This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Cost-effectiveness of DiabetesPrevention InterventionsTargeting High-risk Individualsand Whole Populations: ASystematic ReviewDiabetes Care 2020;43:1593–1616 | https://doi.org/10.2337/dci20-0018
OBJECTIVE
We conducted a systematic review of studies evaluating the cost-effectiveness (CE)of interventions to prevent type 2 diabetes (T2D) among high-risk individuals andwhole populations.
RESEARCH DESIGN AND METHODS
Interventions targeting high-risk individuals are those that identify people at highrisk of developing T2D and then treat them with either lifestyle or metformininterventions. Population-based prevention strategies are those that focus on thewhole population regardless of the level of risk, creating public health impactthrough policy implementation, campaigns, and other environmental strategies.We systematically searched seven electronic databases for studies published inEnglish between 2008 and 2017.We grouped lifestyle interventions targeting high-risk individuals by delivery method and personnel type. We used the medianincremental cost-effectiveness ratio (ICER), measured in cost per quality-adjustedlife year (QALY) or cost saved to measure the CE of interventions. We used the$50,000/QALY threshold to determine whether an intervention was cost-effectiveor not. ICERs are reported in 2017 U.S. dollars.
RESULTS
Our review included 39 studies: 28 on interventions targeting high-risk individualsand 11 targeting whole populations. Both lifestyle and metformin interventions inhigh-risk individuals were cost-effective from a health care system or a societalperspective, with median ICERs of $12,510/QALY and $17,089/QALY, respectively,compared with no intervention. Among lifestyle interventions, those thatfollowed a Diabetes Prevention Program (DPP) curriculum had a median ICER of$6,212/QALY,while those that did not followaDPP curriculumhadamedian ICERof$13,228/QALY. Compared with lifestyle interventions delivered one-on-one orbyahealthprofessional, thoseoffered inagroup settingorprovidedbya combination ofhealth professionals and lay health workers had lower ICERs. Among population-based interventions, taxing sugar-sweetened beverages was cost-saving from boththe health care system and governmental perspectives. Evaluations of otherpopulation-based interventionsdincluding fruit and vegetable subsidies, commu-nity-based education programs, and modifications to the built environmentdshowed inconsistent results.
CONCLUSIONS
Most of the T2D prevention interventions included in our review were found to beeither cost-effective or cost-saving. Our findings may help decision makers setpriorities and allocate resources for T2D prevention in real-world settings.
1Division of Diabetes Translation, Centers forDisease Control and Prevention, Atlanta, GA2College of Nursing and Disability, Aging andTechnology Cluster, University of Central Florida,Orlando, FL3Oak Ridge Institute for Science and Education,Oak Ridge, TN
The findings and conclusions are those of theauthors and do not necessarily represent theofficial position of the Centers for Disease Controland Prevention.
Diabetes is amajor global health issue. In2019, there were an estimated 463 mil-lion adults aged 20–79 years with di-abetes globally (;9.3%of the populationin this age-group), a figure that is pro-jected to increase to 700 million by2045 (1). Health care expenditures at-tributable to diabetes were estimatedat $1.3 trillion in 2015 (2). Fortunately,type 2diabetes (T2D),which accounts for90–95% of the disease burden (3), can beprevented or delayed through nutritionand lifestyle changes as well as throughpharmacologic interventions (4).Approaches to prevent T2D fall under
two categories: targeting individuals athigh risk for developing T2D (high-riskapproaches) and targeting the wholepopulation regardless of the level of risk(population-based approaches). In gen-eral, high-risk individuals are those whohave prediabetes (a health conditionwith a blood glucose level that is higherthan normal but does not reach the levelofdiagnosedT2D)orwhohave risk factorsfordevelopingT2D,suchashavingafamilyhistoryofT2D, beingoverweightorobese,being physically inactive, being 45 yearsold or older, or being a woman with ahistory of gestational diabetes mellitus(5). Interventions targeting high-risk in-dividuals include screening for T2D inclinics and communities and providinglifestyle or pharmacologic interventions.On the other hand, population-based ap-proaches aim to impact public healththrough policy implementation, cam-paigns, and other environmental changestrategies. For example, imposing taxeson sugar-sweetenedbeverages (SSBs) hasbeen proposed as a population-basedapproach to combat T2D and cardiovas-cular disease by the World Health Orga-nization (6). Epidemiological evidence onthe association between added sugarsand T2D incidence and implementationexperiences from Mexico and selectedcities in the U.S. (Berkeley, for example)have led decision makers to explore thefeasibility and effectiveness of scaling upsuch policies (7,8). Some experts suggestthat the goal of reducing the number ofnew cases of T2D in the U.S. and world-wide is likely best achieved through ap-proaches that combine both high-risk andpopulation-based approaches (9,10).T2D prevention approaches, whether
high-risk or population-based, vary interms of intervention effectiveness andcost. However, their cost-effectiveness
(CE) has not been evaluated comprehen-sively or systematically. Most literaturereviews to date have assessed the effi-cacy of T2D prevention approaches onlywithout considering their CE or withfocus on a single strategy (11–15). Forexample, one review and meta-analysisfocused on nutrition education and ex-amined the cost and CE of using dietmodification as a T2D preventive inter-vention (16). Another systematic reviewmeasured the CE of T2D high-risk pre-vention approaches but focused on life-style interventions only (17). A recentstudy reviewed the CE of both lifestyleandmetformin for T2D prevention amonghigh-risk individuals but did not includepopulation-based approaches (18). An-other review evaluated both high-riskand population-based approaches (19);however, it did not examine key featuresof lifestyle interventionsdsuch as inter-vention delivery mode and formatdthat might affect the CE outcome, and itonly included fiscal policies among thepopulation-based approaches. In addition,many new studies on the CE of T2Dprevention interventions that have beenpublished in recent years need to beevaluated in a review.Here, we systematically review the CE
of both high-risk and population-basedapproaches for T2D prevention. The goalis twofold: 1) to update evidence onhigh-risk approaches implemented inreal-world settings, including whetherto screen, whom to screen, and whichformats are best for delivering lifestyleinterventions (in-person vs. virtual, one-on-one vs. group, etc.) and 2) to synthesizeevidence on population-based preventionstrategies.
RESEARCH DESIGN AND METHODS
Literature SearchWe searched the Cumulative Index toNursing and Allied Health Literature(CINAHL), Cochrane databases, ExcerptaMedica (EMBASE), Medical LiteratureAnalysis and Retrieval System Online(MedlinePlus), PsycINFO, Scopus, andSociological Abstracts (Soc Abs) to iden-tify original economic evaluations ofapproaches to prevent T2D publishedin English from January 2008 to July2017. Search keywords included 1) di-abetes, impaired glucose tolerance, andinsulin resistance; 2) expenditure, healthcare cost, and cost of illness; 3) quality-
adjusted life year (QALY), disability-adjusted life year (DALY), and incidence ofdiabetes; and4) cost-effectiveness analysis(CEA), cost-utility analysis, cost-benefitanalysis, and economic evaluation (20).In addition to searching the seven data-bases above, we manually screened thereference lists of all included studies aswell as the table of contents of majordiabetes journals (Diabetes Care, TheLancet Diabetes & Endocrinology, Diabe-tologia, and Diabetes Research and Clin-ical Practice) during the search period.
Study Design for ReviewingInterventions Targeting High-riskIndividualsFollowing the Cochrane Collaboration’sprotocol for systematic reviews (21), twopeople independently reviewed eachstudy for inclusion/exclusion in our re-view, quality assessment, and data ab-straction. We focused on three types ofeconomic evaluations of high-risk ap-proaches to T2D prevention: CE, cost-utility, and cost-benefit analyses. Weincluded studies that reported quanti-tative measures for the CE outcomes.The outcome was the incremental cost-effectiveness ratio (ICER), which is in theform of cost-per-additional QALY gained orcost-per-additional DALY averted.
Quality Assessment of the IncludedStudiesTo assess the quality of included studies,weuseda tool basedonTheBMJauthors’guide for economic studies (22), whichwas used previously (20). In brief, thetool assesses each study based on 13 at-tributes: sources of cost data, sourcesof benefit data, categories of cost data,categoriesof benefit data, analytical timehorizons, study perspectives, model de-scriptions, structure diagrams, currencyand year of the costs, discounting factorfor costs, discounting factor for benefits,ICERs, and sensitivity analyses. Each at-tribute was given one pointdan equalweightdif the study clearly stated it.Weincluded studies with a quality score ofseven and above (20).
Data Abstraction and Cost AdjustmentWe abstracted the following informationfrom each study: publication informa-tion, study objective, prevention ap-proach, comparison, target population,delivery method, provider, analyticaltime horizon, study method, perspective
1594 Systematic Review of CE Analysis of T2D Prevention Diabetes Care Volume 43, July 2020
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]
of the evaluation, and results. We ad-justed ICERs and costs to 2017 U.S. dol-lars using the Consumer Price Index (23).For studies conducted in countries otherthan the U.S., we used the annual ex-change rate from the Federal ReserveBank to convert the foreign currenciesinto U.S. dollars before adjusting themfor inflation (24). In rare cases where thestudy did not report the specific year ofcurrency used to calculate costs, weassumed the costs were calculated1 year before the publication date.Studies were considered cost-effectiveif the ICER was below the $50,000/QALYthreshold (25).
GroupingHigh-riskApproaches toT2DPreventionWe grouped high-risk approaches intofour categories based on their studyobjectives: 1) articles focused on decid-ing whether to screen for prediabetes, 2)articles determining the target popula-tion for screening that would generatethe optimal CE outcomes, 3) articlesevaluating the CE of specific T2D pre-vention interventions, and 4) articlesevaluating the CE of managing gesta-tional diabetes mellitus.To better understand what features
contribute to the CE outcomes of pre-vention interventions, we examined in-terventions from the third categoryabove (those evaluating theCEof specificT2D prevention interventions) and sum-marized the median and range of ICERsfor interventions sharing similar featuresin terms of how the intervention is de-livered (i.e., whether delivered one-on-oneor in a group andwhether conductedin-person or via virtual media, such asinternet or mobile applications) and bywhom (i.e., whether taught by healthcareprovidersor layhealthworkers, suchas trained community health workers ordiabetes educators). The high-risk ap-proaches included lifestyle interventions(translational Diabetes Prevention Pro-gram [DPP] and translational non-DPP)and pharmacologic interventions (met-formin). Translational DPPs refer to nu-trition and physical activity interventionsthat follow the DPP curriculum thattranslated to the real world, such asthose provided in the community orprimary care setting. In contrast, trans-lational non-DPPs are lifestyle interven-tions that do not strictly follow the DPPcurriculum.
Study Design for Population-BasedInterventionsWe modified our study protocol to ac-commodate the methods and resultsreported in studies on population-basedapproaches because many of them didnot use the standard framework for as-sessing CE due to a lack of data. For studyscreening, we included population-basedinterventions if they reported ICERs orif they compared costs given a certainlevel of benefits if benefits were mea-sured as T2D cases prevented or QALYdue to reduction in diabetes. Conse-quently, the result of cost-saving (CS)for population-based interventions shouldbe interpreted with caution, as it couldrefer to a reduction in health costs onlyrather than savings as measured by ICER,which is a negative incremental cost.Quality assessment for population-
based approaches was less restrictiveand reduced to nine scoring attributes(the other four pertained to formalCEA and did not apply in these cases):sources of cost data, sources of bene-fit data, categories of cost data, cate-gories of benefit data, analytical timehorizons, study perspectives, model de-scriptions, currency and year of the costs,and results. Again, we included stud-ies with a quality score of seven andabove.For selected studies, we abstracted
data on publication information, objec-tive, prevention strategy, comparison,target population, analytic time horizon,study method, the perspective of theevaluation, and results.We thengroupedpopulation-based approaches into fourcategories and summarized CE of eachone: 1) implementing fiscal policy, 2)implementing a regulation, 3) promot-ing health by education and informa-tion, and 4) changing the built or foodenvironment.
RESULTS
Figure 1 shows the 39 studies that metour inclusion criteria: 28 articles on high-risk approaches and 11 articles on pop-ulation-based approaches.
High-risk ApproachesTable 1A shows studies arranged chro-nologically and then alphabetically bythe last name of the first author withineach category (26–53). Among thesestudies, the analytic time horizon rangedfrom 1 year to a lifetime. Studies were
evaluated from either a societal perspec-tive or a health care perspective. Moststudies discounted costs and benefitsat 3%. While most of the studies werebased on simulation modeling, eightstudies assessed prevention strategiesusing randomized controlled trials. Re-sults indicate that screening for pre-diabetes and providing interventions,either lifestyle or pharmacologic inter-ventions, is either cost-effective orCS among individuals with a high riskof T2D. The conclusion held for boththe societal and health care perspec-tives and for shorter or longer timehorizons.Seventeen studies (or study arms)
evaluated the CE of specific interven-tions compared with no intervention(status quo or placebo) from a healthcare system perspective (Table 2) (35–37,39,40,42–44,46,49–51). Results indicatethat all interventions were cost-effective,but the magnitude of the ICERs differedby intervention features. Lifestyle in-terventions were more cost-effectivethan metformin interventions, regard-less of analytical time horizon, deliverymethod, media, mode, and personneltype. Among lifestyle interventions,translational DPPwasmore cost-effectivethan translational non-DPP preventionapproaches. The median ICER for trans-lationalnon-DPPwas twiceashighas thatfor translational DPP. Analytical timehorizon also affects CE outcomes; studiesevaluated over a longer time horizonhave a lower ICER. Among lifestyle in-terventions, in-person interventions hadslightly better CE outcomes than virtualinterventions. The median ICERs for in-terventions delivered in groups and forinterventions provided by a combinationof health professionals and trained layhealth workers were less than half ofthose for the one-on-one interventionsor interventions provided by health pro-fessionals alone.
Population-Based ApproachesTable 1B describes the 11 studies thatevaluated 28 population-based approachesto preventing T2D (54–64). Some stud-ies appear in more than one categorybecause they evaluated multiple inter-ventions that were applied to differentcategories. All studieswere evaluated at10 years or longer. More than half ofthese studies (or study arms) assessed theCE of two fiscal policiesdSSB taxation
care.diabetesjournals.org Zhou and Associates 1595
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]
and fruit and vegetable subsidies. Amongthe nine studies (or study arms) thatevaluated the CE of SSB taxation, themost common taxation rate was 20% ofthe total amount paid. All nine studiesused computer-simulation models andused effectiveness outcomes from pub-lished articles. Two studies used a gov-ernmental perspective while the otherseven used a health care perspective. Allnine studies found the SSB tax to be CS.The included studies also evaluated asugar tax, a fruit and vegetable subsidy,and a combination of taxing unhealthyfoods and subsidizing healthy foods andfound large variations in CE outcomes.For nonfiscal policy interventions, suchas a walking group in the community,opening supermarkets to increase foodaccess, and increasing healthy food op-tions in the workplace, most of theinterventions were cost-effective or CSfrom the health care system perspective.However, the CE results were inconsis-tent from the societal perspective. Inaddition, many of these interventionswere only evaluated by one study,such that we were unable to make adefinite conclusion on the CE of theseinterventions.Table 3 summarizes the CE of population-
based approaches. The SSB tax was found
to beCS from thehealth care systemandgovernmental perspectives. The fourstudies (or study arms) that evaluatedthe CE of subsidies for fruits and veg-etables found mixed results, from morecostly with no net health outcomesbenefits to CS. Similarly, the five studiesthat evaluated community-wide inter-ventions also found them to have var-ious CE outcomes from the health caresystem and societal perspective. Inter-ventions of incentive programs and en-vironmental change were cost-effectivefrom the health care system perspective.
CONCLUSIONS
Our systematic review assessed the CE ofapproaches for preventing T2D from39 studies. Three key findings emerged.First, the ICERs of most of the high-riskapproaches were well below the rangethat is generally considered to be cost-effective. Importantly, differences be-tween delivery methods were small,and the group-delivered translationalDPP provided by a combination of healthprofessionals and trained lay healthworkers seemed most cost-effective.Our findings reinforced the fact that in-terventions to prevent T2D among high-risk individuals are highly cost-effectiveand practical in any given setting. Second,
implementing a population-wide SSBtax was CS and has the potential tobenefit a large population. SSB taxa-tion can be considered as an importantpopulation-based policy approach toprevent T2D globally. Third, althoughthere weremany proposed population-based interventions (including subsi-dies for fruits and vegetables, healthpromotion approaches, and environ-mental changes), the CE of these in-terventions needs further investigationwith real-world data in order to draw aconclusion.Our findings for high-risk approaches
are consistent with previous literature inthat lifestyle programsutilizing the trans-lational DPP curriculum are somewhatmore cost-effective than lifestyle inter-ventions that do not follow the DPPcurriculum (17). The translational DPPlifestyle program is widely used in theCenters for Disease Control and Preven-tion (CDC)-led National Diabetes Preven-tion Program (National DPP)da U.S.translational program providing a frame-work and infrastructure for targetinghigh-risk individuals, and this programis covered by several commercial andpublic insurers (65). For example, theCenters for Medicare & Medicaid Serv-ices began covering the CDC-recognizedDPP lifestyle changeprograms in2018 forMedicare beneficiaries (66).
A noteworthy change in the high-riskapproach category is the adoption ofvirtual media for intervention delivery.In recent years, virtual media interven-tions have become available via onlinecounseling calls, emails, and text mes-sages (29,30,39,49). One benefit of vir-tual media is that it reaches individualswho have barriers to in-person interven-tions, such as the elderly and peoplewholive in rural areas. People may takeadvantage of virtual media interventionsto save time and travel expenses (67).Virtual media interventions also allowparticipants to access the program anytime and with a greater frequency (68).Our review found that few studies eval-uated the CE of interventions deliveredvirtually. The results from this limitedevidence show that virtually deliveredprograms were cost-effective but not ascost-effective as the in-person lifestyleprogram as measured by cost per QALY.Additionally, more rigorous studies areneeded to assess the CE of virtuallydelivered programs.
Figure 1—Summary of evidence search and selection for T2D prevention approaches.
1596 Systematic Review of CE Analysis of T2D Prevention Diabetes Care Volume 43, July 2020
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]
Table
1—Des
criptionoftheCEstudiesforhigh-riskandpopulation-b
ase
dT2D
preve
ntionappro
ach
es
A:Interven
tionstargetinghigh-riskindividuals(high-riskap
proaches)
Study
Interven
tion
Target
population
Duration/
analytical
timehorizon
Comparison
Interven
tion
med
iaStudy
method
Discount
rate:cost/
ben
efit
Effectiven
ess
outcomes
Perspective
ICER
,$/QALY
(in20
17US$)
Whether
toscreen
forprediabetes
Colagiuriand
Walker,
2008
/Australia
(27)
Screen
ingfor
prediabetes
and
providelifestyle
interven
tionfor
those
withIGTorIFG
Individualsaged
55–74
yearsandindividuals
aged
45–54
years
withtherisk
ofT2D
10years/10
years
Nointerven
tion
In-person
Simulation
model
3%/0
Red
ucedT2D
inciden
ceby
15%
Healthcare
$56,484/DALY
averted
Gillieset
al.,
2008
/U.K.
(28)
Screen
ingforIGTand
providelifestyle
interven
tionfor
those
screen
edpositive
Individualsat
risk
for
T2D(atleastone:
family
history
of
diabetes,
hypertension,
dyslipidem
ia,CVD,
orBMI.25
)
1year/50
years
Nointerven
tion
In-person
Simulation
model
3.5%
/3.5%
Increased0.17
yearsspen
tdiabetes-free
per
person
Societal
$16,269
Gillieset
al.,
2008
/U.K.
(28)
Screen
ingforIGTand
providemetform
inforthose
screen
edpositive
Individualsat
risk
for
T2D(atleastone:
family
history
of
diabetes,
hypertension,
dyslipidem
ia,CVD,
orBMI.25
)
1year/50
years
Nointerven
tion
In-person
Simulation
model
3.5%
/3.5%
Increased0.11
yearsspen
tdiabetes-free
per
person
Societal
$18,304
Chatterjee
etal.,
2010
/U.S.
(26)
Screen
ingfor
prediabetes,provide
lifestyle
interven
tion
usingDPP
curriculum
forthose
withIGTor
IFG
Individualswithou
tdiabetes:average
age48
years,
averageBMI30
3years/3
years
Nointerven
tion
In-person
Trial
NR
NR
Healthcare
andsocietal
CSfrom
ahealthcare
perspective,cost-
neu
tral
from
asocietal
perspective
Schaufler
and
Wolff,20
10/
Germany(31)
Screen
ingfor
prediabetes,provide
lifestyle
interven
tion
usingDPP
curriculum
forthose
withIGTor
IFG
Individualsaged
35–75
years
3years/
lifetim
eNointerven
tion
In-person
Simulation
model
5%/0
Lived0.8years
longerafter
diagnosis
Healthcare
$998
Schaufler
and
Wolff,20
10/
Germany(31)
Screen
ingfor
prediabetes,provide
metform
inforthose
withIGTorIFG
Individualsaged
35–75
years
3years/
lifetim
eNointerven
tion
In-person
Simulation
model
5%/0
NR
Healthcare
$578
Con
tinu
edon
p.15
98
care.diabetesjournals.org Zhou and Associates 1597
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]
1612 Systematic Review of CE Analysis of T2D Prevention Diabetes Care Volume 43, July 2020
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]
The results of our review also dem-onstrate great potential for population-based interventions to prevent T2D (69).Among fiscal policies, taxing SSBs maybe a better approach than subsidies forhealthy foods for two main reasons: 1)tax policies generated better CE out-comes and 2) evidence supporting taxpolicies was stronger as multiple studiescollectively reached a consistent conclu-sion. From a health care system perspec-tive, SSB taxes would be CS. The SSB taxwould reduceSSBconsumptionat zeroorlittle health intervention costs andwouldalso reduce health care spending. Thenine studies in our review showed howmuch health care costs would be savedfrom SSB taxation. In addition, thesestudies showed that such an interventionwould be CS or cost-effective from thegovernmental perspective. On the otherhand, the ICERs of interventions to pro-mote the consumption of fruits and veg-etables ranged widely. These studiesdiffered in features thatwould change theresults, such as the targeted population(general population vs. participants in the
Supplemental Nutrition Assistance Pro-gram [SNAP]), analytical time horizon(10 years vs. lifetime), and study per-spective (governmental, health care sys-tem, or societal). Although evidenceindicates that an SSB tax could be a CSintervention to prevent T2D, there arepolitical and other considerations thatimpact its implementation in the realworld (70). The uptake of that strategy isdependent on state and local decision-making (70–72).Our study is one of the first to include
articles evaluating the CE of population-based approaches to prevent T2D ina systematic review. The adoption ofpopulation-based approaches could havegreat potential for improving populationhealth. A recent analysis found that only3.1% of U.S. adults without T2D (regard-less of prediabetes status) met T2D riskreduction lifestyle goals in 2007–2012(73), suggesting the need for broaderpublic health efforts to reach the major-ity of the U.S. population for reducingtheir risk of T2D. Individuals at high riskfor T2D could benefit from population-
based prevention efforts in conjunctionwith targeted, high-risk approaches. Forthose who have not been screened forT2D, population-based interventions mayalso slow their progression to T2D andprovide other health benefits from bet-ter nutrition and more physical activity(9).Our findings on the CE of both high-risk
approaches and population-based ap-proaches indicate that investing in T2Dprevention is an efficient use of limitedhealth care and societal resources. Sincethe development of T2D is a result of acombination of multiple risk factors in-cluding genetics, environment, and be-haviors, a combined strategy of both high-risk and population-based approachesmay be the best one to achieve optimaloutcomes of T2D prevention (9,10). In-terventions targeting high-risk individualsare effective and cost-effective amongindividuals at risk for T2D. However, thelow uptake and resource-intensive natureof high-risk approaches limit their applica-tion. In contrast, while population-basedapproaches use “upstream” approachesthat reach a broader population, theirimpact at the individual level is weaker,and the evidence of their effectiveness ismore limited.Based on our review, we suggest two
avenues for the future economic evalu-ation of T2D prevention approaches.The first is to conduct rigorous CEAsusing real-world data on population-based interventions. The studies inthis review generated considerable var-iation in CE, indicating uncertainty aboutthe CE of these interventions. Manystudies are based on simulation model-ing. Although high-quality simulationmodels can generate reliable results,they rely on strong assumptions thatmay or may not be reflected in reality.In contrast, data from empirical studiesdnatural experiments, for exampledaredirectly observed and reflect the “true”behavioral change of the population tointerventions. Although such studies usu-ally last for a couple of years, they areoften the foundation for modeling stud-ies. Additional research that evaluatesthe impact of taxes, subsidies, food label-ing, and other approaches that are al-ready implemented (“natural experiments”)are needed to obtain stronger data. Sec-ond, effectiveandcost-effectivepopulation-based approaches are needed for bothdeveloping and developed countries.
Table 2—Summary of the CE of interventions targeting high-risk individuals forT2D prevention*
Type of lifestyle interventionTranslational DPP‡ 7 $6,212 (CS–$24,368)Translational non-DPP 6 $13,228 (CS–$16,177)
Time horizon of lifestyle intervention,10 years 5 $19,612 ($6,212–$45,358)$10 years 10 $13,779 (CS–$24,368)
Intervention media of lifestyle interventionIn-person 8 $10,956 (CS–$24,368)Virtual 3 $12,510 (CS–$13,944)Combination of both 2 $6,331, $15,122
The delivery setting of in-person lifestyle interventionOne-on-one 5 $15,700 (CS–$24,368)Group 2 CS, $6,212Combination of both 1 $16,177
Provider of in-person lifestyle interventionHealth professionals 5 $15,700 (CS–$24,368)Health professionals and trained lay health workers 3 $6,212 (CS–$16,177)
*Studies included in this table satisfy three conditions: 1) the main objective of a study wasevaluating theCEof an intervention,2) theeffect of the interventionwas comparedwith theeffectof a “status quo” or a placebo scenario, and 3) the evaluation was from a health care systemperspective. †The range of ICER is reported if there are three or more data points. Costs are in2017 U.S. dollars. ‡Translational DPPs refer to diet and physical activity interventions that followthe DPP curriculum that translated to real-world settings, such as provided in the community orprimary care. In contrast, translational non-DPPs are lifestyle interventions that do not strictlyfollow the DPP curriculum.
care.diabetesjournals.org Zhou and Associates 1613
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]
Although we found a disproportionateimbalance in the number of studiespublished involving high-income coun-tries, population-based approaches arestrategies to reach a large scale of thepopulation to address the dramatic in-crease in diabetes prevalence worldwide.Conclusions from this review need to
be interpretedwith caution. First,most ofthe evaluations, especially population-based approaches, utilized simulationmodeling,whichcanbeheavily influencedby assumptions. Unlike data from clinicaltrials, which are directly observed, modeldata are usually from published articles.Even thoughmanymodels useddata fromclinical trials for the initial years of inter-ventions, theymustmakeassumptionsonthe persistence of costs and effectivenessbeyondthe trial studyperiod tosimulate alonger time horizon. Because of thesecontraints, in our review, we tried torely on evidence if it was consistentacross multiple modeling studies. Sec-ond, in order to include as many studieson population-based interventions aspossible, we used somewhat “looser”quality criteria for these studies. Manyof the population-based approaches did
not conduct formal CEA. As a result, theCS results fromthese studiesneeds tobebetter understood, as they were a sim-ple comparison of costs given a certainlevel of health benefit. Also,many of theCE results were estimated from govern-mental or health care system perspec-tives rather than a societal perspective.Third, the societal perspective de-fined in population-based approacheswas not as inclusive as it was for high-risk approaches. Some cost categorieswere not included in the societal per-spective, such as productivity loss ortime cost. Fourth, we compared theCE of interventions based on the me-dian ICERs without explicitly consid-ering other study information, suchas the evaluation method and rigorous-ness of data. This comparison followsprevious literature (17) but may notreflect a real difference in CE. Fifth,our results provide information for de-cision makers to choose among inter-ventions based on CE criteria only.Manyother issues such as health equality,acceptability, and feasibility shouldalso be considered in real-world decision-making.
Evidence from our review indicatesthat investing in T2D prevention, usingeither high-risk approaches or population-based approaches, is an efficient use ofhealth care and societal resources. Giventhe enormous cost associatedwith T2D, ifhealth care resources are limited, thenprevention is a highly efficient use ofsuch resources. Interventions targetinghigh-risk individuals with group-deliv-ered translational DPP lifestyle interven-tion, provided by a combination of healthprofessionals and trained lay health work-ers, was more cost-effective comparedwith one-on-one interventions providedby health professionals solely; however,all interventions targeting high-risk in-dividuals were cost-effective. Amongpopulation-based approaches, the SSBtaxation saves costs and resourcesof thehealth care system and government.Therefore, expansion of insurance-cov-ered, professional, and lay-deliveredgroup DPPs with a simultaneous insti-tution of SSB taxation can be consideredas a priority to stem the rising tide ofT2D. A combined approach that targetsboth high-risk individuals and the wholepopulation could be a policy choice for
Table 3—Summary of the CE of population-based T2D prevention approaches
InterventionStudyarm, n Perspective CE outcome*
Penny-per-ounce, 10–20%, or $0.5/liter tax on SSB 9 Health care andgovernmental
CS
Tax sugar for $0.99/100 mL of ice cream and $0.9/100 gof all other products 1 Health care CS
30% subsidy for the consumption of fruits and vegetables among SNAP beneficiaries 4 Health care,governmental,and societal
CS to worse healthand more cost
A bundled policy of taxing $1.45/100 g of saturated fat, $0.32/1 g of sodium, $0.5/liter of SSB,$0.99/100 mL of ice cream, $0.9/100 g of sugar of all other products, and subsidizing $0.15/100 g of fruits and vegetables 1 Health care CS
Ban on using SNAP dollars for SSB purchases 1 Governmental CS
Community-wide programs for health education (newspaper column, booklet, television news,talks, seminars, workshops, and diet and cooking classes) and physical activity promotion(organized walking events, worksite exercise programs, financial incentives, home visits, andphone calls)
5 Health careand societal
CS to notcost-effective
Themassmediacampaign, including televisionadvertising, advertisements inprintmedia,a toll-free telephone line for community-level support, and marketing of campaign merchandise 1 Health care CS
Targeted incentive program, including distributing tailoredmaps of local walking paths and busschedules, using merchandise as incentive or reward for reducing the use of cars, andencouraging use of pedometers
2 Health care CS tocost-effective
Internet intervention, includinggivingaccess tophysical activity informationandadvice throughwebsite and e-mail 1 Health care Cost-effective
Environmental change, including building bicycle paths, extending hours at recreation facilities,opening fitness centers, increasing the convenient supply of healthy foods, nutritioninformation pamphlets placed on dining tables, color-coded labeling for foods, opening newsupermarkets, and increasing healthy food options in workplace cafeterias
3 Health careand societal
CS tocost-effective
*For studies that did not conduct formal CEA, CS indicates that the intervention reduces cost.
1614 Systematic Review of CE Analysis of T2D Prevention Diabetes Care Volume 43, July 2020
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]
preventing T2D in the U.S. and probablyin other high-income countries.
Acknowledgments. Thiswork is a collaborationbetween the Centers for Disease Control andPrevention and the American Diabetes Associ-ation. The authors thank the external and in-ternal reviewers for their valuable commentsduring the reviewprocess. The authors thank RuiLi (CDC) for generously sharing materials fromher previous review and providing guidance,William Thomas (CDC) for his timely help withthe literature search, and Clarice G. Conley (CDC)for her editorial assistance.Duality of Interest. No potential conflicts ofinterest relevant to this article were reported.Author Contributions. X.Zho. and P.Z. de-signed the research. X.Zho. analyzed data, in-terpreted results, and drafted the manuscript.K.R.S. made a critical revision of the manuscript.X.Zho., K.R.S., B.P.N., S.J., and K.K.P. screenedstudies and abstracted data. B.P.N., X.Zha., A.L.A.,and P.Z. provided important intellectual contentto the manuscript.
References1. International Diabetes Federation. IDF Dia-betes Atlas, 9th ed. Brussels, Belgium, Interna-tional Diabetes Federation, 20192. Bommer C, Heesemann E, Sagalova V, et al.The global economic burden of diabetes in adultsaged 20-79 years: a cost-of-illness study. LancetDiabetes Endocrinol 2017;5:423–4303. Centers for Disease Control Prevention. Na-tional Diabetes Statistic Report, 2020. Atlanta,GA, Centers for Disease Control and Prevention,US Department of Health and Human Services,20204. American Diabetes Association. 3. Preventionor delay of type 2 diabetes: Standards ofMedicalCare in Diabetesd2020. Diabetes Care 2020;43(Suppl. 1):S32–S365. American Diabetes Association. 2. Classifica-tion and diagnosis of diabetes: Standards ofMedical Care in Diabetes–2020. Diabetes Care2020;43(Suppl. 1):S14–S316. Thow AM, Downs SM, Mayes C, Trevena H,Waqanivalu T, Cawley J. Fiscal policy to improvediets and prevent noncommunicable diseases:from recommendations to action. Bull WorldHealth Organ 2018;96:201–2107. Colchero MA, Guerrero-Lopez CM, Molina M,Rivera JA. Beverages sales in Mexico before andafter implementation of a sugar sweetened bev-erage tax. PLoS One 2016;11:e01634638. Falbe J, Thompson HR, Becker CM, Rojas N,McCulloch CE, Madsen KA. Impact of the Berke-ley excise tax on sugar-sweetened beverageconsumption. Am J Public Health 2016;106:1865–18719. Albright AL, Gregg EW. Preventing type 2diabetes in communities across the U.S.: theNational Diabetes Prevention Program. Am JPrev Med 2013;44(Suppl. 4):S346–S35110. Pan American Health Organization. Popula-tion and individual approaches to the preventionand management of diabetes and obesity. 2011.Accessed 11May 2020. Available fromhttps://www.paho.org/hq/dmdocuments/2012/DMPLAN-ENGLISH.pdf
11. Dunkley AJ, Bodicoat DH, Greaves CJ, et al.Diabetes prevention in the real world: effective-ness of pragmatic lifestyle interventions for theprevention of type 2 diabetes and of the impactof adherence to guideline recommendations:a systematic review and meta-analysis. DiabetesCare 2014;37:922–93312. Ali MK, Echouffo-Tcheugui J, WilliamsonDF. How effective were lifestyle interventionsin real-world settings thatweremodeledon theDiabetes Prevention Program?Health Aff (Mill-wood) 2012;31:67–7513. Aziz Z, Absetz P, Oldroyd J, Pronk NP,Oldenburg B. A systematic review of real-worlddiabetes prevention programs: learnings fromthe last 15 years. Implement Sci 2015;10:17214. NeamahHH, Sebert KuhlmannAK, TabakRG.Effectiveness of programmodification strategiesof thediabetespreventionprogram: a systematicreview. Diabetes Educ 2016;42:153–16515. Haw JS, Galaviz KI, Straus AN, et al. Long-term sustainability of diabetes prevention ap-proaches: a systematic review andmeta-analysisof randomized clinical trials. JAMA Intern Med2017;177:1808–181716. SunY,YouW,AlmeidaF,EstabrooksP,DavyB.The effectiveness and cost of lifestyle interven-tions including nutrition education for diabetesprevention: a systematic review and meta-anal-ysis. J Acad Nutr Diet 2017;117:404–421.e3617. Li R, Qu S, Zhang P, et al. Economicevaluation of combined diet and physical ac-tivity promotion programs to prevent type 2diabetes among persons at increased risk:a systematic review for the Community Pre-ventive Services Task Force. Ann Intern Med2015;163:452–46018. Roberts S, Barry E, Craig D, Airoldi M, BevanG, Greenhalgh T. Preventing type 2 diabetes:systematic review of studies of cost-effective-ness of lifestyle programmes and metformin,with and without screening, for pre-diabetes.BMJ Open 2017;7:e01718419. International Diabetes Federation. Cost-Effective Solutions for the Prevention of Type 2Diabetes. Brussels, Belgium, International Diabe-tes Federation, 201620. Li R, Zhang P, Barker LE, Chowdhury FM,Zhang X. Cost-effectiveness of interventions toprevent and control diabetesmellitus: a systematicreview. Diabetes Care 2010;33:1872–189421. Higgins JPT, Thomas J, Chandler J, et al. (Eds.).Cochrane Handbook for Systematic Reviews ofInterventions [Internet], 2019. Version 6.0. Avail-able from https://www.training.cochrane.org/handbook. Accessed 15 October 201922. Drummond MF, Jefferson TO; The BMJ Eco-nomic Evaluation Working Party. Guidelines forauthors and peer reviewers of economic submis-sions to the BMJ. BMJ 1996;313:275–28323. U.S. Bureau of Labor Statistics. Consumerprice index (Medical care in U.S. city average, allurban consumers, not seasonally adjusted). Ac-cessed 28 August 2017. Available from https://www.bls.gov/cpi/data.htm24. Board of Governors of the Dederal ReserveSystem. Foreign exchange rates (G.5A). Accessed28 August 2017. Available from https://www.federalreserve.gov/releases/g5a/25. Braithwaite RS, Meltzer DO, King JT Jr, LeslieD, Roberts MS. What does the value of modernmedicine say about the $50,000 per quality-
adjusted life-year decision rule? Med Care2008;46:349–35626. Chatterjee R, Narayan KM, Lipscomb J,Phillips LS. Screening adults for pre-diabetesand diabetes may be cost-saving. DiabetesCare 2010;33:1484–149027. Colagiuri S, Walker AE. Using an economicmodel of diabetes to evaluate prevention andcare strategies inAustralia.HealthAff (Millwood)2008;27:256–26828. Gillies CL, Lambert PC, Abrams KR, et al.Different strategies for screening and preventionof type 2 diabetes in adults: cost effectivenessanalysis. BMJ 2008;336:1180–118529. Liu X, Li C, Gong H, et al. An economicevaluation for prevention of diabetes mellitusin a developing country: a modelling study. BMCPublic Health 2013;13:72930. Neumann A, Schwarz P, Lindholm L. Esti-mating the cost-effectiveness of lifestyle inter-vention programmes to prevent diabetes basedon an example from Germany: Markov model-ling. Cost Eff Resour Alloc 2011;9:1731. Schaufler TM,WolffM. Cost effectiveness ofpreventive screening programmes for type 2diabetes mellitus in Germany. Appl HealthEcon Health Policy 2010;8:191–20232. Breeze PR, Thomas C, Squires H, et al. Theimpact of type 2 diabetes prevention programmesbased on risk-identification and lifestyle inter-vention intensity strategies: a cost-effective-ness analysis. Diabet Med 2017;34:632–64033. Zhuo X, Zhang P, Selvin E, et al. AlternativeHbA1c cutoffs to identify high-risk adults fordiabetes prevention: a cost-effectiveness per-spective. Am J Prev Med 2012;42:374–38134. Zhuo X, Zhang P, Kahn HS, Gregg EW. Cost-effectivenessofalternativethresholdsofthefastingplasma glucose test to identify the target popula-tion for type 2 diabetes prevention in adultsaged$45years.DiabetesCare2013;36:3992–399835. Diabetes Prevention Program ResearchGroup. The 10-year cost-effectiveness of lifestyleintervention or metformin for diabetes preven-tion: an intent-to-treat analysis of the DPP/DPPOS. Diabetes Care 2012;35:723–73036. Herman WH, Edelstein SL, Ratner RE, et al.;Diabetes Prevention Program Research Group.Effectiveness and cost-effectiveness of diabetesprevention among adherent participants. Am JManag Care 2013;19:194–20237. Hoerger TJ, Crouse WL, Zhuo X, Gregg EW,Albright AL, Zhang P. Medicare’s intensive be-havioral therapy for obesity: an exploratory cost-effectiveness analysis. Am J Prev Med 2015;48:419–42538. Hollenbeak CS, Weinstock RS, Cibula D,Delahanty LM,Trief PM.Cost-effectivenessofSHINE:a telephone translation of the Diabetes PreventionProgram. Health Serv Insights 2016;9:21–2839. Leal J, Ahrabian D, Davies MJ, et al. Cost-effectiveness of a pragmatic structured educa-tion intervention for the prevention of type 2diabetes: economic evaluation of data from theLet’s Prevent Diabetes cluster-randomised con-trolled trial. BMJ Open 2017;7:e01359240. Lin J, Zhuo X, Bardenheier B, et al. Cost-effectiveness of the 2014 US Preventive Serv-ices Task Force (USPSTF) recommendations forintensivebehavioral counseling interventions foradults with cardiovascular risk factors. DiabetesCare 2017;40:640–646
care.diabetesjournals.org Zhou and Associates 1615
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]
41. NeumannA,LindholmL,NorbergM,SchofferO,Klug SJ, Norstrom F. The cost-effectiveness ofinterventions targeting lifestyle change forthe prevention of diabetes in a Swedish primarycare and community based prevention pro-gram. Eur J Health Econ 2017;18:905–91942. Palmer AJ, Tucker DM. Cost and clinicalimplications of diabetes prevention in an Aus-tralian setting: a long-term modeling analysis.Prim Care Diabetes 2012;6:109–12143. Peels DA, Hoogenveen RR, Feenstra TL, et al.Long-term health outcomes and cost-effective-ness of a computer-tailored physical activityintervention among people aged over fifty:mod-elling the results of a randomized controlled trial.BMC Public Health 2014;14:109944. Png ME, Yoong JS-Y. Evaluating the cost-effectiveness of lifestyle modification versusmetformin therapy for the prevention of diabe-tes in Singapore. PLoS One 2014;9:e10722545. Saha S, Carlsson KS, GerdthamU-G, et al. Arelifestyle interventions in primary care cost-effective?–An analysis based on a Markov model,differences-in-differences approach and the Swed-ish Bjorknas study. PLoS One 2013;8:e8067246. Smith KJ, Bryce CL, Hsu HE, et al. Peerreviewed: cost-effectiveness analysis of effortsto reduce risk of type 2 diabetes and cardiovas-cular disease in southwestern pennsylvania,2005-2007. Prev Chronic Dis 2010;7:A10947. van Wier MF, Lakerveld J, Bot SD, ChinapawMJ, Nijpels G, van Tulder MW. Economic eval-uation of a lifestyle intervention in primary careto prevent type 2 diabetes mellitus and cardio-vascular diseases: a randomized controlled trial.BMC Fam Pract 2013;14:4548. Wilson KJ, Brown HS 3rd, Bastida E. Cost-effectiveness of a community-based weightcontrol intervention targeting a low-socioeconomic-statusMexican-originpopulation.HealthPromotPract 2015;16:101–10849. Wong CK, Jiao F-F, Siu S-C, et al. Cost-effectiveness of a short message service inter-vention toprevent type2diabetes from impairedglucose tolerance. J Diabetes Res 2016;121958150. ZhuoX, Zhang P, Gregg EW, et al. A nationwidecommunity-based lifestyle program could delay orprevent type 2 diabetes cases and save $5.7 billionin 25 years. Health Aff (Millwood) 2012;31:50–6051. Feldman I, Hellstrom L, Johansson P. Het-erogeneity in cost-effectiveness of lifestyle coun-seling for metabolic syndrome risk groups
-primary care patients in Sweden.Cost Eff ResourAlloc 2013;11:1952. Kolu P, Raitanen J, Puhkala J, Tuominen P,Husu P, Luoto R. Effectiveness and cost-effectivenessof a cluster-randomized prenatal lifestyle coun-seling trial: a seven-year follow-up. PLoS One2016;11:e016775953. Oostdam N, Bosmans J, Wouters MG,Eekhoff EM, van Mechelen W, van PoppelMN. Cost-effectiveness of an exercise programduring pregnancy to prevent gestational diabe-tes: resultsof aneconomicevaluationalongsidearandomised controlled trial. BMC PregnancyChildbirth 2012;12:6454. Wang YC, Coxson P, Shen Y-M, Goldman L,Bibbins-Domingo K. A penny-per-ounce tax onsugar-sweetened beverages would cut healthand cost burdens of diabetes. Health Aff (Mill-wood) 2012;31:199–20755. Basu S, Seligman H, Bhattacharya J. Nutri-tional policy changes in the supplemental nutri-tion assistance program: a microsimulation andcost-effectiveness analysis. Med Decis Making2013;33:937–94856. Mekonnen TA, Odden MC, Coxson PG, et al.Health benefits of reducing sugar-sweetenedbeverage intake in high risk populations ofCalifornia: results from the cardiovascular disease(CVD) policy model. PLoS One 2013;8:e8172357. Manyema M, Veerman JL, Chola L, TugendhaftA, Labadarios D, HofmanK.Decreasing the burdenof type 2 diabetes in South Africa: the impact oftaxing sugar-sweetened beverages. PLoS One2015;10:e014305058. Sanchez-Romero LM, Penko J, Coxson PG,et al. Projected impactofMexico’s sugar-sweetenedbeverage tax policy on diabetes and cardiovasculardisease: a modeling study. PLoS Med 2016;13:e100215859. Veerman JL, Sacks G, Antonopoulos N,Martin J. The impact of a tax on sugar-sweetenedbeverages on health and health care costs:a modelling study. PLoS One 2016;11:e015146060. Breeze PR, Thomas C, Squires H, et al. Cost-effectiveness of population-based, community,workplace and individual policies for diabetesprevention in the UK. Diabet Med 2017;34:1136–114461. Cobiac LJ, TamK, Veerman L, Blakely T. Taxesand subsidies for improving diet and populationhealth in Australia: a cost-effectiveness model-ling study. PLoS Med 2017;14:e1002232
62. Choi SE, Seligman H, Basu S. Cost effec-tiveness of subsidizing fruit and vegetable pur-chases through the supplemental nutritionassistance program. Am J Prev Med 2017;52:e147–e15563. Roux L, Pratt M, Tengs TO, et al. Cost effec-tiveness of community-based physical activityinterventions. Am J Prev Med 2008;35:578–58864. Cobiac LJ, Vos T, Barendregt JJ. Cost-effectiveness of interventions to promote phys-ical activity: amodelling study. PLoSMed2009;6:e100011065. Ely EK, Gruss SM, Luman ET, et al. A nationaleffort to prevent type 2 diabetes: participant-level evaluation of CDC’s National DiabetesPrevention Program. Diabetes Care 2017;40:1331–134166. Pear R Medicare proposal takes aim at di-abetes. New York Times, 23 March 2016;Sect. A,p. 1267. Moin T, Ertl K, Schneider J, et al. Womenveterans’ experience with a web-based diabetesprevention program: a qualitative study to in-form futurepractice. JMed InternetRes2015;17:e12768. Tate DF, Lyons EJ, Valle CG. High-tech toolsfor exercise motivation: use and role of tech-nologies such as the internet, mobile applica-tions, social media, and video games. DiabetesSpectr 2015;28:45–5469. Jacobson MF, Krieger J, Brownell KD. Poten-tial policy approaches to address diet-relateddiseases. JAMA 2018;320:341–34270. Pomeranz JL, Zellers L, Bare M, PertschukM. State preemption of food and nutritionpolicies and litigation: undermining govern-ment’s role in public health. Am J Prev Med2019;56:47–5771. Pomeranz JL, Pertschuk M. Sugar-sweetenedbeverage taxation in the USA, state preemp-tion of local efforts. Public Health Nutr 2019;22:19072. Hagenaars LL, Jeurissen PP, KlazingaNSJPhn: Sugar-sweetened beverage taxationin 2017: a commentary on the reasons be-hind their quick spread in the EU comparedwith the USA. Public Health Nutr 2019;22:186–18973. Siegel KR, Bullard KM, Imperatore G, et al.Prevalence of major behavioral risk factors fortype 2 diabetes. Diabetes Care 2018;41:1032–1039
1616 Systematic Review of CE Analysis of T2D Prevention Diabetes Care Volume 43, July 2020
For advance media review only. Content embargoed until 3:00 p.m. EST, June 19, 2020. For more information, contact: [email protected]