Impact of screening and early detection of impaired fasting glucose tolerance and type 2 diabetes in Canada: a Markov model simulation

Figures & data

Figure 1 Markov state-transition model with the seven proposed health states and possible transitions to be performed once for each screening method.

Notes: Whether type 2 diabetes and impaired fasting glucose are diagnosed or undiagnosed determines if the patients receive relevant treatments or interventions and whether associated costs and utilities are applied for each time interval. Dotted circles represent starting states.
Abbreviation: IFG, impaired fasting glucose; T2D, type 2 diabetes mellitus.

Figure 2 Markov model simulation for the impact of screening and early detection of impaired fasting glucose tolerance and type 2 diabetes in Canada.

Notes: Two Markov models are run, one for a no screening strategy and the other for a prediabetes and diabetes screening strategy. Appropriate costs for treatment or lifestyle modification programs are applied for those detected by screening as IFG or T2DM, and those who are clinically diagnosed with T2DM because of diabetic complications. Conventional screening is denoted by “Yes”.
Abbreviations: IFG, impaired fasting glucose; TP, true positive test result; TN, true negative; FN, false negative; FP, false positive; T2DM, type 2 diabetes mellitus.

Table 1 Risk factors for type 2 diabetes1

Risk factor

Age 40 years

First-degree relative with type 2 diabetes

Member of high-risk population (eg, people of Aboriginal, Hispanic, South Asian, Asian, or African descent)

History of IGT or IFG

Presence of complications associated with diabetes

Vascular disease (coronary, cerebrovascular, or peripheral)

History of gestational diabetes mellitus

Hypertension

Dyslipidemia

Overweight

Abdominal obesity

Polycystic ovary syndrome

Acanthosis nigricans

Schizophrenia

Note:

1 Risk factors were generated by the Canadian Diabetes Association 2008 Clinical Practice Guidelines for the Prevention and Management of Diabetes in Canada.⁵

Abbreviations: IGT, impaired glucose tolerance; IFG, impaired fasting glucose.

Table 2 Base-case input values for the Markov model1

Parameter	Base-case value
Demographic variables
Age at start of screening program (years)	40
Prevalence of undiagnosed T2DM	5% (1.4%–50%)^Citation3,Citation12
Prevalence of undetected IFG	12%^Citation3
Transition rates
Normal to IFG (per 100 person-years)	1.66 (0.08)^Citation15,Citation18
IFG to T2DM (per 100 person-years)	1.96 (0.25)^Citation19
Test efficiency
We assumed a sensitivity and specificity of almost 100% for two consecutive tests for T2DM and for IFG (for T2DM, sensitivity 83%, specificity 91%; for IFG only, sensitivity 53%, specificity 63%)^Citation12,Citation13,Citation20,Citation21
Costs $ (per person)2
Screening test	30
Diagnostic test	45
Average annual direct medical costs (for diabetes with complications)	5687 (5371–6343)^Citation16
Average annual direct medical costs (for diabetes without complications)	2210^Citation16
Average annual lifestyle modification costs	500^Citation14,Citation22,Citation23
Utilities
Utility of undetected T2DM	0.79 (0.02)^Citation3
Utility of screen detected T2DM	0.79 (0.02)^Citation3
Utility of T2DM with complications	0.77 (0.03)^Citation24

Note:

1 Costs were reported in Canadian dollars from year 2010 and were standardized by inflation indices.

2 Adjusted according to the laboratory and physician fee schedules.¹⁶

Abbreviations: T2DM, type 2 diabetes mellitus; IFG, impaired fasting glucose.

Figure 3 Cost-effectiveness comparison between conventional screening for type 2 diabetes mellitus and no screening.

Notes: Conventional screening strategy (square) is compared with no screening (triangle). The dominated strategy is highlighted by a circle (ie, conventional screening).

Table 3 Sensitivity analysis and variation in frequency of screening

Screening strategy	Cost ($)	Incremental cost ($)	Effect (QALYs)	Incremental effectiveness (QALYs)	Cost/effect ($/QALYs)	Incremental C/E (ICER)
Once every 3 years (basecase)
Yes1	16,800		7.387		2281	16,800
No	20,500	3600	7.081	−0.306	2890	(Dominated)
Once every 5 years
Yes1	15,653		7.40		2116
No	20,466	4812	7.08	−0.31	2890	(Dominated)
Once every year
Yes1	17,500		7.379		2367
No	20,500	3000	7.081	−0.298	2890	(Dominated)

Note:

1 “Yes” represents screening at stated frequency for normal subjects and once every year for subjects with prediabetes.

Abbreviations: QALYs, quality-adjusted life-years; C/E, cost/effectiveness; ICER, incremental cost-effectiveness ratio.

Figure 4 Cost-effectiveness acceptability curve for conventional screening for type 2 diabetes compared with “no screening” strategy.

Abbreviation: WTP, willingness to pay.

Canadian Diabetes Association2008 Clinical Practice Guidelines for the Prevention and Management of Diabetes in CanadaCanad J Diab200832 Suppl 1S14S17

 Google Scholar

GilliesCLLambertPCAbramsKRDifferent strategies for screening and prevention of type 2 diabetes in adults: cost effectiveness analysisBr Med J200833676541180118518426840

 PubMed Web of Science ®Google Scholar

WilsonSERosellaLCLipscombeLLManuelDGThe effectiveness and efficiency of diabetes screening in Ontario, Canada: a population-based cohort studyBMC Public Health201010150620727141

 PubMed Web of Science ®Google Scholar

RasmussenSGlumerCSandbaekALauritzenTBorch-JohnsenKProgression from impaired fasting glucose and impaired glucose tolerance to diabetes in a high-risk screening programme in general practice: the ADDITION study, DenmarkDiabetologia200650229329717143605

 PubMed Web of Science ®Google Scholar

MeigsJMullerDNathanDBlakeDAndresRBaltimore Longitudinal StudyThe natural history of progression from normal glucose tolerance to type 2 diabetes in the Baltimore longitudinal study of agingDiabetes20035261475148412765960

 PubMed Web of Science ®Google Scholar

BonoraEKiechlSWilleitJOberhollenzerFEggerGPopulation-based incidence rates and risk factors for type 2 diabetes in white individuals: the Bruneck studyDiabetes20045371782178915220202

 PubMed Web of Science ®Google Scholar

HuxJEIvisFFlintoftVBicaADiabetes in Ontario: Determination of prevalence and incidence using a validated administrative data algorithmDiabetes Care200125351251611874939

 PubMed Web of Science ®Google Scholar

DaviesMTringhamJJarvisJSkinnerTCFarooqiAMKhuntiKSystematic screening for type 2 diabetes mellitus: results of a large population based study targeting those with conventional risk factorsDiabet Med200522 Suppl 299

 Google Scholar

ZhangYEngelgoanMValdezREfficient cutoff points for three screening tests for detecting undiagnosed diabetes and pre-diabetesDiabetes Care20052861321132515920046

 PubMed Web of Science ®Google Scholar

O’ReillyDHopkinsRBlackhouseGHuxJDolovichLDevelopment of an Ontario Diabetes Economic Model (ODEM) and Application to a Multidisciplinary Primary Care Diabetes Management ProgramReport Prepared for the Ontario Ministry of Health and Long-term Care2006

 Google Scholar

PalmerAJRozeSValentineWJSpinasGAShawJEZimmetPZIntensive lifestyle changes or metformin in patients with impaired glucose tolerance: modeling the long-term health economic implications of the diabetes prevention program in Australia, France, Germany, Switzerland, and the United KingdomClin Ther200426230432115038953

 PubMed Web of Science ®Google Scholar

AvenellABroomJBrownTSystematic review of the long-term effects and economic consequences of treatments for obesity and implications for health improvementHealth Technol Assess2004821iiiiv1182

 PubMed Web of Science ®Google Scholar

HernanWHBrandleMZhangPDiabetes Prevention Program Research GroupCosts associated with the primary prevention of type 2 diabetes mellitus in the diabetes prevention programDiabetes Care2003261364712502656

 PubMed Web of Science ®Google Scholar

ClarkePGrayAHolmanREstimating utility values for health states of type 2 diabetic patients using the EQ-5D (UKPDS 62)Med Decis Making200222434034912150599

 PubMed Web of Science ®Google Scholar