Screening for type 2 diabetes: do screen-detected cases fare better?

Area of Science:

• Public Health
• Epidemiology
• Preventive Medicine

Background:

• Diabetes mellitus is a significant public health concern with substantial morbidity and mortality.
• Early detection and intervention are crucial for managing diabetes and preventing complications.
• Population-based screening programs aim to identify individuals with diabetes before symptomatic presentation.

Purpose of the Study:

• To compare health outcomes between individuals with screen-detected diabetes and those with clinically detected diabetes.
• To investigate the impact of screening participation on diabetes detection and subsequent health outcomes.

Main Methods:

• A population-based cohort study in Västerbotten County, Sweden, utilizing data from the Västerbotten Intervention Programme.
• Inclusion of 142,037 individuals eligible for 10-year interval diabetes screening (Oral Glucose Tolerance Tests).
• Identification and comparison of 1024 screen-detected and 8642 clinically detected diabetes cases (1992-2013) using Cox regression analysis, adjusting for confounders.

Main Results:

• Screen-detected diabetes cases were diagnosed at an average of 4.6 years younger than clinically detected cases.
• Clinically detected individuals experienced significantly worse health outcomes, including higher all-cause mortality (HR 2.07) compared to screen-detected individuals.
• Higher mortality was observed in clinically detected non-participants (HR 2.31) versus prior screening participants (HR 1.70), with similar trends for cardiovascular disease, renal disease, and retinopathy.

Conclusions:

• Individuals diagnosed with diabetes through screening exhibit better health outcomes, characterized by earlier diagnosis and reduced mortality and complications.
• The findings suggest a potential benefit of organized diabetes screening programs in improving patient prognosis.
• Further research is needed to disentangle the contributions of earlier treatment versus biases (healthy user, lead time, length time) to these observed differences.