Impact of Early Diagnosis and Pharmacological Management of Prediabetes and Diabetes on the Middle- and Long-term Outcomes for Patients Attending Australian General Practices

Abstract

Background Early diagnosis and management of prediabetes and diabetes have been top priorities for Australian primary care over the past decades. However, there is a lack of national evidence about whether activities undertaken by Australian general practitioners (GPs) regarding screening, diagnosis, and management of prediabetes and diabetes are consistent with current guidelines. Aims This thesis investigates (1) the epidemiology of diabetes and prediabetes in general practice and whether people at higher risk of diabetes are more likely to be screened for diabetes than those not at risk; (2) differences in monitoring and control of clinical parameters in patients with past or newly recorded diabetes; (3) whether patients with a recent diabetes diagnosis achieve better glycaemic control with early metformin therapy compared with delayed pharmacological management; and (4) whether patients with prediabetes managed with metformin achieve better glycaemic control than those not receiving that medication. Methods The four studies in this thesis used a national electronic health record (EHR) database (MedicineInsight) containing data on diagnoses, laboratory results, and prescriptions, collected between 2011 and 2018 from 662 general practices across Australia. To attend to the first thesis aim, we used a cross-sectional design that projected the prevalence of undiagnosed or diagnosed diabetes and prediabetes in general practice (2016–2018), and explored the sociodemographic and clinical profile of diabetes screening among patients at risk of diabetes. For the second aim, we used a retrospective cohort design that allowed us to identify 101,875 ‘regular’ patients (at least one consultation each year from 2015 to 2018) with past recorded diabetes (diabetes recorded in 2015 and/or 2016) and 9,236 with newly recorded diabetes (diabetes recorded in 2017 but not in previous years). Based on laboratory results reported in 2018, two groups of outcomes were assessed: (1) diabetes monitoring, based on whether a result for HbA1c, blood pressure (BP), total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglycerides, estimated glomerular filtration rate, or albumin-to-creatinine ratio was available in the EHR; and (2) ‘well-controlled’ diabetes, defined as HbA1c ≤7.0%, BP ≤140/90 mmHg, and total cholesterol <4.0 mmol/L. Differences in the frequency of these binary outcomes between those with past or newly recorded diabetes were examined using logistic regression models. For the third aim, we again used a longitudinal cohort design and included data from 27,027 ‘regular’ patients with incident diabetes (at least three consultations, including one the year before and one the year after the first recorded diabetes diagnosis) who were first diagnosed with diabetes between 2012 and 2017. Augmented inverse probability weighting (AIPW) was used to estimate the average treatment effect (ATE) of early (<3 months), timely (3–6 months), delayed (6–12 months), or no metformin prescription within 12 months of diagnosis on glycaemic parameters (HbA1c or fasting blood glucose [FBG] levels). The ATE was estimated at 3–6, 6–12, 12–18, or 18–24 months after exposure or diagnosis. For the fourth aim, we used a similar approach, but the cohort included 4,770 regular patients with incident prediabetes diagnosed between 2012 and 2017. AIPW was used to estimate the ATE of metformin prescription on glycaemic parameters (HbA1c or FBG levels), at 6–12, 12–18, or 18–24 months after exposure. Results Paper 1 (addressing thesis aim 1): 7.5% (95%CI 7.3;7.8) of adult patients attending Australian general practices had a recorded diagnosis of diabetes, 0.7% (95%CI 0.6;0.7) of prediabetes, and 0.3% (95%CI 0.3;0.3) had unrecorded diabetes/prediabetes (elevated glucose levels without a recorded diagnosis) during 2016–2018. Patients with unrecorded diabetes/prediabetes had clinical characteristics similar to patients with recorded diabetes, except for a lower prevalence of overweight/obesity among the former (55.5% and 69.9%, respectively). Dyslipidaemia was 1.8 times higher (36.2% vs 19.7%) and hypertension was 15% more likely (38.6% vs 33.8%) among patients with prediabetes than those with diabetes. The rate of diabetes screening in the past 3 years among people at high risk of diabetes was 55.2% (95%CI 52.7;57.7), with lower rates among young or elderly males, patients with prediabetes, or patients who were prescribed antipsychotics. Paper 2 (addressing thesis aim 2): In 2018, HbA1c was monitored in 45.2% (95%CI 42.6;47.7) of patients with past diabetes and 39.4% (95%CI 37.1;41.7) of patients with recent diabetes (adjusted odds ratio 0.78, 95%CI 0.74;0.83). Monitoring of HbA1c, BP, and total cholesterol levels was no better among smokers, or patients with hypertension or cardiovascular disease (CVD) than among patients without these risk factors. HbA1c control was achieved by 54.4% (95%CI 53.4;55.4) and 78.5% (95%CI 76.8;80.1) of monitored patients with past and recent diabetes, respectively (adjusted odds ratio 3.11, 95%CI 2.84;3.41). Irrespective of whether they had past or newly recorded diabetes diagnosis, less than 20% of patients had all three clinical parameters controlled (i.e., HbA1c, BP, and total cholesterol levels). Patients with a history of CVD were more likely to have the three clinical parameters controlled than those without a history of CVD, especially among those with newly (adjusted odds ratio 2.43, 95%CI 1.85;3.19) rather than past recorded diabetes (adjusted odds ratio 1.39, 95%CI 1.30;1.49). Paper 3 (addressing thesis aim 3): Compared with patients with incident diabetes who were not managed with metformin (i.e., the group with the lowest baseline glycaemic levels), the corresponding ATE for HbA1c at 18–24 months was +0.04% (95%CI –0.05;0.10) for early treatment, +0.24% (95%CI 0.11;0.37) for timely treatment, and +0.29% (95%CI 0.20;0.39) for delayed treatment. Similar results were observed for FBG levels. Paper 4 (addressing thesis aim 4): Despite having higher baseline HbA1c levels, patients with incident prediabetes who were managed with metformin had similar mean HbA1c levels at 6–12 months (ATE 0.00, 95%CI −0.05;0.05) or 12–18 months (ATE −0.02, 95%CI −0.09;0.06) as patients not managed with antidiabetic medications. However, at 18–24 months, patients with prediabetes who received metformin had lower mean HbA1c levels (ATE −0.09, 95%CI −0.16;0.00) than those who were unexposed. The analysis of FBG levels provided consistent results. Conclusions and clinical implications In Australian general practice, diabetes screening among high-risk populations can be improved for patients with prediabetes and those treated with antipsychotics, as these people visit their GP on average five times per year. Moreover, the monitoring of clinical parameters among patients with diabetes is currently suboptimal, as only half of the patients with diabetes had a record of their blood glucose levels being checked over the preceding 12 months. Additionally, 80% of all patients monitored did not reach the recommended HbA1c, BP, and total cholesterol targets. Diabetes screening among high-risk individuals, and monitoring and reaching target levels of critical clinical parameters are essential to minimise the health and economic impact of diabetes progression. International initiatives show it is feasible to improve diabetes screening, monitoring, and management, as these activities could be performed during the annual diabetes ‘cycle of care’ in general practice. In primary care settings, early metformin therapy helped patients with diabetes achieve better and more stable glycaemic parameters. Furthermore, metformin therapy for patients with incident prediabetes with high baseline glycaemic levels could help prevent further deterioration of their glycaemic parameters. This finding may influence diabetes and prediabetes management guidelines, as none of them currently recommend using antidiabetic medications to treat prediabetes.