Health economic considerations of screening for early type 1 diabetes

Abstract

Health economics of screening for early type 1 diabetes (T1D) is an emerging area of research. Historically, screening for early T1D has been limited to prospective cohort studies of the aetiology and natural history of the disease as well as trials of secondary prevention. In the past few years, successful large‐scale studies have expanded the screening to the general population, providing data needed to estimate the cost‐effectiveness of such programs and their economic viability as they become part of public health preventive services. This review covers available results of health economics analyses from early estimates through state transition cohort simulation models relying on birth cohort studies to real‐world experience from general population screening and, finally, predictions to a national scale. The cost‐effectiveness analyses of screening for early T1D to date suggest that the financial investment in such programs is justified by the health benefits and savings achieved. Screening can be cost‐effective if it reduces the incidence of diabetic ketoacidosis at diagnosis and decreases healthcare costs associated with long‐term complications. However, the cost‐effectiveness varies based on population prevalence of T1D, frequency of diabetic ketoacidosis at clinical diagnosis, screening test accuracy and cost as well as healthcare system efficiency. All these factors influence the HbA1c at diagnosis and glycaemic control as well as insulin requirements later in the course of the disease. Earlier interventions aiming at delaying insulin dependency may further improve the cost‐effectiveness.

Plain Language Summary

Type 1 diabetes (T1D) is one of the most frequent severe chronic diseases of childhood, increasing in numbers with age, affecting in lifetime 1 in 100 people. It is caused by the immune system attacking insulin‐producing cells in the pancreas. Stage 1 of T1D is diagnosed when multiple markers of the attack (islet autoantibodies) are detected. The disease progresses to stage 2 when most of the insulin is lost and blood sugar levels become abnormal. Stage 3 is defined by very high blood sugars leading often to diabetic ketoacidosis (DKA) and need of insulin treatment to safe life. This progression happens over a period of months or years, providing an opportunity for screening to detect the disease early and slow the progression with drug therapy as well as to provide education that prevents DKA. Early diagnosis and start of insulin therapy when HbA1c is just above 6.5% helps to avoid multiorgan complications that increase with age and account for most of the direct medical cost and lost workforce participation and productivity associated with T1D. Studies following high‐risk youth for decades provided early estimates of the cost‐effectiveness of screening for stage 1 and stage 2 T1D. Screening cannot be limited to relatives of people with T1D because they represent only 10‐15% of new T1D diagnoses. In recent years, successful programs such as Fr1da in Germany or Autoimmunity Screening for Kids (ASK) in USA expanded the screening to the general population, providing data needed to estimate cost‐effectiveness as they become part of public health preventive services. Testing for islet autoantibodies dramatically reduced the incidence of DKA at diagnosis and may decrease healthcare costs associated with long‐term complications; however, an extended follow‐up of screening participants is necessary. Screening for early T1D can be combined with screening for celiac disease, high cholesterol levels, or other conditions detectable by blood test, further improving cost‐effectiveness. The widely accepted routine newborn screening for about 30 rare diseases, affecting 1 in 600 infants, costs $125‐150 per child. In contrast, 1 in 30 children has early T1D or celiac disease that could be detected for less than $50 per child screened. Results of the programs in Germany and USA indicate that the cost‐effectiveness of screening depends on population prevalence of T1D, frequency of DKA, screening test accuracy and cost as well as healthcare system efficiency. These analyses allow national‐scale predictions of the number of high‐risk youths and the resources necessary to successfully monitor them during stage 1 and 2. Screening at age 2 and 6 years in Germany would annually yield a similar number of children at early stage T1D to those currently diagnosed at stage 3 and reduce DKA at diagnosis by 61%. If all USA children were screened today, it is estimated that 350,000 would be diagnosed with stage 1 and 70,000 with stage 2 T1D. The cost‐effectiveness of analyses of screening for early T1D suggest that financial investment in such programs is justified by the observed health benefits and savings.

1. BACKGROUND

Type 1 diabetes (T1D) is a chronic autoimmune disease in which progressive destruction of pancreatic islet β‐cells leads to insulinopenia, hyperglycaemia, acute metabolic complications, multi‐organ chronic complications and premature death. The earliest stages of T1D can be detected by testing for islet autoantibodies to insulin, glutamic acid decarboxylase (GAD), tyrosine phosphatase islet antigen‐2 (IA‐2) and zinc transporter type 8 (ZnT8). ¹ Persistent presence of one or more of these autoantibodies marks the beginning of the autoimmune phase of T1D. ² After a variable time, metabolic disease ensues with an inconspicuous onset of dysglycaemia, gradually progressing to overt hyperglycaemia and dependence on life‐saving therapy with exogenous insulin. ³ , ⁴ Current staging of early T1D is summarized in Table 1. ⁵

TABLE 1.

Stages of early type 1 diabetes (FPG, fasting plasma glucose; OGTT, oral glucose tolerance test). ⁵

Stage	Islet autoantibody	Glycaemic status	Symptoms
Immune activation	Single autoantibody	Normoglycaemia FPG <5.6 mmol/L (<100 mg/dL) 2‐h OGTT <7.8 mmol/L (<140 mg/dL) HbA1c <5.7% (<39 mmol/mol)	No symptoms
Stage 1 T1D	≥2 autoantibodies
Stage 2 T1D	≥1 a autoantibody	Dysglycaemia, one more of the following: FPG 5.6–6.9 mmol/L (120–125 mg/dL) 2‐h OGTT values 7.8–11.1 mmol/L (140–199 mg/dL) HbA1c 5.7–6.4% (39–47 mmol/mol) or ≥ 10% increase	No symptoms
Stage 3 T1D	≥1 a autoantibody	Hyperglycaemia b , one or more of the following: FPG ≥7.0 mmol/L (≥126 mg/dL) 2‐h OGTT ≥11.1 mmol/L (≥200 mg/dL) HbA1c ≥6.5% (48 mmol/mol) random plasma glucose ≥11.1 mmol/L (≥200 mg/dL) with overt symptoms;	Polyuria, polydipsia, weight loss, fatigue

^a Some persons with confirmed persistent prior multiple autoantibody positivity may revert to single autoantibody status or negative status in stage 2. ⁴⁶

^b In the absence of unequivocal hyperglycaemia, diagnosis requires two abnormal results from different tests, which may be obtained at the same time (e.g., A1c and FPG), or the same test at two different time points.

Islet autoantibody screening and early detection of pre‐symptomatic T1D can offer significant benefits but are associated with certain costs (Table 2). The benefits emphasized in health economics analyses include the prevention of diabetic ketoacidosis (DKA) and associated acute neurologic, neurocognitive, vascular and kidney complications, as well as hospitalization at clinical diagnosis. Early awareness and start of insulin therapy when HbA1c is just above 6.5% help to avoid severe glucotoxicity and further loss of β‐cell function, which, in turn, improves glycaemic control and long‐term outcomes. ⁶ , ⁷ These benefits have tangible economic implications, with reductions in healthcare costs and increased workforce participation and productivity. ⁸ , ⁹ Social determinants of health, including parental education, income and health insurance that provides access to optimal diabetes care, predict both DKA at diagnosis and long‐term diabetes complications. These factors, in addition to race/ethnicity (often used as a proxy variable), should be included in health economics analyses of screening for early T1D.

TABLE 2.

Benefits and costs of screening for early type 1 diabetes.

Potentials benefits of screening

Prevention of DKA at diagnosis of clinical T1D, its short‐term neurologic, vascular and kidney complications, as well as long‐term neurocognitive sequela. Elimination of hospitalization at diagnosis, reduction of the associated direct and indirect cost. Early control of hyperglycaemia and reduction of glucotoxicity‐induced loss of endogenous insulin secretion, lowering the risk of recurrent DKA and severe hypoglycaemia. Better long‐term glycaemic control lowering the risk of long‐term micro‐ and macrovascular complications and neuropathy. Gradual effective education regarding diabetes signs and symptoms, monitoring of glycaemia, optimal diet and physical activity prior to initiation of insulin therapy. Improved quality of life of affected children and their relatives at the time of clinical diagnosis and following initiation of insulin therapy. Opportunity to receive approved disease‐modifying therapy in autoantibody‐positive individuals to preserve β‐cell function and delay insulin dependence. Opportunity to participate in clinical trials of interventions aimed at halting or slowing progression to insulin dependence.

Costs of screening

Education of health care providers. Raising community awareness about benefits of screening, information materials. Laboratory cost of screening and confirmation. Provider time to complete screening and confirmation. Results notification, explanation and documentation in health medical records. Monitoring for development of dysglycaemia and progression to hyperglycaemia. Caregiver early education to prevent DKA and hospitalization. Psychosocial support for families with increased anxiety. Loss of caregiver income due to increased health care utilization for monitoring. Potential problems with health insurance, credit ratings and job prospects. Currently, most primary health care and endocrinology services are not prepared for early T1D screening. Significant infrastructure investment will be needed. Cost of healthcare and caregivers caused by false negative screening results.

2. METHODS AND MODELLING

Historically, screening for early T1D has been limited to studies of the aetiology and natural history of the disease ¹⁰ as well as trials of secondary prevention. ¹¹ These studies oversampled first‐degree relatives of people with T1D because of their elevated genetic risk and higher participation in research. However, ~85% of persons diagnosed with T1D lack a family history of T1D. ¹² , ¹³ Health economics modelling based on cohort studies of first‐degree relatives is likely to be biased. They may underestimate benefits due to the lower frequency of DKA at diagnosis among relatives compared to the general population. ¹⁴ On the other hand, models based on screening relatives may overestimate compliance with screening protocols and underestimate the cost of monitoring children identified through screening.

In the past few years, successful large‐scale studies such as Fr1da in Germany ¹⁵ or ASK in the USA ⁸ expanded the screening to the general population, providing data needed to estimate the cost‐effectiveness and economic viability of such programs as they become part of public health preventive services. Importantly, both studies have demonstrated significant reductions in DKA and improved long‐term glycaemia.

Disease‐modifying therapies are becoming available for use in early T1D, holding the promise of delaying progression to insulin dependence. In 2022, the FDA approved teplizumab (Tzield®, Sanofi), which demonstrated greater than a 32‐month median delay of progression from stage 2 to Stage 3 T1D. ¹⁶ , ¹⁷ Such therapies, capable of delaying insulin dependence, may be of particular value in children. While the therapy has been already administered to hundreds of people at stage 2 T1D in the USA and a growing number of people in other countries, it is too early to estimate its potential impact on the cost‐effectiveness of screening and standard metrics such as Quality‐Adjusted Life Years (QALYs) and Incremental Cost‐Effectiveness Ratios (ICERs). Appropriate data are being collected, and the inclusion of the impact of disease‐modifying therapies on screening cost‐effectiveness should be included alongside other key factors, e.g., population prevalence of T1D, frequency of DKA at diagnosis, screening test accuracy and healthcare system efficiency.

This review will cover available results of health economics analyses from early estimates through state transition cohort simulation models relying on birth cohort studies to real‐world experience from general population screening and predictions to a national scale.

2.1. Early estimates of cost‐effectiveness based on sole benefit of preventing DKA at diagnosis

Multiple studies of children at high genetic risk have demonstrated that islet autoantibody screening, followed by education of caregivers regarding signs/symptoms of diabetes and periodic monitoring of glycaemia, can prevent potentially life‐threatening DKA and hospitalization at diagnosis (Table 3). A report based on data available in 2015 compared the cost of population screening with the benefit of preventing DKA in children younger than 5 years. ¹⁸ The cost of screening included one‐time HLA typing of the general population, followed by islet autoantibody testing in children genetically at high risk every 6 months until age 5 years, similar to the protocol of The Environmental Determinants of Diabetes in the Young (TEDDY) Study. ¹⁹ The potential benefits of screening include reductions in lost parental income, medical expenses, morbidity and mortality. The authors concluded that screening in this format, for the sole purpose of reducing the cost of DKA at the onset of T1D, was not economically viable unless HLA testing and autoantibody testing could be performed for less than $1 and $0.03, respectively. The major limitation of that analysis was its narrow focus on the cost of avoiding DKA at diagnosis while ignoring the long‐run cost‐effectiveness. Several longitudinal analyses have since demonstrated sustained improvement of glycaemic control from early detection and education, ⁶ , ²⁰ , ²¹ suggesting that the benefits of early diabetes screening may not materialize until later stages of the disease. ²² , ²³ , ²⁴ The 2015 analysis was also narrowly focused on very young children (<5 years), while most children develop clinical T1D after 8 years of age, and 60% of all T1D patients are diagnosed as adults.

TABLE 3.

Evidence that screening, monitoring and education can prevent DKA at diagnosis of clinical T1D.

Study	Frequency of DKA	References
Children at high genetic risk followed by research studies
DAISY (Colorado) DAISY, TEDDY, TrialNet (Colorado) DIPP (Finland) TEDDY (Colorado, Georgia, Florida, Washington, Sweden, Finland, Germany)	3.3% 4.9% 5.0% 6.1%	Barker J., 2004 47 Sooy M., 2024 48 Hekkala A., 2018 49 Jacobsen L., 2022 50
Children identified through general population screening programs
Fr1da (Germany)	5.6%	Ziegler A‐G., 2020 15
ASK (Colorado)	3.8%	Rewers M., IDS 2023
Children diagnosed with T1D without prior screening
Sweden Finland Germany	22% 23% 24%	Wersäll J., 2021 51 Hekkala A., 2018 49 Kamrath C., 2020 52
USA	40%–62%	Rewers A., 2015 14 Alonso G., 2020 53

2.2. Health economics of screening estimates derived from birth cohort studies

Birth cohort studies, such as BABYDIAB, DAISY, DIPP, DEWIT, DiPiS and TEDDY, initiated more than 25 years ago, have provided invaluable data regarding the natural history of T1D as well as genetic and environmental determinants of T1D risk. ¹⁹ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ These studies in Europe and the USA have followed children at increased genetic risk for the development of islet autoimmunity and progression to diabetes. Observations from these cohorts ² led to the current staging of T1D (Table 1), ³⁰ forming consensus regarding optimal age and frequency of screening. ³¹ , ³² While screening at ages 2–4 and 4–6 years is necessary to detect more than 70% of stage 1 and stage 2 patients, an additional screening at 9–11 years can improve the sensitivity to over 80%. Notably, key input data for the modelling of screening cost‐effectiveness have been derived from close collaboration among these cohort studies using harmonized combined data ² , ³³ allowing assessment of the performance of islet autoantibodies ³⁴ and genetic markers ³⁵ , ³⁶ in screening for T1D.

Analysis based on DAISY and TEDDY data (Mark Trusheim, personal communication) compared islet autoantibody screening followed by glucose testing (i.e., A1c, OGTT) and therapeutic intervention to intercept T1D with a strategy using genetic risk score (GRS) screening followed by annual autoantibody, glucose testing and therapeutic intervention. Using benchmarks for participation in newborn screening (85%) and patient follow‐up (80%), as well as rapid T1D progression in some children, the GRS screening strategy would intercept at dysglycaemia 40% of children progressing to T1D by age 15 years. The strategy using autoantibody screening at age 2 and 5 years would intercept 26%. GRS scenario short‐term net incremental costs totaled $640 million or $153 000 per T1D case intercepted. Autoantibody screening incremental costs totaled $400 million, resulting in $150 000 per T1D case intercepted. For the GRS strategy to break even in terms of costs, GRS testing would need to cost $5, and the total cost of the autoantibody test would need to be $12.

A recent report based on data from the GPPAD‐02, TEDDY and Type 1 Diabetes Intelligence studies modelled the performance of a hypothetical screening for islet autoantibodies at ages 2, 6 and 10 years with or without prescreening for high‐risk HLA genotypes. ³⁶ In pilot studies, for instance, the Virginia PrIMeD project, ³⁷ genetic pre‐selection encountered challenges of low recall rates, loss to follow‐up, socioeconomic biases and limited applicability across diverse ancestries. Integration of genetic testing into routine healthcare will be very challenging in fragmented systems such as that in the USA. The authors conclude that the ultimate success of any screening program will depend on maximizing public and healthcare‐provider engagement, ensuring high participation and addressing socioeconomic and demographic disparities. Digital health infrastructure may help compliance, but brings additional concerns related to the privacy of genetic data and screening results.

2.3. Models based on the results from large‐scale general population screening programs

2.3.1. Autoimmunity screening for kids (ASK)

Between 2017–2025, the ASK study in Colorado, USA screened >40 000 children aged 1–17 years for islet autoantibodies as well as transglutaminase autoantibodies – a marker of celiac disease. In a 2020 analysis ⁸ that included data for 10 029 children and adolescents screened by July 2018, the study reported cost and cost‐effectiveness of just screening for early T1D (celiac disease was not considered). The ethnically diverse study population was representative of the underlying population of Denver, with 51% Hispanics, 35% non‐Hispanic whites, 8% non‐Hispanic African Americans and 6% other races. Only 5% of the participants had a first‐degree relative with T1D. Cases were defined as children and adolescents with multiple islet autoantibodies (10‐year risk of progression to clinical diabetes of 70%) ² or a single high‐affinity autoantibody detected by both radiobinding and electrochemiluminescent assays after confirmation. ³⁸

A bridge model was used to simulate participant progression from islet autoantibody positivity to a diagnosis of Stage 3 T1D from ages 2–30 years. Screening and follow‐up costs were assumed for all those screened up until age 18 and simulated patients were followed until age 30 to track diagnosis of Stage 3. Follow‐up and monitoring costs until age 30 years were included: communicating results, two follow‐up visits a year until age 18 years, and repeat screening panel costs for all positive persons identified over the duration of the model time horizon. A single high‐affinity islet autoantibody was detected in 0.53% of the participants and multiple autoantibodies in 0.48% (n = 101).

The total research costs for all children screened for ASK were $471 000 and estimated at $1 417 500 if it were done as a routine care screening. Cost per child screened was $47 for ASK screening and $141 for routine screening. Cost per case detected was $4700 for ASK research screening versus no screening and $14 000 for routine screening versus no screening. The main driver of total cost differences was the negotiated screening price of $15 versus the CMS laboratory fee schedule price of $138. Of note, cost per case detected does not provide a comprehensive assessment of the value of early diabetes screening, given the potential impact of screening on lifetime survival and cost outcomes.

In a separate lifetime simulation model from age ≥30 years, only for those diagnosed with T1D, the impact of changes in HbA1c on clinical and economic outcomes was estimated from early detection of diabetes. The model estimated the long‐run lower HbA1c benefits needed to offset upfront screening costs by avoiding long‐term diabetic complications and associated health care expenditures. Costs and outcomes were discounted at 3% per year. Using Colorado‐specific data, the model assumed an 80% progression rate to Stage 3 T1D, a 46% prevalence of DKA at diagnosis in routine care without screening, and the subsequent population‐level HbA1c average of 9.1% (76 mmol/mol). ⁶

Projected total discounted direct medical cost of diabetes over a lifetime for an average screening participant diagnosed with Stage 3 T1D was $260 256, including screening and monitoring – $71 781, DKA at diagnosis – $3078, and cost of treatment of diabetes and its complications – $250 000. The quality‐adjusted life‐year (QALY) gained over a lifetime horizon is the recommended metric of effectiveness for cost‐effectiveness studies because it combines quality and quantity of life that can be used to compare the value of interventions across diseases. ³⁹

Discounted QALYs over a lifetime for these people were projected at 33 QALYs per person, on average. To achieve value thresholds of $50 000–$100 000 per QALY, screening costs would need to be offset by cost savings through a minimum 20% relative reduction in DKA events at diagnosis in addition to a 0.1% (1.1 mmol/mol) subsequent improvement in HbA1c over a lifetime compared with no screening. Greater reductions in DKA events at diagnosis of at least 40% and a subsequent improvement in HbA1c of 0.3% (3.3 mmol/mol) led to cost savings for ASK screening. Cost savings were not achieved from avoiding DKA events alone, consistent with a previous report. ¹⁸ These analyses did not include any of the “social” costs of late diagnosis, such as missed work time and lost productivity, which may account for up to 40% of the total cost of T1d lifetime.

The study concluded that the cost‐effectiveness of routine screening will largely depend on the cost of laboratory testing, and further effort is needed to develop affordable high‐throughput methods for the detection of islet autoantibodies. While the ASK discounted laboratory research price of $15 may not be sustainable in real‐world mass screening, the costs of a routine screening would be completely offset with a 40% decrease in DKA and the laboratory price of $80.

2.3.2. Fr1da

The Fr1da study in Bavaria, Germany screened >90 000 children aged 1.75 to <6.0 years, between 2015–2019. ¹⁵ Families of children who screened positive for ≥2 islet autoantibodies were invited to participate in a programme of diabetes education, metabolic staging, assessment of psychological stress and prospective follow‐up for progression of their T1D status, until July 31, 2019. ⁹ Stage 1 T1D was present in 0.22%, stage 2 in 0.02% and Stage 3 in 0.03% of the children screened. The screening was performed by primary care paediatricians during a well‐baby check. During follow‐up, DKA at diagnosis of Stage 3 T1D was present in only 3.2% of the children. The Fr1da study is still going on demonstrating sustainability of this screening model.

Analysis of the costs associated with this part of the Fr1da study indicated that it was cost‐effective as a public health program. ⁹ Data were analysed with models that mimic procedures during the screening process. The primary model resembled screening for presymptomatic T1D in standard care in Germany. It included only 50% of the costs associated with obtaining informed consent in the Fr1da and assumed that negative screening results would be communicated to families only if a second blood sample was requested to confirm the initial screening results. However, it assumed that all children diagnosed with early T1D would receive metabolic staging and diabetes education as part of standard care. Alternative models assumed higher or lower provider time needed to complete the screening, as well as higher or lower laboratory costs, as they can be expected in standard care for various scenarios.

The costs per child screened in the Fr1da study were euro 28.17 (95% CI 19.96; 39.63) and the costs per child diagnosed with presymptomatic type 1 diabetes were euro 9117 (6460; 12 827). ⁹ Assuming a prevalence of presymptomatic type 1 diabetes of 0.31%, as in the Fr1da study, the estimated costs in standard care in Germany would be euro 21.73 (16.76; 28.19) per screened child and euro 7035 (5426; 9124) per diagnosed child. Of the projected screening costs, euro 12.25 would be the costs in the medical practice, euro 9.34 for coordination and laboratory, and euro 0.14 for local diabetes clinics. The alternative models showed that the provider's time rather than laboratory cost was the main predictor of screening cost.

Comparison of ASK and Fr1da results points to important cross‐country differences in the economic determinants of screening programs. The cost per screened child in a research setting and in routine care was higher in the USA than in Germany, consistent with generally more expensive medical and preventive care costs in the USA. ⁴⁰ However, standard reimbursement fees for provider time and laboratory services remain to be established in either country and will be subject to stringent negotiations. As of now, it appears that, in contrast to Germany, laboratory fees rather than provider time reimbursement are likely to drive the cost of screening in the USA. Preliminary comparisons of the costs of early T1D screening vs. standard newborn screening suggested that in both countries early T1D screening would detect more cases of preventable disease at a lower cost.

2.4. Population‐scale estimates and implementation challenges

Large‐scale general population screening studies conducted in the USA and Germany have led to population‐scale estimates of the number of children that could be detected in these countries and associated costs.

2.4.1. USA

Approximately 0.3% of children born each year will develop T1D by age 15, and twice as many will develop T1D later in life. ⁵ Thus, nearly 11 000 of the 3.59 million children born in the U.S. annually will likely progress to T1D by their 15th birthday. Identifying these children remains a barrier to prevention. Multiple research studies have shown the feasibility of screening and identifying those in the early stages of type 1 diabetes, ⁴¹ but none have yet been translated to the public health setting.

The Autoimmunity Screening for Kids (ASK) program in Colorado found 0.5% of general population children to be positive for multiple islet autoantibodies and approximately as many positive for a single islet autoantibody, confirmed by two methods at two or more time points. ⁴² Extrapolating ASK findings, if all 70 million U.S. children 1–17 years of age were screened today, an estimated 700 000 would be found with islet autoantibodies and a high risk for type 1 diabetes. Half, or 350 000, would be found to express multiple islet autoantibodies and an estimated 70 000 of those would be at stage 2 T1D (dysglycaemia) (Table 4).

TABLE 4.

Estimated number of people with early T1D in the USA.

	USA
	Children	Adults
Stage 1 Prevalent persons	280 000	400 000
Stage 2 Prevalent persons	70 000	70 000
Stage 3 New patients/year	25 000	35 000
Stage 3 Prevalent patients	150 000	1 400 000

TEDDY study has followed from birth 8667 children at high genetic risk selected from the general population, with only 10.6% being a first‐degree relative of a person with T1D. ¹⁰ The cumulative incidence of confirmed persistent islet autoimmunity has rapidly increased in the initial 6 years of life and more slowly between 6–15 years, with very few new seroconversions in older teenagers (Figure 1, solid black line). The results of TEDDY appear to indicate that nearly all diagnoses of T1D in adults follow the initiation of islet autoimmunity in childhood. Data from the initial 35,000 children screened by ASK (unpublished) allowed us to estimate the age‐specific prevalence of islet autoimmunity up to the age of 17 years (Figure 1, dashed line). It appears to peak at 12–13 years of age, followed by a decline in older groups due to progression to clinical T1D (red solid line ⁴³ ), scarcity of new seroconversions and remission in some cases. The data for adults is sparse, which presents a challenge in modeling the lifetime cost‐effectiveness of screening.

FIGURE 1.

Estimated age‐specific prevalence of confirmed persistent islet autoantibodies in the general population, USA. (1) Cumulative incidence of islet autoimmunity extrapolated to the general population from TEDDY results (Rewers M et al. Diabetes Care 2025 https://doi.org/10.2337/dc24‐2886 figure 2A,B). (2) Estimated prevalence of multiple and single islet autoantibodies accounting for progression to Stage 3 T1D and a 25% remission. Prevalence of persistent islet autoantibodies confirmed by two methods observed in Colorado population screened by ASK. (3) Cumulative incidence of Stage 3 T1D derived from Rogers MAM et al. BMC Med 2017:15;199 Table 1).

To put screening for early T1D in a broader context of paediatric screening, the routine newborn screening for ~30 rare diseases costs the U.S. ~$125–$150 per child, depending on the panel requested. ⁴⁴ Fewer than 1 in 600 infants have one of these rare diseases detectable by testing a blood sample. In contrast, one in 30 children has early type 1 diabetes or celiac disease; the ASK program detects these conditions in participating children at <$50 per child screened.

2.4.2. Germany

Prevalence of children with early T1D observed in Fr1da was used to predict the number of children identified and eligible for follow‐up in a hypothetical nation‐wide general population screening at ages 2 and 6 years. ⁴⁵ Among 76 322 children screened by Fr1da at 2 years, 0·21% were positive for multiple islet autoantibodies. An additional 0.19% of initially negative children (n = 6812) developed multiple islet autoantibodies when rescreened approximately 4 years later. These estimates were extrapolated to 800 000 children aged 2 and 800 000 children aged 6 living in Germany. Screening at ages 2 and 6 years, with a 90% participation, would annually yield 2880 screening positive young children, compared to approximately 3200 children under the age of 18 years diagnosed yearly with clinical T1D nation‐wide. At the observed rates of progression and 2880 early T1D children identified annually, 15 210 (95% CI 10 964–21 537) children with early T1D would enter monitoring after 10 years of screening. The proposed monitoring would include random plasma glucose, biannual HbA1c and less frequent OGTT or CGM for an average of 6 years before progressing to clinical diabetes. In comparison, there are currently around 32 000 people younger than 18 years with clinical T1D in paediatric diabetes care in Germany. The authors estimated that 66% of the children diagnosed in Germany annually would be identified through this screening. This could eventually reduce the number of cases of DKA at clinical onset by 61%, from 960 to 376 cases annually, and hospitalization time would be reduced by 18% from 35 200 to 28 828 days. While these estimates need to be interpreted with caution, this report clearly sets an example of how to approach planning for the transition from research to public health preventive service.

3. CONCLUSIONS AND POLICY IMPLICATIONS

In conclusion, the cost‐effectiveness analyses of screening for early T1D suggest that the financial investment in such programs is justified by the observed health benefits and savings. Early disease‐modifying interventions aiming at delaying insulin dependency may further improve the outlook. However, the real‐world results may differ by state and healthcare systems as well as human resource time, often intangible, needed to instigate screening and properly monitor high‐risk children. Further careful analysis of the resource use, cost and the outcomes is needed. Screening can be cost‐effective if it reduces the incidence of diabetic ketoacidosis at diagnosis and decreases healthcare costs associated with long‐term complications. Other social benefits (productivity, life quality) can make this even more societally attractive. The cost‐effectiveness varies based on population prevalence of T1D, frequency of DKA at clinical diagnosis, screening test accuracy and its cost as well as healthcare system efficiency.

CONFLICT OF INTEREST STATEMENT

The author declares no conflicts of interest.

PEER REVIEW

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.16522.

Rewers M. Health economic considerations of screening for early type 1 diabetes. Diabetes Obes Metab. 2025;27(Suppl. 6):69‐77. doi: 10.1111/dom.16522

DATA AVAILABILITY STATEMENT

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.