Strengths and limitations
The strength of this study lies in its utilisation of real-world data obtained from primary healthcare settings, providing insights that can better reflect routine clinical practices. Incorporating socioeconomic factors, such as country of origin and income, adds depth to the analysis by recognising elements beyond biological factors that may influence insulin prescription decisions. This comprehensive approach acknowledges the multifaceted nature of healthcare decision making.
The main limitation of this study is the inability to exclude other diabetes subtypes that necessitate early insulin prescription. Certain diabetes subtypes, specifically late autoimmune diabetes in adults (LADA), are often misdiagnosed as T2DM, present in people with normal BMI, and may require insulin therapy sooner than in people with T2DM.30,31 In this study there were challenges in identifying individuals with possible diagnoses of LADA owing to the absence of specific coding in the dataset in which they are usually classified as individuals with T2DM. Recognising and distinguishing these subtypes is crucial, as their unique clinical characteristics may influence both the timing and necessity of insulin prescription. The observed inverse relationship between a higher BMI and a reduced likelihood of insulin prescription revealed a weight-related consideration in clinical decision making.32,33 Although the well-documented phenomenon of insulin-induced weight gain offers a plausible explanation for this association, it is essential to acknowledge that lower BMI may also be indicative of LADA that necessitates earlier insulin initiation.34
Specific ethnic background was not available in the current dataset; therefore, country of origin was used as a proxy for ethnic group. Although this approach provides some insight into cultural and demographic differences, it may not fully capture the complexity of ethnic background. Additionally, the study considered one measurement for each covariate at index time, even though many of them are dynamically changing over time, such as HbA1c levels. Moreover, SGLT2 (sodium glucose co-transporter 2) inhibitors and GLP-1 (glucagon-like peptide-1) receptor agonists have recently begun receiving reimbursement in the Netherlands, which has led to their increased use in clinical practice. Thus, the use of SGLT2 inhibitors and GLP-1 receptor agonists was not accounted for in this study.
This study was observational in nature and aimed to explore factors associated with the initiation of insulin therapy. The study was not designed to establish causality, and the findings should be interpreted with caution as observational associations. Although the study adjusted for a range of relevant variables, some factors — such as comorbidities, use of other medications, and alcohol consumption — were not included in the analysis. Future studies could address these limitations by including more detailed information on ethnic group to better understand cultural influences on treatment decisions. They could also consider changes in patient characteristics over time and account for the use of newer medications such as SGLT2 inhibitors and GLP-1 receptor agonists.
Comparison with existing literature
Glycaemic control directly, and diabetes-related complications indirectly, play a role in insulin initiation for individuals with T2DM.35 Poor glycaemic control, indicated by elevated HbA1c levels, aligns with an increased likelihood of transitioning to insulin therapy according to established clinical guidelines.2,22,36 Timely initiation of insulin is often indicated when oral antidiabetic agents fail to achieve adequate glycaemic control. The HbA1c target for individuals in this study cohort (ages 40–79 years) depends on age, disease duration, and treatment stage. For patients aged <70 years, the recommended target is ≤53 mmol/mol. In patients aged >70 years, the target remains ≤53 mmol/mol during the initial stage of medication, but increases to 54–58 mmol/mol for those with a disease duration of <10 years, and 54–64 mmol/mol for those with a disease duration of ≥10 years.22 Maintaining HbA1c below these thresholds signifies better glycaemic control and reduces complications associated with elevated blood glucose levels. As expected, an elevated HbA1c was associated with a 54% higher hazard of insulin initiation within 5 years. These findings are consistent with other studies observing elevated baseline HbA1c levels in the group that initiated insulin.13,20,37 This emphasises the importance of glycaemic control shortly after commencing treatment with metformin in predicting treatment trajectory and diabetes progression in individuals with T2DM.
Additionally, eGFR serves as a critical biomarker affecting insulin prescription.38 Lower eGFR indicates impaired kidney function, a common complication in individuals with diabetes. Providers may initiate insulin therapy earlier in individuals with compromised renal function, especially when other antidiabetic medications are contraindicated because of critically low eGFR levels, to ensure safe and effective treatment.39 In the current analysis, a higher eGFR was associated with a decreased probability of using insulin, in concordance with the anticipated outcome.
The observed correlation between younger current age and the risk of early insulin initiation aligns with the guidelines’ age consideration in targeted HbA1c levels and the existing literature on the natural progression of T2DM.40 Individuals diagnosed at a younger age, such as in their early 40s, often experience a more aggressive form of diabetes, leading to greater beta-cell dysfunction and increased insulin resistance over time.41,42 Additionally, challenges in changing longstanding dietary habits can exacerbate the condition. Prolonged exposure to poor dietary choices, often established early in life, may lead to persistent difficulties in glycaemic control, necessitating more intensive therapeutic interventions, such as insulin, at an earlier stage.43
Factors beyond glycaemic control and comorbidities can also influence insulin initiation in a primary care setting. Individuals with a migration background were less likely to receive insulin prescriptions in this study, despite research indicating that ethnic minorities, including Surinamese, have a higher prevalence of the disease, poorer glycaemic control, and an increased risk of complications.44 These patients also tend to experience more severe disease progression at a younger age compared with other ethnic groups, which typically contributes to a higher likelihood of earlier insulin initiation in these populations.45,46
The complexity of tailoring treatment plans for diverse patient populations is evident when ethnicity intersects with other variables. In diabetes management, ensuring patient adherence to insulin therapy is critical.47 Healthcare providers must assist patients in understanding and following insulin use instructions, as misuse poses serious risks. Such instructions require ample time for patients with limited health literacy and language barriers.11 The National Dutch General Practitioners Association recommends special consideration for certain patient groups, such as those with limited education, limited health literacy, and migration backgrounds, when planning insulin initiation.22 This approach can introduce unconscious bias when providers become hesitant to initiate insulin among such groups, even if they have elevated HbA1c levels. Clinical inertia, which refers to the delay in initiating or intensifying treatment despite established guidelines, plays a significant role in the decision-making process for initiating insulin therapy in this study.48 Ethnic disparity in the healthcare system is common worldwide.49 For instance, in high-deprivation areas such as Glasgow, South Asian patients are less likely to receive insulin compared with other ethnic groups.50,51 Studies in the US and UK using cross-sectional designs and randomised control trials also demonstrate a decreased likelihood of insulin prescriptions among those from ethnic minority groups.18,20,52 These disparities are not solely because of variations in illness duration or control, suggesting deeper cultural or physician perceptions of different ethnic groups.53 Insulin initiation is often delayed until significant diabetes-related complications arise, potentially exposing disadvantaged patient groups to higher risks.17
Furthermore, the current study found that patients with lower income were more likely to receive insulin prescriptions within 5 years after initiation of metformin. It is well established that individuals’ SES significantly influences their health and overall quality of life.54 For those with T2DM, lower SES is linked to higher rates of complications and mortality.55 This relationship is attributed to the tendency of individuals with limited education and social class achievements to require higher insulin doses and struggle with controlling their blood sugar levels, compared with peers who are better educated and with higher SES. The prevalence of obesity, smoking, and physical inactivity is higher among low-income individuals.56 This, combined with possible limited access to healthcare services and lower health literacy, can lead to poor glycaemic control and the need for insulin treatment earlier.49
The seemingly contradictory findings that individuals with migration backgrounds are less likely to receive insulin prescriptions, coupled with the observation that patients with lower SES are more likely to receive insulin treatment, raises important considerations regarding potential barriers in healthcare delivery. One of these barriers could be the language barrier, as physicians may encounter challenges in effectively communicating insulin initiation plans with patients with limited proficiency in the dominant language. Studies have shown that language barriers can hinder patient–provider communication, leading to misunderstandings about treatment recommendations and reduced adherence to therapy instructions.57 This issue is particularly relevant in culturally diverse healthcare settings, where patients from various ethnic backgrounds may have different linguistic preferences and levels of health literacy.58 Therefore, addressing language barriers through targeted interventions in future studies, such as interpreter services or culturally sensitive communication strategies, is essential to ensure equitable access to insulin therapy and optimise diabetes management outcomes.
