Practical Use of Electronic Health Records Among Diabetes Patients in Research

0
0
0

Practical Use of Electronic Health Records Among Patients with Diabetes in Scientific Research

Electronic health records (EHRs), defined as digital repositories of patient information collected during routine clinical care, have revolutionized epidemiological research by enabling large-scale, cost-effective analyses that reflect real-world clinical practices. Unlike traditional randomized controlled trials (RCTs), which are resource-intensive and often exclude broader patient populations, EHR-based studies provide insights into disease progression, treatment outcomes, and risk factor associations across diverse demographics. This approach, termed real-world study, leverages big data to inform clinical guidelines, drug safety monitoring, and public health policies. For chronic conditions like diabetes, EHRs offer longitudinal tracking of biomarkers, treatments, and complications, making them indispensable for understanding disease dynamics.

Key Advantages of EHRs in Diabetes Research

EHRs integrate multifaceted data, including demographics (age, sex, race), clinical measurements (HbA1c, blood pressure, lipid profiles), medical history, medication use, and billing information. This comprehensive data allows researchers to:

  1. Reflect Real-World Clinical Practice: EHRs capture routine care patterns, avoiding the selection bias inherent in RCTs, which often enroll healthier, more compliant participants.
  2. Enable Large-Scale Cohort Studies: For example, Sweden’s National Diabetes Register (NDR) includes over 450,000 patients with type 2 diabetes, enabling retrospective analyses of risk factors like smoking, HbA1c, and statin use on mortality and cardiovascular outcomes.
  3. Track Longitudinal Variability: Repeated measures of biomarkers (e.g., HbA1c fluctuations) provide insights into sustained control versus transient improvements, crucial for diabetes management.
  4. Reduce Costs: EHRs eliminate the need for expensive data collection processes, making research feasible for low-resource settings.

Global Applications of EHRs in Diabetes Studies

Europe: The Swedish National Diabetes Register

Sweden’s NDR, established in 1996, exemplifies successful EHR utilization. With data from 40,000 type 1 and 450,000 type 2 diabetes patients, the NDR links risk factors (blood pressure, HbA1c, lipids) to outcomes via national death and disease registries. A landmark study of 271,174 type 2 diabetes patients identified smoking, physical inactivity, elevated HbA1c, and non-use of statins as top mortality predictors. Patients maintaining target risk factor ranges exhibited mortality and cardiovascular risks comparable to the general population. However, the NDR has yet to explore the impact of biomarker variability over time, a critical gap for future research.

United States: Diverse EHR Ecosystems

  1. PCORnet and REACHnet: The National Patient-Centered Clinical Research Network (PCORnet) aggregates EHR data from 11 clinical networks, covering 100 million Americans. REACHnet, a PCORnet partner, facilitated a cohort of 67,544 type 2 diabetes patients, revealing inverse associations between HDL cholesterol and stroke risk, and between BMI and stroke risk (supporting the “obesity paradox”).
  2. Kaiser Permanente: Serving 12.3 million members, Kaiser’s integrated system enables studies on bariatric surgery outcomes, gestational diabetes, and drug efficacy. For instance, bariatric surgery was linked to diabetes remission and reduced microvascular complications, even after relapse.
  3. Louisiana State University (LSU) Hospital System: LSU’s EHRs from seven public hospitals tracked 35,406 low-income type 2 diabetes patients, highlighting racial disparities in cardiovascular risk control. Key findings included U-shaped mortality risks associated with BMI (e.g., higher mortality for BMI <30 or ≥35 kg/m² in Black patients) and differential impacts of HbA1c on heart failure and stroke.

China: Emerging EHR Infrastructure

China’s EHR adoption began in the early 2000s but faces challenges due to fragmented healthcare systems and population mobility. Regional initiatives, like a Guangzhou cohort of coronary heart disease patients, demonstrated associations between fasting glucose, BMI, and mortality. The National Metabolic Management Center, launched by the Chinese Association of Clinical Endocrinologists, aims to standardize diabetes care and EHR utilization. However, national registries remain underdeveloped, limiting large-scale studies.

Designing Target Trials Using EHRs

EHRs can emulate clinical trials by structuring data to address causal questions, such as drug effectiveness. Key steps include:

  1. Eligibility Criteria: Extract baseline data (diagnosis dates, lab results) to mirror trial inclusion/exclusion criteria.
  2. Treatment Assignment: Use propensity score matching to balance confounding factors between intervention and control groups.
  3. Follow-Up Periods: Define start (e.g., first prescription) and end points (e.g., death, outcome diagnosis).
  4. Outcome Validation: Link EHRs to registries (e.g., death records) to confirm endpoints.
  5. Handling Missing Data: Exclude baseline missing data; impute follow-up gaps using statistical methods.

Successful emulations include replicating the CAROLINA and LEADER trials, which assessed cardiovascular outcomes of glucose-lowering drugs. However, China’s high patient mobility and disjointed EHR systems complicate trial emulation, necessitating integrated platforms combining insurance claims, pharmacy data, and medical records.

Limitations of EHRs in Research

  1. Data Quality Issues: Inaccurate entries, inconsistent measurements, and misdiagnoses require rigorous validation (e.g., chart reviews).
  2. Missing Data: Unmeasured confounders (diet, physical activity) and loss to follow-up due to patient mobility limit causal inference.
  3. System Heterogeneity: Varied EHR formats across hospitals complicate data pooling. In China, provincial disparities in healthcare access further exacerbate this issue.
  4. Ethical and Privacy Concerns: Secure data sharing protocols are essential to protect patient confidentiality.

Future Directions and Conclusion

EHRs hold immense potential for diabetes research, particularly in studying biomarker variability, treatment adherence, and health disparities. The Swedish NDR and U.S. networks like PCORnet offer blueprints for scalable, collaborative EHR systems. In China, centralized registries and standardized EHR formats are critical for advancing real-world diabetes research.

Resource sharing and multi-institutional collaborations will enhance data richness and generalizability. Projects like China’s National Metabolic Management Center underscore the importance of integrating EHRs into routine care. As EHR systems evolve, they will increasingly support precision medicine, policy-making, and global health strategies, ultimately improving outcomes for millions of diabetes patients worldwide.

doi.org/10.1097/CM9.0000000000000784

Was this helpful?

0 / 0