Our results indicate that, despite the prescription of novel antidiabetic drugs, still half of participants treated for diabetes do not achieve adequate control. Our results also suggest that participants with newly diagnosed diabetes achieve much better control than participants with established diabetes.

The decrease in sulfonylureas in our study is consistent with several other studies [11,12,13,14,15] and may be due to the growing awareness of hypoglycaemia and cardiovascular risk associated with these drugs [16]. Similarly, the decrease in thiazolidinediones in our study is consistent with other studies [11, 12, 14]. The slight decrease in insulin therapy in our study contrasts with increased use in other studies [12, 13] and might be due to difficulties in controlling diabetes, as indicated by the relatively low control rates consistently found in all surveys. Biguanides decreased (for individual drugs) or remained stable (for antidiabetic drug class) in our study, whereas they increased in other studies [11,12,13,14,15]. As most national and international guidelines propose metformin as a first-line drug treatment for T2D [6, 7], the reasons for this decrease are hard to explain. The increase of DPP4 or SGLT2 inhibitors and of GLP1 analogues in our study is consistent with several other studies [11,12,13,14,15], although in one study the use of DPP4 inhibitors tended to decrease [14]. Still, an increase in the association of biguanides with DPP4 inhibitors was found, which agrees with the international guidelines [7]. Conversely, the low association of GLP1 analogues with biguanides was unexpected; a possible explanation would be that SGLT2 inhibitors were preferred to GLP1 analogues, as both combinations (metformin + GLP1 analogues or metformin + SGLT2 inhibitors) are recommended by the Swiss Society for Endocrinology and Diabetes [6]. Overall, our results suggest that the Swiss guidelines tend to be implemented in clinical practice, although among a relatively small number of people.

Hence, and as observed elsewhere, our results indicate that in Switzerland, the prescription of sulfonylureas and thiazolidinediones is decreasing, being replaced by GLP1 analogues, SGLT2 and DPP4 inhibitors. Although the combination of DPP4 and SGLT2 inhibitors with metformin increased, the overall use of metformin- and insulin-based therapies decreased, in contrast to other studies [12, 13]. Differences in outcomes may stem from variations in clinical guidelines, prescribing practices, cost-effectiveness and regional healthcare priorities. Sociodemographic and patient-related factors, such as insulin resistance and comorbidities, also influence treatment choices.

Despite the increasing availability of new antidiabetic drugs, there was little to no improvement in diabetes control, with one half of treated participants not achieving adequate FPG levels. Our findings are in agreement with the literature [4, 17,18,19,20,21,22]. Poor medication adherence contributes to lower diabetes control. A review of 71 studies estimated that only half (51.2%) of patients adhere to their medication [23]. Poorer diabetes control (higher HbA1c levels) has been shown to be associated with earlier intensification of treatment [24, 25]. The most poorly controlled diabetics are already under stronger medication, but their control may remain poor despite increased treatment.

Little is known how management of newly diagnosed diabetes and risk factor control have evolved over time. In our study, diet and BMI did not differ between newly diagnosed and established diabetics, suggesting that lifestyle changes, such as increased physical activity and weight loss, are not effectively implemented to control blood glucose levels. People newly diagnosed with diabetes may be more motivated to adhere to recommended medications, while those with long-term diabetes may relax their monitoring. Nevertheless, in a systematic review of 27 studies, duration of diabetes was not associated with medication adherence [26].

Maintaining long-term treatment for diabetes is essential to prevent cardiovascular complications [27] and poor medication adherence is associated with increased HbA1c levels, emergency department visits and hospitalizations [28, 29]. Side effects, mainly gastrointestinal and weight gain, hinder adherence to treatment [30]. Patients also lack information to better understand their condition, as well as the benefits and risks of treatment [31]. Our results show that, despite the availability of novel, potent antidiabetic drugs, no significant improvement in diabetes control was found. Hence, a closer monitoring of patients with diabetes should be performed, focusing on a healthy lifestyle, weight control, and adequate compliance to treatment.

Our results also show that participants with newly diagnosed diabetes had better blood glucose control. Clinicians should explore this window of opportunity to provide guidance regarding lifestyle changes and patient-centred care that addresses individual concerns and needs, to increase the likelihood of adequate diabetes control in the future. Metformin, as a first-line therapy for T2D, is widely endorsed by clinical guidelines due to its efficacy, safety profile, and cost-effectiveness [6, 7]. Deviations from these guidelines raise concerns about clinicians’ preferences, patient factors, or gaps in knowledge. Addressing non-adherence could improve patient outcomes by prioritizing evidence-based treatments, reducing complications, and optimizing resources.

Contrary to expected, participants on insulin or on the newer and more potent antidiabetic drugs such as GLP-1 analogues or SGLT2 inhibitors did not present with a better control of their condition. Possible explanations include a reverse causation, those drugs being prescribed to participants with difficulties in controlling their diabetes, or to the small number of participants receiving those drugs, leading to a low statistical power. Further, both DPP4 and SGLT2 inhibitors have a lower risk of hypoglycaemia and potential cardiovascular and renal protective effects, while their impact on glycaemic control may not be as potent as that of sulfonylureas, particularly in patients who are more insulin-resistant or those with advanced diabetes.

There are several strengths to this study. Firstly, the importance of different antidiabetic drugs in the management and control of diabetes in Switzerland was analysed over a period of eighteen years. Secondly, a wide range of socio-demographic covariates were analysed to better understand the facilitating and inhibiting factors for diabetes control.

There are also some limitations to this study. Firstly, it was conducted in a single location, and results might not apply to the entire country or to other settings. Still, the poor control of blood glucose levels is consistent with a previous study conducted in Geneva [22], and studies abroad [32, 33], suggesting that better control among patients with newly diagnosed diabetes might be a general trend. Secondly, the sample size was relatively small, which would have reduced the statistical power. It would be interesting to replicate our study in a larger sample. Thirdly, it was not possible to assess the exact posology of the antidiabetic drugs or the compliance of the participants towards their treatment. This could lead to a possible information bias, with some participants being considered as treated while not taking their medication. Fourthly, other comorbidities beyond hypertension, such as renal impairment, were not considered, yet they can impact the pharmacokinetics and efficacy of certain diabetes medications, influencing treatment outcomes. Fifthly, although the number of excluded participants was small in the first two surveys, it increased significantly in the other two. This might lead to an information bias if excluded participants had different treatment and control levels or received different antidiabetic medications than included ones. Still, as the characteristics of included and excluded participants were globally similar, it can be expected that those differences in management, if present, would be minor. Notwithstanding, it would be of importance if other studies could replicate our findings. Finally, we relied on FPG levels < 7 mmol/L to define diabetes control, which is a rather high level. Had we considered a lower level, the control rates would have been even lower than reported.

We conclude that, in a population-based sample of French-speaking Switzerland, at least half of participants treated for diabetes do not achieve adequate control, despite the availability of novel antidiabetic drugs. Participants with newly diagnosed diabetes achieve better control than participants with established diabetes.