Gestational diabetes mellitus (GDM) is an increasing cause of morbidity in women and their offspring. Screening and intervention can reduce perinatal and most likely also long-term diabetes consequences. There have been many economic studies, but not recently systematically compared. We conducted a systematic search and abstraction of cost-effectiveness and cost–utility studies from 2002 to 2014. We standardized all findings to 2014 US dollars. We found that cost-effectiveness ratios varied widely. Most variation was found to be due to differences in geographic setting, diagnostic criteria and intervention approaches, and outcomes (e.g., inclusion or exclusion of long-term type 2 diabetes risk and associated costs). We concluded that incorporation of long-term benefits of GDM screening and treatment has huge impact on cost-effectiveness estimates. Based on the large methodological heterogeneity and varying results in the existing body of evidence, we find it unreasonable to outline any global recommendations. For future economic studies, we recommend inclusion of long-term outcomes and adaptation to local preferences, as well as examination of the impact of the diagnostic criteria recently proposed by the International Association of Diabetes in Pregnancy Study Groups (IADPSG).
Introduction
With the rapidly increasing focus on reducing the overwhelming global burden of noncommunicable diseases, there is growing interest in exploring and assessing the impact of “new” links between health conditions in order to identify efficient interventions. One such link is between gestational diabetes (GDM) and the long-term health of the pregnant woman and her offspring. The prevalence of GDM is increasing globally, so identifying and providing good care for women with GDM could have a substantial impact on population health in both high- and low-income countries.
GDM and type 2 diabetes mellitus (T2DM) share risk factors such as obesity and other metabolic disturbances. As the T2DM epidemic expands, GDM will increase in parallel . Current diagnostic and therapeutic recommendations for GDM are founded on the growing body of evidence of the short-term risks associated with GDM. These include for the offspring up to a 3.5-fold higher risk of macrosomia (high birth weight) , elevated risk of shoulder dystocia, and hypoglycemia , and for the women, increased risk of cesarean section, polyhydramnios, gestational hypertension, and preeclampsia . The recently published International Association of Diabetes in Pregnancy Study Groups (IADPSG) diagnostic criteria , which are based on data from a large multicenter and multinational prospective study , are lower than some commonly used diagnostic criteria and therefore lead to a two-to threefold higher prevalence of GDM . The World Health Organization (WHO) very recently approved the IADPSG criteria . The increase in GDM risk factors and the changed diagnostic criteria are both pushing GDM prevalence upwards.
Often less visible are the long-term risks which potentially have a much larger effect on population health. Women with GDM have a more than sevenfold increased risk of T2DM (pooled relative risk (RR) = 7.43) , with cumulative incidence up to 70% in 3 years in some populations . The risk of T2DM is most pronounced for women with high glucose values at the diagnostic oral glucose tolerance test (OGTT) during pregnancy or postpartum continuation of glucose intolerance , greater body mass index (BMI), use of insulin during pregnancy (as a marker of severity of GDM), and earlier gestational age at diagnosis of GDM . The offspring of mothers with hyperglycemia during pregnancy also carry an increased risk of prediabetes and T2DM later in life, four to eight times that of offspring born to mothers with normal glucose tolerance during pregnancy , with a 20-year cumulative incidence of up to 21% .
GDM is increasingly mentioned as a possible contributor to the T2DM epidemic . A modeling study from Canada estimated GDM as a forerunner of as many as 30% of T2DM cases in a high-prevalence diabetes population . In a 2009 report from The Lancet’s Stillbirths Series steering committee, detection and management of GDM were on a priority list of interventions to reduce stillbirth . Lately, the potential key role for gestational diabetes is being discussed at highest political level: the United Nations General Assembly and other UN bodies .
So while the association between GDM and the long-term health of the mother and her offspring is accepted, we are only beginning to understand which interventions might be effective at reducing the burden, and at what cost . To institute appropriate political action, decision makers at national and global levels therefore need guidance from reliable information on the cost and cost-effectiveness of GDM screening and treatment.
Methods
We reviewed cost-effectiveness research identified in a systematic literature search of articles published from 2002 to 2014 ( Table 1 ). The time period was based on an initial literature search showing that formal quantification of the cost-effectiveness of GDM was not done before this period. Furthermore, we excluded older studies since changes in population characteristics, screening, and treatment regimes were considered to have changed significantly from that period.
| Literature sources | Search phrases | Inclusion criteria |
|---|---|---|
| MEDLINE Embase Cochrane NHS-CRD-HTA-database Google Scholar and Google | Gestational diabetes OR hyperglycemia pregnancy AND screening AND Economics OR Cost-effectiveness OR Cost benefit OR Economic evaluation | Full economic evaluations Publication period: 2002–2014 Language: English, German, French, or Scandinavian language |
Using the definition of full economic evaluations by Drummond et al. , we included cost-effectiveness analyses (CEA), cost–utility analyses (CUA), and cost–benefit analyses (CBA). We included only those studies including the following essential elements: costs, consequences, and competing alternatives. Furthermore, we excluded studies not written in English, German, French, or Scandinavian languages.
The included studies were initially scrutinized regarding health economic perspective. As economic evaluation focuses on optimizing utility across all sectors, we regarded “societal” as the preferred perspective. The ideal comprehensive economic evaluation on screening for GDM was seen as a combination of all clinically used screening types, populations, and diagnostic thresholds compared to no screening or other screening approaches or criteria. For alternatives to be comparable, we converted diagnostic thresholds from milligram per deciliter to millimoles per liter by dividing milligrams per deciliter by 18.
We assessed whether identified costs reflected the specified perspective and time horizon. We preferred economic costing, in which donated or subsidized inputs were valued at their market price. We evaluated the validity of the effectiveness values from an epidemiological perspective, though being aware of a trade-off between efficacy and effectiveness when applying results from an experimental to a clinical setting. Both costs and consequences can, based on the idea of time preference, be discounted, and this is generally preferred .
Average or total costs do not reflect the trade-off in marginal cost and health gain, and we therefore focused on the cost of producing an extra health unit. Thus, incremental analysis was considered essential when estimating cost-effectiveness. There exists no consensus on threshold for willingness to pay , so estimates of cost-effectiveness incorporating willingness to pay were regarded as controversial . The included studies generally did not display stochastic data on the estimates of cost-effectiveness. As the screening alternatives examined varied widely between studies, we synthesized the estimates of cost-effectiveness qualitatively by focusing on the intensity of the included screening alternatives in each evaluation. The incremental cost-effectiveness ratio (ICER) represents the ratio of incremental costs to incremental benefits of screening. Results from studies not performing incremental analyses were excluded in the synthesis of cost-effectiveness estimates. To make study results comparable, we inflated all incremental cost estimates into 2014 US dollars using data from the United States Bureau of Labor Statistics and OANDA historical exchange rates .
Results
Characteristics of included studies
We screened 100 studies identified through the literature search. We excluded 19 studies, where the topic was not on screening for GDM, and then 69 studies that could not be defined as full economic evaluations. On this basis, we included 12 full economic evaluations ( Table 2 ) . However, one of these was a brief report, whilst two of these where only abstracts . Three studies could readily be classified as cost analyses, and thereby as partial economic evaluations. As it can be argued that found cases of GDM express the effectiveness of the screening programs, we chose to consider them as CEAs. One further study was classified as CEA . The remaining eight studies used quality-adjusted life years (QALYs) or disability-adjusted life-years (DALYs) and were therefore classified as CUAs. No CBAs were identified. Only four authors used a societal perspective, so only these evaluations were able to incorporate all potential social utility or disutility of the screening into their results. In three evaluations , it was not possible to identify the perspective. It was therefore difficult to judge the scope of these evaluations.
| Study | Evaluation type (consequences) | Perspective |
|---|---|---|
| Poncet B, Touzet S, Rocher L et al. | CEA (found cases of GDM, macrosomia, prematurity, hypertensive disorders, and perinatal mortality) | NS |
| Larijani B, Hossein-nezhad A, Rizvi SW et al. | CEA (found cases of GDM) | NS |
| Nicholson WK, Fleisher LA, Fox HE, Powe NR | CUA (QALY maternal and neonatal) | Societal |
| Ayach W, Costa RA, Calderon Ide M, Rudge MV | CEA (found cases of GDM) | NS |
| Thung S, Pettker C, Funai E | CUA (QALY) | Health-care sector |
| National Collaborating Centre for Women’s and Children’s Health (UK) | CUA (QALY) | Health-care sector |
| Lee S, Pettker C, Funai E et al. | CUA (QALY) | Societal |
| Meltzer SJ, Snyder J, Penrod JR et al. | CEA (found cases of GDM) | Not stated |
| Round JA, Jacklin P, Fraser RB et al. | CUA (QALY) | Health-care sector |
| Mission JF, Ohno MS, Cheng YW, Caughey AB | CUA (QALY) | Societal |
| Werner EF, Pettker CM, Zuckerwise L et al. | CUA (QALY) | Health-care sector |
| Marseille E, Lohse N, Jiwani A et al. | CUA (DALY) | Societal |
Two evaluations approached being comprehensive with respect to inclusion of screening alternatives. Thus, these evaluations included both one- and two-step models of several test types while combining different risk factors for GDM. However, these evaluations did not include 3h 100-g OGTT as an alternative, just as the tests were not examined using different diagnostic criteria. Five evaluations did not include a no-screening alternative. We found large variation in the stated test criteria ( Table 3 ), also the definition of risk factors varied to some degree ( Table 4 ).
| FBG or RBG 0.a: BG ≥5.0 (0h) 0.b: BG ≥7.0 (0h) 0.d: BG ≤5.1: further testing, 5.1–6.9: GDM, ≥7.0: manifest DM (0h) (IADPSG criteria) | 1h 50-g GCT 1.a: BG >7.2 (1h) 1.b: BG >7.8 (1h) 1.c: BG >10.3 (1h) |
| 2h 75-g OGTT 2.a: BG >5.5 or 8 (0/2h) 2.b: BG >8.0 (2h) 2.c: BG >6.0 (2h) 2.d: BG >7.0 (2h) 2.e: BG >7.8 (2h) 2.f: ≥2 BG >5.3, 10.6, or 8.9 (0/1/2h) 2.g: BG ≥8.9 (2h) 2.h: BG ≥5.1, 10.0, or 8.5 (0/1/2h) (IADPSG criteria) 2.i: BG 7.8–11.0: IGT, >11.0: GDM (ADA criteria) | 3h 100-g OGTT 3.a: BG >5.3, 10.0, 8.6, or 7.8 (0/1/2/3h) (modified CC criteria) 3.b: BG >7.2 (3h) 3.c: BG >7.8 (3h) 3.d: BG >5.3, 10.0, 8.6, or 8.0 (0/1/2/3h) (modified CC criteria) 3.e: ≥2 BG ≥5.3, 10.0, 8.6, or 7.8 (0/1/2/3h) (CC criteria) 3.f: ≥2 BG >5.8, 10.5, 9.1, or 8.0 (0/1/2/3h) (NDDG criteria) |
| Study | Screening population | Definition of risk factors | Screening test type (test criteria; see Table 3 ) |
|---|---|---|---|
| Poncet B, Touzet S, Rocher L et al. | Alternative 1) high-risk screening Alternative 2-3) universal screening | Age > 35 yrs Family history of DM (first-degree relatives) BMI > 27 kg/m 2 Previous GDM, preeclampsia or fetal death after 3 months of gestation or child >4000 g | Alternative 1-2) two-step: 1h 50-g GCT (1.a) ± 100-g 3h OGTT (3.a) Alternative 3) one-step: 2h 75-g OGTT (2.a) |
| Larijani B, Hossein-nezhad A, Rizvi SW et al. | Alternative 1-2) universal screening Alternative 3-4) high-risk screening | Age > 35 yrs Family history of DM (first-degree relatives) Obesity (ND) Previous poor obstetric outcome (ND), polyhydramnios, or macrosomia (ND) Glucosuria | Alternative 1, 4) two-step: 1h 50-g GCT (1.a) ± 3h 100-g OGTT (3.b) Alternative 2-3) two-step:1h 50-g GCT (1.a) ± 3h 100-g OGTT (3.c) |
| Nicholson WK, Fleisher LA, Fox HE, Powe NR | Alternative 1-3) universal screening Alternative 4) no screening | Alternative 1) two-step: 1h 50-g GCT (1.b) ± 3h 100-g OGTT (3.d) Alternative 2) one-step: 2h 75-g OGTT (2.b) Alternative 3) one-step: 3h 100-g OGTT (3.d) | |
| Ayach W, Costa RA, Calderon Ide M, Rudge MV | Alternative 1-3) universal screening | Alternative 1) one-step: FBG (0.a) and risk factors present Alternative 2) one-step: 1h 50-g GCT (1.b) Alternative 3) one-step: 3h 100-g OGTT (3.e) | |
| Thung S, Pettker C, Funai E | Alternative 1-2) universal screening (as no risk-factors stated) Alternative 3) no screening | Alternative 1) two-step: 1h 50-g GCT (1.a) ± 3h 100-g OGTT (NS) Alternative 2) two-step: 1h 50-g GCT (1.b) ± 3h 100-g OGTT (NS) | |
| National Collaborating Centre for Women’s and Children’s Health (UK) | Alternative 1-20) 20 alternative combining high-risk screening vs. universal screening 21) no screening | Age ≥30 yrs BMI ≥27 kg/m 2 Family history of DM Previous GDM, macrosomia (ND), fetal death with no apparent cause, recurrent miscarriages or malformations | Alternative 1-20) 20 alternatives combining one-step vs. two-step: FBG, RBG, 1h 50-g GCT vs. 2h 75-g OGTT (NS) |
| Lee S, Pettker C, Funai E et al. | Alternative 1) universal screening (as no risk factors stated) Alternative 2) universal screening (as no risk factors stated) Alternative 3) universal screening (as no risk factors stated) | Alternative 1) one-step: 2h 75-g OGTT (2.c) Alternative 2) one-step: 2h 75-g OGTT (2.d) Alternative 3) one-step: 2h 75-g OGTT (2.e) | |
| Meltzer SJ, Snyder J, Penrod JR et al. | Alternative 1-3) universal screening | Alternative 1) two-step: 1h 50-g GCT (1.c) ± 3h 100-g OGTT (3.e) Alternative 2) two-step: 1h 50-g GCT (1.c) ± 2h 75-g OGTT (2.f) Alternative 3) one-step: 2h 75-g OGTT (2.f) | |
| Round JA, Jacklin P, Fraser RB et al. | Alternative 1-7) high-risk screening Alternative 8) no screening | Age > 25 yrs BMI ≥ 27 kg/m 2 Family history of DM High-risk ethnic group (ND) | Alternative 1) one-step: 2h 75-g OGTT (2.g) Alternative 2) one-step: FBG (0.b) Alternative 3) one-step: RBG (NS) Alternative 4) one-step: 1h 50-g GCT (NS) Alternative 5) two-step: RBG (NS) ± 2h 75-g OGTT (2.f) Alternative 6) two-step: FBG (0.b) ± 2h 75-g OGTT (2.f) Alternative 7) two-step: 1h 50-g GCT (NS) ± 2h 75-g OGTT (NS) |
| Mission JF, Ohno MS, Cheng YW, Caughey AB | Alternative 1-2) universal screening | Alternative 1) two-step: 1h 50-g GCT (1.b) ± 3h 100-g OGTT (either 3.e or 3.f) Alternative 2) one-step: 2h 75-g OGTT (2.h) | |
| Werner EF, Pettker CM, Zuckerwise L et al. | Alternative 1-2) universal screening Alternative 3) no-screening | Alternative 1) two-step: 1h 50-g GCT (1.a) ± 3h 100-g OGTT (3.a) Alternative 2) two-step: FBG (0.d) ± 2h 75-g OGTT (2.h) | |
| Marseille E, Lohse N, Jiwani A et al. | Alternative 1) universal screening Alternative 2) no screening | Alternative 1) one-step: 2h 75-g OGTT (NS) |
Costing
We found consistency between identified costs and perspectives where these were stated . Only one study included overheads or capital costs, whereas just two of the evaluations applying lifelong horizon included costs for treatment of T2DM. The remaining CUAs using lifelong horizon instead focused on the short-term cost of screening, treatment of GDM, and associated complications on short term. Except for two evaluations , we found a lack of information on how costs had been measured. Similarly, only one study explicitly stated that market prices were adjusted to costs using cost-to-charge ratios, but elsewhere we found it unclear whether the remaining costing was based on opportunity costing or market prices.
Measurement and valuation of consequences
Three studies based the effectiveness of screening on parallel cohort studies , whereas seven studies based the effectiveness on published literature . However, only two studies stated how the point estimates from the literature were weighted and pooled. Two studies had no information on methods or sources for the establishment of effectiveness, why we could not judge the validity of the established effectiveness in these studies.
The CUAs respectively applied the two-dimensional QALY and DALY, thereby combining lifetime expectancy with quality of life , as QALY incorporates utility and DALY disability associated with given health states. Except for three studies , we could not identify sources for the lifetime assumptions. Only two CUAs had no information on valuation of consequences. The remaining CUAs based the valuation of consequences on published literature. We found information supporting that one study based the utility on maternal and neonatal preferences, while another based it on maternal preferences. It was not stated which methods were used for the measurement of the neonatal preferences. However, as these authors respectively applied a societal and health-care sector perspective, it could be argued that the preferences should have been based on broader population. We could not find such information on the remaining CUAs, so it was not possible to uncover which population the utilities in the remaining CUAs were based on. Thus, we found it difficult to judge the consistency between stated perspective and consequences in these evaluations. As this is more of a theoretical discussion, the implications of this will not be elaborated in this article. Valuation of consequences was not relevant to the CEAs as there was no utility incorporated into the one-dimensional outcomes’ found cases of GDM, macrosomia, prematurity, perinatal mortality, or hypertensive disorders .
Estimates of cost-effectiveness
Three of the studies did not perform incremental analysis and only presented average costs of the screening alternatives. As these studies thereby focused on the most cost-saving screening alternatives, the trade-off in overseen health gains was not reflected in these evaluations. Thus, we chose not to include the results from these evaluations in the synthesis of the estimates of cost-effectiveness. Except from these three, all authors performed incremental analyses allowing cost-effectiveness estimates to be outlined.
In the studies performing incremental analysis, we found that one-step 2h 75-g OGTT and two-step 1h 50-g GCT ±100-g OGTT were the screening types most frequently found to be cost-effective. Two authors calculated ICERs using consecutive referencing , whereas the remaining authors used the same referent on all alternatives. We also found variation in the way estimates of cost-effectiveness were presented, though the authors primarily presented aggregated estimates of cost-effectiveness. However, one study focused on several outcome parameters, and therefore assessed the cost-effectiveness for these outcomes separately, whereas another study calculated the cost-effectiveness ratios for maternal and neonatal outcomes separately. Similarly, one study presented both the short- and the long-term cost-effectiveness ratios. Finally, one study incorporated a maximum willingness to pay threshold at 20,000 UK £ into their cost-effectiveness estimates, while their results were only displayed as a distribution of most cost-effective screening alternative within this threshold. Overall, we identified a large variance in the scale of the stated ICERs, and this tendency was further exaggerated by our inflation into 2014 US dollars ( Table 6 ).
| Study | Key cost components | Cost sources/data handling | Consequence sources/data handling | Key input for decision modeling |
|---|---|---|---|---|
| Poncet B, Touzet S, Rocher L et al. | Screening tests Obstetrical care Management of GDM Delivery care Sick leave | Prospective study ( n = 120) Hospital in Rhône-Alpes, France French key-letters costing system (≈DRG), literature, and expert opinions | Clinical effectiveness: published literature | GDM prevalence Screening refusal rate Risk factors Test effectiveness GDM treatment success |
| Discounting (NS) | Decision modeling Discounting (NS) | |||
| Larijani B, Hossein-nezhad A, Rizvi SW et al. | Test materials Time for testing (staff) | Source for cost valuation (NS) | Clinical effectiveness: Cohort ( n = 2416), four university hospitals in Teheran, Iran, period or retro-/prospective design: NS | |
| Discounting (NS) | Prevalence estimate on patient-level data Discounting (NS) | |||
| Nicholson WK, Fleisher LA, Fox HE, Powe NR | Treatment of maternal/neonatal complications (direct and indirect costs) | The 2003 Medicare resource-based relative value units/The Maryland Health Care Commission Database/The Bureau of Labor Statistics | Clinical effectiveness and utilities: published literature Life expectancy: The National Centre for Health Statistics | GDM prevalence Test effectiveness Hypertensive disease Polyhydramnios Cesarean/vaginal delivery and associated complications Neonatal hypoglycemia Macrosomia Shoulder dystocia Morbidity/infant death |
| Market prices adjusted using cost-to-charge-ratios Inflated into 2003 US dollars using consumer price index Discounted at 3% rate per yr | Decision modeling Discounting at 3% rate per yr | |||
| Ayach W, Costa RA, Calderon Ide M, Rudge MV | Laboratory costs (ND) | Public reimbursements | Clinical effectiveness: Prospective cohort ( n = 341) in university hospital in Matto Grosso do Sul, Brazil, July 1997–December 1999 | |
| Discounting (NS) | Prevalence estimate on patient-level data Discounting (NS) | |||
| Thung S, Pettker C, Funai E | Identified costs (ND) | Units/quantity/sources (NS) | Clinical effectiveness, lifetime expectancy, and utilities (NS) | |
| Discounting (NS) | Decision modeling Discounting (NS) | Perinatal death Shoulder dystocia | ||
| National Collaborating Centre for Women’s and Children’s Health (UK) | Testing (ND) Treatment of GDM (staff’s time, materials, medicine) Treatment of GDM associated complications (ND) | National Health Service/literature | Clinical effectiveness and utilities: published literature Lifetime expectancy (NS) | |
| Inflated into 2006 UK Pounds using Retail Price Index Discounting (NS) | Decision modeling Discounting (NS) | Test effectiveness/acceptability GDM treatment Induction of labor/cesarean section Stillbirth/neonatal death Shoulder dystocia Bone fracture Nerve palsy Jaundice Hypoglycemia | ||
| Lee S, Pettker C, Funai E et al. | Treatment of DM Treatment long-term child complications Preeclampsia management | Units/quantity/sources (NS) | Clinical effectiveness: published literature Lifetime expectancy and utilities (NS) | Maternal death due to preeclampsia Permanent brachial plexus injury |
| Discounting (NS) | Decision modeling Discounting (NS) | |||
| Meltzer SJ, Snyder J, Penrod JR et al. | Test materials Time (staff/women) | Statistics Canada/Canada Revenue Agency | Clinical effectiveness: RCT ( n = 1594) in university hospital in Montreal, Quebec, Canada, January 2001–2002 | |
| Inflated into 2002 Canadian dollars Discounting (NS) | Prevalence estimate on patient-level data Discounting (NS) | |||
| Round JA, Jacklin P, Fraser RB et al. | Test materials Time for testing (staff) Treatment of GDM Treatment of GDM-associated complications | Literature/National Health Service data | Clinical effectiveness and utilities: published literature Lifetime expectancy: 80 yrs (reference NS) | Test effectiveness/acceptability GDM-treatment Preeclampsia Induction of labor/cesarean section Neonatal jaundice Admission to neonatal nursery |
| Inflated to 2009 UK pounds Discounting at 3.5% rate per yr | Decision modeling Discounted at 3.5% rate per yr | |||
| Mission JF, Ohno MS, Cheng YW, Caughey AB | Treatment/diagnosis of GDM (direct and indirect costs) Admission for GDM-related conditions omitted | Published literature | Clinical effectiveness, utilities, life time expectancy: published literature | |
| Inflated into 2012 US dollars using consumer price index Discounting (NS) | Decision modeling Discounting at 3% rate per yr | Test effectiveness GDM treatment Preeclampsia Mode of delivery Shoulder dystocia Macrosomia Brachial plexus injury Hypoglycemia Hyperbilirubinemia NICU admission Maternal/neonatal death | ||
| Werner EF, Pettker CM, Zuckerwise L et al. | Testing (ND) Prenatal care Treatment of GDM-associated complications (short and long term) Prevention of DM | Published literature/Medicare reimbursements | Clinical effectiveness, utilities, and lifetime expectancy: published literature | Test effectiveness Preeclampsia Shoulder dystocia Preterm birth Cesarean delivery Stillbirth NICU admission Maternal diabetes postpartum |
| Inflated into 2011 US dollars using consumer price index Discounting at 3% rate per yr (sensitivity analyses: 1–5% rate per yr) | Decision modeling Discounting at 3% rate per yr (sensitivity analyses: 1–5% rate per yr) | |||
| Marseille E, Lohse N, Jiwani A et al. | Testing (materials, staff) GDM treatment, and intervention Treatment of T2DM | Personal communications/published literature | Clinical effectiveness and utilities: Published literature Lifetime expectancy: country-specific WHO life tables from 2009 | GDM prevalence Test performance Effectiveness of treatment Full set of perinatal adverse events Maternal/offspring T2DM |
| Converted into 2011 International dollars Discounting at 3% rate per yr | Decision modeling Discounting at 3% rate per yr |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
