Highlights
- Clinical decision support systems (DSS) for type 1 diabetes (T1D) do not yield uniform glycemic benefits across all populations; efficacy is highly dependent on baseline patient characteristics.
- Informative DSS (iDSS), which provides feedback for decision-making rather than direct commands, significantly reduced time above range for users with lower diabetes-related knowledge and higher baseline HbA1c.
- Psychosocial factors, specifically diabetes-related emotional distress and hypoglycemia worry, are primary predictors of low technology engagement.
- User preference plays a critical role in adherence, with 40% of participants preferring informative feedback over prescriptive recommendations.
The Challenge of Personalizing Type 1 Diabetes Management
The management of type 1 diabetes has undergone a technological revolution over the last decade. The widespread adoption of continuous glucose monitoring (CGM) and advanced insulin pump therapy has provided clinicians and patients with an unprecedented volume of data. However, data alone does not equate to better outcomes. The cognitive burden of interpreting glucose trends, calculating insulin doses, and accounting for variables like exercise and carbohydrate intake remains a significant hurdle for many patients. This gap has led to the development of Decision Support Systems (DSS)—algorithmic tools designed to translate raw data into actionable clinical insights.
Despite the theoretical promise of DSS, real-world studies often report low user engagement and modest improvements in glycemic control. The mismatch between technological capability and clinical efficacy suggests that “human factors”—the psychological, educational, and behavioral attributes of the user—may be the missing link in the success of these digital health interventions. A recent randomized controlled trial by Pavan et al. investigates this interplay, offering a critical look at why some patients thrive with DSS while others do not.
Study Design and Methodology
The researchers conducted a robust randomized controlled trial involving 53 adults with type 1 diabetes. The cohort was balanced between those using multiple daily injections (MDI) and those using insulin pump therapy, with all participants utilizing CGM. The study employed a three-period crossover design, where each participant underwent three distinct 2-month interventions in a randomized order:
1. No DSS (Control)
Participants managed their diabetes using standard-of-care methods without the assistance of specialized decision-support software.
2. Informative DSS (iDSS)
The iDSS provided summary feedback and retrospective analysis. Instead of telling the user exactly what to do, it highlighted patterns and provided the necessary information for the user to make their own informed decisions. This approach functions as a pedagogical tool, aiming to increase the user’s self-efficacy and understanding of their own glycemic patterns.
3. Prescriptive DSS (pDSS)
The pDSS took a more direct approach, recommending specific treatment actions, such as precise bolus amounts or adjustments to basal rates. This system was designed to reduce the user’s cognitive load by providing a clear, actionable directive for immediate therapy optimization.
The primary outcomes were CGM-derived glycemic metrics, including Time in Range (TIR), Time Above Range (TAR), and Mean Sensor Glucose. Crucially, the researchers also conducted exploratory analyses to correlate these outcomes with psychosocial variables, including diabetes-related knowledge, emotional distress, and hypoglycemia worry.
Key Findings: The Paradox of Technology and Engagement
The aggregate results of the study initially appeared underwhelming: there were no statistically significant differences in primary glycemic outcomes across the three interventions for the total study population. However, when the data were stratified by human factors, a more nuanced and clinically significant picture emerged.
The Knowledge Gap and iDSS Efficacy
For participants with lower baseline diabetes-related knowledge and those starting with a higher hemoglobin A1c, the use of iDSS resulted in a significant reduction in hyperglycemia. Specifically, these users saw a 6% reduction in average time spent above 180 mg/dl (p < 0.001) when using the informative system compared to no DSS. This suggests that for individuals who struggle with the foundational principles of diabetes management, a system that explains "the why" behind the data is more effective than one that simply provides a prescription.
The Impact of Psychosocial Barriers
The study identified a clear negative correlation between psychological distress and technology use. Participants reporting higher levels of emotional distress (p < 0.001) and significant worry about hypoglycemia (p < 0.01) showed markedly lower engagement with the DSS modules. This indicates that when a patient is overwhelmed by the emotional burden of their condition or paralyzed by the fear of low blood sugar, even the most advanced algorithm may fail because the user lacks the mental bandwidth or trust to interact with it.
Preference and Adherence
Engagement was significantly higher when participants were using the system they personally preferred (p < 0.01). Interestingly, 40% of the cohort preferred the iDSS over the pDSS, highlighting that a substantial portion of the T1D population values autonomy and learning over automated recommendations.
Clinical Interpretation: DSS as a Learning Tool
The finding that iDSS outperformed pDSS in specific subgroups is a critical takeaway for clinicians. In the current medical climate, there is a strong push toward full automation (such as closed-loop systems). While automation is life-changing for many, it can sometimes act as a “black box” that leaves the user disconnected from their own physiology. The iDSS model serves as a cognitive bridge, helping patients with lower health literacy to recognize patterns and gain the confidence needed for effective self-management.
This study suggests that DSS should not be viewed merely as a replacement for human decision-making, but as a scaffold for behavioral change. For a patient with a high A1c and limited knowledge, the transition to advanced technology may need to be preceded or accompanied by tools that prioritize education and feedback over mere automation.
Expert Commentary: Addressing the Distress Barrier
Medical technology often assumes a “rational actor” model of the patient—the idea that if we provide the right data, the patient will perform the right action. The findings by Pavan et al. debunk this assumption by highlighting the role of diabetes-related distress. If a patient is in a state of burnout, a DSS that provides more notifications or more tasks may actually exacerbate the problem, leading to technology abandonment.
Clinicians must screen for psychological readiness before prescribing complex decision-support tools. Addressing hypoglycemia worry through behavioral therapy or more conservative CGM alerts may be a necessary prerequisite to successfully implementing a prescriptive DSS. Furthermore, the preference data suggests that a one-size-fits-all approach to diabetes tech is destined to fail. Personalization must extend beyond the insulin algorithm to the user interface and the style of feedback provided.
Conclusion
The study “Human factors in the use and efficacy of decision support technologies for type 1 diabetes” provides a vital roadmap for the next generation of diabetes care. It demonstrates that the efficacy of digital health tools is inextricably linked to the user’s educational background and psychological state. While prescriptive systems offer convenience, informative systems may offer a more powerful path to long-term glycemic improvement by acting as a learning tool for those who need it most.
To unlock the full potential of these technologies, the clinical community must move toward a holistic model of care that integrates psychological support with technological intervention. Only by addressing the human element can we ensure that the promise of decision support is realized for all patients, not just those who are already highly engaged and knowledgeable.
Funding and References
This study was supported by various clinical research grants focused on diabetes technology. For further details on the trial design and full data sets, refer to ClinicalTrials.gov.
Reference:
Pavan J, Nass R, Fabris C, et al. Human factors in the use and efficacy of decision support technologies for type 1 diabetes: evidence from a randomized controlled trial. Diabetes Res Clin Pract. 2026 Jan;231:113049. doi: 10.1016/j.diabres.2025.113049. Epub 2025 Dec 10. PMID: 41380778.