# The borderline effect for diabetes: when no difference makes a difference

## TL;DR

This paper shows that being labeled as prediabetic may increase the risk of developing diabetes, even when health differences are minimal.

## Contribution

The study introduces the 'borderline effect' where diagnostic labels influence health outcomes despite similar initial health status.

## Key findings

- Participants labeled as prediabetic perceived higher risk than those with similar but differently labeled results.
- Patients labeled as prediabetic were more likely to develop diabetes than those with similar A1c levels but a normal label.

## Abstract

We hypothesized that people at the borderline of being labeled as “prediabetic” based on A1c blood test results, who initially face equivalent risks of developing diabetes but who are labeled differently, would be more likely to develop diabetes when labeled as “prediabetic” as a result of the label. Study 1 served to establish the psychological effect of the prediabetes label: we surveyed 260 participants on Amazon Mechanical Turk to test whether risk perception significantly increased when comparing A1c test results that differed by 0.1% and led to different diagnostic labels (5.6 and 5.7%) but did not significantly increase when comparing those that differed by 0.1% but received the same label (5.5%/5.6 and 5.7%/5.8%). Study 2 explored whether labels are associated with different rates of developing diabetes when the initial difference in A1c results suggests equivalent risk. Using data from 8,096 patients, we compared patients whose initial A1c results differed by 0.1% and found those who received results labeled as prediabetic (A1c of 5.7%) were significantly more likely to develop diabetes than patients whose initial results were labeled as normal (5.6%). In contrast, patients whose initial results differed by 0.1% but who received the same “normal” label (5.5 and 5.6%) were equally likely to develop diabetes. These preliminary results suggest that diagnostic labels may become self-fulfilling, especially when the underlying pathology of patients receiving different labels does not meaningfully differ.