screening. The final grading of the DR depends

on the summed interpretation of the photographs and the

recorded ophthalmoscopic gradings. From the Reti-

nopathy Disease Severity Scale, diabetic subjects were

classified according to the most severe changes in the

worse eye [11]. In addition, legal blindness was defined

as a best-corrected visual acuity of 0.1 (6/60) or worse in

the better eye [8].

A pilot study was performed in 50 randomly selected

type 2 diabetic patients to set up the consistent diagnosis

of DR among ophthalmologists. The kappa statistic of

0.73 (95%CI: 0.48—0.98) between the observers re-

vealed a good interobs erver reliability.

2.3. Cost-Benefit Analysis of Screening for

Diabetic Retinopathy

2.3.1. Markov Decision Model of Screening for

Diabetic Ret inopathy

In this study, the cost-benefit analysis tool of screening

for DR among subjects with type 2 diabetes was based

on TreeAge software (DATA 3.5, Tree-Age, Inc., Wil-

liamstown MA) for medical decision analysis. A deci-

sion analysis using the Markov Decision Model was

constructed to compare different screening regimes for

DR with no screening group (see Figure 2). The as-

sumption of no screening group was that except eye

screening, diabetic patients still received routine medical

care until they become blind. Accord ing to the theory of

stochastic process, the Markov chain model is deter-

mined by both the initial state and the transition matrix.

The model starts from the decision to screen or not to

screen and the overall expected value is based on ex-

pected values of end nodes rather than all nodes. For

each decision, there are six states of disease natural his

C. C. Peng et al. / HEALTH 1 (2009) 304-312

306

Table 1. Cost assumptions, willingness-to-pay value, transition probabilities in decision analysis

of screening for diabetic retinopathy.

Parameter Value

Annual direct cost (NT dollars)

Screening cost1 2,298

Drug cost2 10,857

Regular clinics fee3 509

Laser photocoagulation4 10,970

Vitrectomy5 10,840

Total 35,474

Annual indirect cost (NT dollars)

Gross Domestic Product, GDP 452,168

Willingness-to-pay value(NT dollars) [8]

No DR 440.1331.6

NPDR 450.0298.8

PDR 683.3285.5

Legal blindness 822.2192.2

Total 468.9327.7

Annual transition probability (%) [4]

No DRMild NPDR 7.37

Mild NPDRModerate NPDR 19.37

Moderate NPDRSevere NPDR 17.41

Severe NPDRPDR 28.95

PDRLegal blindness 21.10

1Screening cost includes clinician’s fee, vision examination, pupil dilation, slit lamp contact mirror fundus-

copy, funduscopic exam, HbA1c, SMA-12 test , and manpower cost.

2According to the drug usage distribution from Taiwanese Association of Diabetes Educators (TADE) study

in 2004 and the payment of N atio nal Heal th Insurance.

3Regular clinics fee includes clinician’s fee and pharmacist’s fee

4Laser photocoagulation cost includes panretinal photocoagulation, two fundus color photos, and fluorescein

angiography (FAG).

5Vitrectomy cost includes vitretomy, two fundus color photos, and fluorescein angiography (FAG).

tory of DR including NDR, mild NPDR, moderate

NPDR, severe NPDR, PDR, and blindness. The initial

state distribution is based on the results of the present

study. Transition probabilities from one state to another

representing the disease natural history of DR were de-

rived from our empirical estimation, that is, the annual

transition probabilities for each stage to the next are as

follows: mild NPDR to moderate NPDR 19.37%, mod-

erate NPDR to severe NPDR 17.41%, severe NPDR to

PDR 28.95%, and PDR to blindness 21.10% [4]. For

each scenario, we calculated the expected probability of

patients aggregate experience that is accumulated in

each state during 10-year foll ow-up.

2.3.2. An Empirical Survey for the Estimation of

Cost and Willingness to Pay

Costs in the present study include direct and indirect cost.

Direct costs include cost of DR screening, drug cost,

cost of regular clinic fee, and treatment cost (for exam-

ple, laser photocoagulation and surgery). Indirect cost

includes only productivity loss of the patient because of

time taken off work for treatment. The average time

taken off work for treatment depends on the professions’

opinion. In addition, WTP was assessed by the following

question: “What is the most price that you would be

willing to pay for routine screening for DR that reduces

the risk of fully blindness?” The WTP amounts

C. C. Peng et al. / HEALTH 1 (2009) 304-312

307

Table 2 . Cost-benefit analysis using the human-capital approach of different screening programs for diabetic reti-

nopathy.

Screening group Non-screening groupNet cost

Annual screening

Direct cost

Screening cost 21,900 0 21,900

Other cost 113,403 131,219 -17,816

Total 135,303 131,219 4,084

Indirect cost 36,704 333,911 -297,207

Total (Direct + Indirect cost) 172,007 465,130 -293,123

Biennial screening

Direct cost

Screening cost 10,950 0 10,950

Other cost 116,216 131,219 -15,003

Total 127,166 131,219 -4,053

Indirect cost 70,435 333,911 -263,476

Total (Direct + Indirect cost) 197,601 465,130 -267,529

3-yearly screening

Direct cost

Screening cost 7,300 0 7,300

Other cost 118,832 131,219 -12,387

Total 126,132 131,219 -5,087

Indirect cost 107,629 333,911 -226,282

Total (Direct + Indirect cost) 233,761 465,130 -231,369

4-yearly screening

Direct cost

Screening cost 5,475 0 5,475

Other cost 121,199 131,219 -10,020

Total 126,674 131,219 -4,545

Indirect cost 144,729 333,911 -189,182

Total (Direct + Indirect cost) 271,403 465,130 -193,727

5-yearly screening

Direct cost

Screening cost 4,380 0 4,380

Other cost 123,330 131,219 -7,889

Total 127,710 131,219 -3,509

Indirect cost 180,479 333,911 -153,432

Total (Direct + Indirect cost) 308,189 465,130 -156,941

C. C. Peng et al. / HEALTH 1 (2009) 304-312

308

Table 3. The benefit-cost ratio estimates of different screening programs for diabetic retinopathy.

Benefit-cost ratio

Payer’s perspective Society perspective

Annual screening 0.81 (17,816/21,900) 14.38 ((17,816+297,207)/21,900)

Biennial screening 1.37 (15,003/10,950) 25.43 ((15,003+263,476)/10,950)

3-yearly screening 1.70 (12,387/7,300) 32.69 ((12,387+226,282)/7,300)

4-yearly screening 1.83 (10,020/5,475) 36.38 ((10,020+189,182)/5,475)

5-yearly screening 1.80 (7,889/4,380) 36.83 ((7,889+153,432)/4,380)

Figure 1. The procedure of cost-benefit analysis of screening for diabetic retinopathy among type 2 diabetics in Kinmen.

Cost-benefit analysis

Cost-benefit analysis of diabetic

retinopathy screening

Laser photocoagulation patients

with severe DR

Routine follow up for patients

with mild or moderate DR

725 Type 2 diabetes received DR

screening

Screening for

diabeti retin opathy

(1999-2002)

DR screening for Type 2 diabetics in

1999-2002 (n=97 1)

Loss to follow-up

(n=246)

Routine treatment and follow

up by region hosp ital Type 2 diabetes

(n=1,123)

WHO (1999) criteria

1994-1998

Migrated out or death

(n=152)

Screening for type 2

diabetes

(1991-1993)

1991-1993

A community-based screening

regime f or type 2 diabetes (n=12,489)

C. C. Peng et al. / HEALTH 1 (2009) 304-312

309

mildNPDR

NDR

others others

others

others ot hers

others

PDR

PDR PDR

PDR

onNDR

onND

onNDR

Blindness

Blindness Blindness

Blindness

Blindness Blindness

severeNPDR

ModerateNPDR

severeNPDR

ModerateNPDR

severeNPDR

ModerateNPDR

severeNPDR

MildNPDR

moderateNPDR

MildNPDR

moderateNPDR

MildNPDR

moderateNPD

MildNPDR

moderateNPDR

MildNPDR

moderateNPDR

MildNPDR

moderateNPD

MildNPDR

moderateNPD

mildNPD

mildNPDR

Moderate

PDR

Severe

PDR PDR PDR

Blindness Blindness Blindness

Severe

Moderate

PDR

mildNPDR

others others PDR

Blindness PDR

Blindness

severeNPDR

on-

Screening

Group

Screening

Group

Scr eening for

Retinopathy in

Type ІІ DM patients

mildNPDR

moderateNPDR

NDR

others

onND

others

onNDR

mildNPDR

ModerateNPDR

sever eNPDR

ModerateNPDR

severeNP DR

severeNPDR

ModerateNPDR

severeNPDR

sever eNPDR

ModerateNPDR

Figure 2. Markov decision model for two options, screening and non-screening for diabetic

retinopathy.

for a routine screening for DR were elicited by dis-

crete-choice, that is, subjects were presented a single

price for a screening program that would yield a speci-

fied health change. Subjects either accept or reject the

price. By randomly varying the price across a number of

different subsamples, the mean WTP could be estimated

[8]. To maintain consistency of interview quality, all

information on WTP measurements was also collected

by one well-trained interviewer. All costs are expressed

as New Taiwan Dollars (NTD).

2.3.3. Cost-Benefit Analysis and Discount Rate

Using the human-capital approach, net cost (saving) for

different screening programs of DR, taking direct cost

and indirect cost into account, was calculated. Bene-

fit-cost ratios were calculated as the reduction blindness

costs divided by the cost of the screening programs. Us-

ing WTP approach, net present value (NPV) was also

calculated on the basis of the total benefit (calculated by

the WTP method) minus screening cost of DR. To take

time preference into account, that is, receiving benefit

Openly accessible at

C. C. Peng et al. / HEALTH 1 (2009) 304-312

310

earlier and incurring cost later, we discounted all costs

and benefits to the present value at 5% annually.

3. RESULTS

The annual direct cost, annual indirect cost, WTP value,

and annual transition probability in decision analysis of

DR screening are shown in Table 1. Direct costs include

screening cost, drug cost, regular clinics’ fees, laser

photocoagulation, and vitrectomy. Indirect cost repre-

sents lost productivity according to patient’s disease

state, estimated using the Gross Domestic Product (GDP)

value in 2004.

Table 2 shows total discounted direct costs and indirect

costs using the human capital approach. Annual screening,

biennial screening, 3-yearly screening, 4-yearly screening,

and 5-yearly screening regimens could save NTD17,816,

NTD15,003, NTD12,387, NTD10,020, and NTD7,889

per case in discounted direct costs (except screening cost),

res p e ct iv e l y. From payer’s perspective, the discounted net

total direct costs for DR screening program were NTD

4,084, NTD-4,053, NTD-5,087, NTD-4,545, and NTD-

3,509 for annual screening, biennial screening, 3-yearly

screening, 4-yearly screening, and 5-yearly screening, re-

spectively. The discounted indirect costs saved per case by

DR screening program were NTD297,207, NTD 263,476,

NTD226,282, NTD189,182, and NTD153,432 for annual

screening, biennial screening, 3-yearly screening, 4-yearly

screening, and 5-yearly screening, respectively. Thi s yi el d

NTD293,123 (annual screening), NTD267,529 (biennial

screening), NTD231,369 (3-yearly screening), NTD

193,727 (4-y e a r l y s c ree ning ) , an d NTD156,941 (5-yea r l y

screening) net saving per case due to DR screening pro-

gram from the societal perspective.

As Table 3 shows, DR screening programs could save

NTD14.38 (annual screening), NTD25.43 (biennial

screening), NTD32.69 (3-yearly screening), NTD36.38

(4-yearly screening), and NTD36.83 (5-yearly screening)

from the societal viewpoint and save NTD0.81 (annual

screening), NTD1.37 (biennial screening), NTD1.70 (3-

yearly screening), NTD1.83 (4-yearly screening), and

NTD1.80 (5-yearly screening) in discounted costs for

each NTD dollar incurred in DR screening programs

from health care payer’s perspective.

The average estimate of WTP in order to reduce

blindness as shown in Table 1 is NTD468.9. Translating

this figure into benefit yields NTD4,689 (annual screen-

ing), NTD2,344.5 (biennial screening), NTD1,406.7 (3-

yearly screening), NTD1,172.3 (4-yearly screening), and

NTD 937.8 (5-yearly screening) benefit per case due to

DR screening during 10-year follow-up. The NPV of the

screening programs, taking indirect cost into account,

were NTD-167,318, NTD-195,256.5, NTD-232,354.3,

NTD-270,230.7, and NTD-307,251.2 of annual screen-

ing, biennial screening, 3-yearly screening, 4-yearly

screening, and 5-yearly screening, respectively.

4. DISCUSSION

4.1. Implications of Cost-Benefit

Analysis for Diabetic Retinopathy

Screening

In Taiwan, few population-based studies have attempted

to quantify the cost and benefit of DR screening pro-

grams. The present study uses cost-benefit analysis to

assess whether a DR screening program against non-

screening group is worthwhile in Taiwan from different

perspectives. The results indicate that indirect costs play

an important role in the evaluation of the DR screening

program. Annual screening program could save the most

(NTD297,207) per case in discounted indirect costs

compared with non-screening group. From health care

payer’s perspective, the discounted net cost for annual

screening was NTD4,084 per case. This indicates that

the benefit from the annual screening program of DR

can not outweigh the cost incurred in the DR screening

program from health care payer’s perspective. Taking

indirect cost into account, the NTD293,123 net saving

per case suggest the annual screening program is rather

worthwhile from the societal perspective.

In addition, using the WTP approach, the present

study shows it is not worthwhile to have screening for

DR from the perspective of WTP due to the negative

result of NPV value. It should be noted that the WTP

approach is a contingent-valuation method that reflects

consume surplus of getting DR screening. Since the

mean estimate of WTP (NTD468.9) for the DR screen-

ing program was far below the current expense for di-

rect costs of DR screening (NTD2,190) per case. This

suggests that they could not get any surplus from the

purchase of screening program. This accounts for why

the results of NPV are negative. If the estimate of WTP

is raised to NTD2,500, this means that if people have to

pay only NTD2,190 for benefits they value at

NTD2,500 then they get a surplus of NTD310 from the

purchase of screening for DR. Results from the WTP

suggest that the amount diabetics in Kinmen are willing

to pay for the screening program is lower than th e bene-

fit they value. In terms of consumer decision based on

the perspective of WTP, it may not be worthwhile to

launch a DR screening program.

4.2. The Efficiency and Advantage for

Routine Diabetic Retinopath y

Screening

The surveillance and treatment of diabetes-related com-

plications should be part of routine care in all type 2

diabetes [4]. The benefits of DR screening rest on the

additional time patients have to obtain treatment. If one

wishes to reduce the loss of vision associated with DR,

C. C. Peng et al. / HEALTH 1 (2009) 304-312

311

then performing the screening is more important than

the type of screening used [12,13]. During the first year

of DR screening, savings associated with preventing one

single case of blindness could cover the cost of the

screening program [12]. The threat of blindness is less

severe in type 2 diabetes because DR progresses more

slowly than in type 1 diabetes. Although the eye care

program saved 21 sight-years [14], it was less efficient

in patients with type 2 diabetes.

The optimal screening interval is determined by the

disease natural history with regard to the screening pol-

icy [4]. Many evidence-based studies have showed that

screening for and treating DR is extremely efficiency

and cost-effective. From the health insurer’s viewpoint,

routine screening and treatment of eye disease in dia-

betic patients cost USD 3190 per QALY saved [15].

Duration of blin dness drops by 0.48 and 0.13 years with

increase in year of onset of the disease whereas effec-

tiveness decreases in type 2 diabetic patients [16]. Pre-

vention screening programs aimed at improving eye

care for type 2 diabetic patients results in both highly

cost-effective health care and substantial federal budg-

etary savings [15,17]. In addition, our previous studies

have demonstrated that annual DR screening is the most

effective and efficient screening schedule for reducing

blindness compared to other screening intervals [4,17].

To take both cost and efficacy into consideration, many

organizations such as the National Committee for Qual-

ity Assurance, through Health Employer Data and In-

formation Set (HEDIS) measures also recommends that

annual eye examinations be used not only as a general

guideline, but also as a quality standard in all patients

with diabetes [18]. Better diabetes management such

more strict adherence by diabetic patients and ophthal-

mologists to best practice guidelines could have enor-

mous protective impacts in cases of visual loss caused

by DR [19].

4.3. Methodological Considerations

From the methodological viewpoint, although using a

population-based follow-up study design could reduce

selection bias and increase statistical power, using pri-

mary information and calculating both direct and indi-

rect costs help us estimate the true benefit of DR

screening more closely than which has been possible

before. There are still some critiques raised from this

study. First, only three-year follow-up period, we could

not have enough sample size to predict all of the effects

of DR screening on disease variations. Second, we did

not explicitly consider the sensitivity and specificity of

the DR screening tests. Previous studies demonstrated

that indirect ophthalmoscopy performed by ophthal-

mologists has a sensitivity of approximately 85% [20],

but this may approach 100% with newer slit lamp

biomicroscopic techniques [21]. Retinal photography, an

alternative detection method for DR among diabetic

patients, has an overall sensitivity of approximately

85% [20]. That means the accuracy of DR diagnosis

could be accepted. Third, although the Kappa value for

the agreement of interobserver reliability seemed ac-

ceptable [22], non-differential misclassification-bias

identification still could have occurred. Fourth, we did

not estimate the influence of covariates such as duration

of type 2 diabetes or HbA1c level . Because severe and

very severe NPDR without macular edema is associated

with high risk of progression to PDR, 10—50% of those

with type 2 diabetes and this level of NPDR will de-

velop PDR within 1 year [23]. Further long-term studies

should be conducted to clarify whether patients with

better glycemic control or in an early stage of DR could

benefit from less frequent screening intervals. Finally, it

should be noted that the estimates used in this analysis

were based on relatively small samples, that is, the ag-

gregate estimates may reflect a reasonable population,

but not all Chinese with type 2 diabetes. Further study

of those inadequ ately represented is needed.

5. CONCLUSIONS

In conclusion, this study revealed that it is worthwhile to

launch a routine DR screening program of Chinese type

2 diabetes for blindness prevention from the societal

perspective but not from consumer decision based on the

perspective of WTP.

6. ACKNOWLEDGEMENTS

This study was also supported by the grants from the National Science

Council (NSC-95-2314-B-350-002-MY3) and (NSC-98-2314-B-350-

002-MY3).

REFERENCES

[1] Sanchez-Thorin, J.C. (1998) The epidemiology of diabetes

mellitus and diabetic retinopathy. Intern Ophthalmol Clin,

38(2), 11-18.

[2] Chen, M.S., Kao, C.S., Chang, C.J., Wu, T.J., Fu, C.C.

and Chen, C.J. (1992) Prevalence and risk factors of dia-

betic retinopathy among non-insulin-dependent diabetic

subjects. Am J Op h t h a lm o l, 114(6), 723-730.

[3] Tung, T.H., Chen, S.J., Liu, J.H., et al. (2005) A commu-

nity-based follow-up study on diabetic retinopathy

among type 2 diabetics in Kinmen. Eur J Epidemiol,

20(4), 317-323.

[4] Tung, T.H., Chen, S.J., Shih, H.C., et al. (2006) Assess-

ing the natural course of diabetic retinopathy: A popula-

tion-based study in Kinmen. Taiwan, Ophthalmic Epide-

miol, 13(5), 327-333.

[5] Bala, M.V., Mauskopf, J.A. and Wood, L.L. (1999) Will-

ingness to pay as a measure of health benefits. Pharma-

coeconomics, 15(1), 9-18.

C. C. Peng et al. / HEALTH 1 (2009) 304-312

312

[6] Jampel, H.D., Schwartz, G.F., Robin, A.L., Abrams, D.A.,

Johnson, E. and Miller, R.B. (2003) Patient preferences

for eye drop characteristics: A willingness-to-pay analy-

sis. Arch Ophthalmol, 121(4), 540-546.

[7] Olsen, J.A. and Smith, R.D. (2001) Theory versus prac-

tice: A review of “willingness-to-pay” in health and

health care. Health Econ, 10(1), 39-52.

[8] Shih, H.C., Chou, P., Chen, S.J., et al. (2007) A commu-

nity-based study of the willingness to pay associated with

screening for diabetic retinopathy among type 2 diabetes

in Kinmen, Taiwan. J Epidemio l , 17(6), 186-193.

[9] Chou, P., Liao, M.J., Kuo, H.S., Hsiao, K.J. and Tsai, S.T.

(1994) A population survey on the prevalence of diabetes

in Kin-Hu, Kinmen. Diabetes Care, 17(9), 1055-1058.

[10] World Health Organization (1999) Definition, diagnosis

and classification of diabetes mellitus and its complica-

tions: Report of a WHO consultation. Part 1. Diagnosis

and classification of diabetes mellitus. Geneva, World

Health Organization.

[11] Wilkinson, C.P., Ferris 3rd, F.L., Klei n, R.E., et al. (2003)

Proposed international clinical diabetic retinopathy and

diabetic macular edema disease severity scales. Oph-

thalmology, 110(9), 1677-1682.

[12] García-Serrano, J.L., Serrano-Laborda, D., López-Pozas,

M., Cabello-Aparicio, C. and Castro-Rosales, L. (2007)

The Granada diabetic retinopathy study: Direct screening

of 8,244 patients. I Arch Soc Esp Oftalmol, 82(11),

681-689.

[13] Harvey, J.N., Craney, L., Nagendran, S. and Ng, C.S.

(2006) Towards comprehensive population-based

screening for diabetic retinopathy: Operation of the

North Wales diabetic retinopathy screening programme

using a central patient register and various screening

methods. J Med Screen, 13(2), 87-92.

[14] Ackerman, S.J. (1992) Benefits of preventive programs

in eye care are visible on the bottom line. A new nation-

wide effort to improve eye care for people with diabetes

gets backing from a study on the cost-effectiveness of

screening for retinopathy. Diabetes Care, 15(4), 580-581.

[15] Javitt, J.C. and Aiello, L.P. (1996) Cost-effectiveness of

detecting and treating diabetic retinopathy. Ann Intern

Med, 124(1 Pt 2), 164-169.

[16] Polak, B.C., Crijns, H., Casparie, A.F., et al. (2003)

Cost-effectiveness of glycemic control and ophthal-

mological care in diabetic retinopathy. Health Policy,

64(1), 89-97.

[17] Tung, T.H., Shih, H.C., Chen, S.J., et al. (2008) Eco-

nomic evaluation of screening for diabetic retinopathy

among Chinese type 2 diabetics: A community-based

study in Kinmen, Taiwan. J Epidemiol, 18(5), 225-233.

[18] Health Employer Data and Information Set (1997) Ver-

sion 3.0. Washington, DC: National Committee for Qual-

ity Assurance, 2.

[19] Sloan, F.A., Grossman, D.S. and Lee, P.P. (2009) Effects

of receipt of guideline recommended care on onset of

diabetic retinopathy and its progression. Ophthalmology,

116(8), 1515-1521.

[20] Klein, R., Klein, B.E, Neider, M.W., Hubbard, L.D.,

Meuer, S.M. and Brothers, R.J. (1985) Diabetic reti-

nopathy as detected using ophthalmoscopy, a nonmydri-

atic camera and a standard fundus camera. Ophthalmol-

ogy, 92(4), 485-491.

[21] Ekoe, J.M., Zimmet, P. and Williams, R. (2001) The epi-

demiology of diabetes mellitus: An international perspec-

tive. John Wiley & Sons, Inc.

[22] Byrt, T. (1996) How good is that agreement. Epidemiol-

ogy, 7(5), 561.

[23] Aiello, L.P., Gardner, T.W., King, G.L., et al. (1998) Dia-

betes retinopathy. Diabetes Care, 21(1), 143-156.