Project Summary
This research project is in progress. PCORI will post the research findings on this page within 90 days after the results are final.
One of PCORI’s goals is to improve the methods that researchers use for patient-centered outcomes research. PCORI funds methods projects like this one to better understand and advance the use of research methods that improve the strength and quality of comparative effectiveness research.
What is the project about?
Medicine nonadherence occurs when patients don’t take medicine as directed. If patients take the wrong dose or forget a dose, medicines may not work well or may cause health problems.
Many research studies assume that patients take their medicine as directed. Others make assumptions about how and when nonadherence happens. For example, studies might use data from electronic health records, or EHRs, that capture when and how much medicine a patient received. But EHRs don’t show whether a patient took his or her medicine, or if the patient took it as directed.
The research team is developing and testing new methods to help researchers who use EHR data avoid making incorrect assumptions about whether patients use medicines as directed after filling a prescription.
How can this project help improve research methods?
Results may help researchers improve studies that use EHR data to compare how well medicines work, and help them better understand the health effects of nonadherence.
What is the research team doing?
The research team is developing new methods that build on methods researchers currently use to estimate the effect of medicine refill patterns on patients’ health in studies using EHR data. The team is testing these new methods on a data set they create to see how well the methods work compared with existing methods. Then the team is testing the new methods using real EHR data from two studies comparing medicines for type 2 diabetes. Finally, the team is developing computer programs so that other researchers can use the new methods.
Research methods at a glance
| Design Elements | Description |
|---|---|
| Goal |
Aim 1. Develop new methods for studying the health effects of prescription fill and refill regimens in studies using EHR data with time-dependent confounding. Aim 2. Evaluate and compare the new methods with existing methods using simulation and analyses of data from two EHR-based studies of type 2 diabetes. Aim 3. Create and distribute free software that automates analyses of EHR data using the new methods. |
| Approach | Theoretical derivations, simulation, secondary analyses, software development |
COVID-19-Related Study
Comparing the Effect of Medicines to Reduce Blood Sugar Levels on COVID-19 Risk among Patients with Diabetes
Results Summary
In response to the COVID-19 public health crisis in 2020, PCORI launched an initiative to enhance existing research projects so that they could offer findings related to COVID-19. The initiative funded this study and others.
What was this COVID-19 study about?
Type 2 diabetes is a long-term illness that causes blood sugar levels to rise. Diabetes increases patients’ risk for severe illness from COVID-19. Medicines to keep blood sugar levels low might reduce the risk for severe COVID-19. But questions remain about when to start such medicines.
In this study, the research team compared the effect of starting a new medicine at one of four blood sugar levels on risk for COVID-19 and other risks. The blood sugar levels were 7 percent, 7.5 percent, 8 percent, and 8.5 percent. For each level, the team looked at data from electronic health records, or EHRs, to see the number of:
- COVID-19 infections
- Patients who had a hospital stay due to COVID-19
- Patients who used a ventilator due to COVID-19
- Patient deaths in the hospital due to COVID-19
The research team did the comparisons using two statistical methods: inverse probability weighting, or IPW, and targeted learning, or TL. The methods imitate randomized controlled trials using EHR data.
What were the results?
Patients who started a medicine at lower blood sugar levels had a lower risk of getting COVID-19 compared with patients who started a new medicine at higher blood sugar levels or who didn’t start a new medicine. The TL and IPW methods showed similar results.
Using both TL and IPW, this study found that starting a medicine didn’t affect patients’ risk for a hospital stay due to COVID-19.
Who was in the study?
The study included EHR data for 7,199 patients with type 2 diabetes. All received care from a health system in California. Among patients, 41 percent were White, 25 percent were Asian, 5 percent were Black, and 5 percent were another race; 24 percent were Hispanic or Latino. The average age was 59, and 59 percent were men.
What did the research team do?
The research team reviewed EHR data from January 1, 2018, to May 31, 2021. They identified patients who received medicines to lower their blood sugar. Patients’ blood sugar levels were at one of four levels when they started an additional medicine to lower blood sugar. The team compared those who started a medicine to patients who didn’t start a medicine. The team took into account patient traits that may affect COVID-19 outcomes, such as:
- Vaccination status
- Other health conditions
- Neighborhood income levels
Patients with type 2 diabetes, doctors, pharmacists, and health system administrators helped plan the study.
What were the limits of the study?
Few patients required ventilators or died from COVID-19. As a result, the research team couldn’t determine the effect of medicines for these outcomes.
How can people use the results?
Patients and their doctors can use the results when considering treatments for type 2 diabetes to reduce the risk of severe COVID-19.
Professional Abstract
In response to the COVID-19 public health crisis in 2020, PCORI launched an initiative to enhance existing research projects so that they could offer findings related to COVID-19. The initiative funded this study and others.
Background
Patients with type 2 diabetes are at increased risk for developing severe COVID-19 and experiencing poor health outcomes. Better blood glucose control may reduce the risk for and severity of COVID-19 among patients with diabetes. One way to improve blood glucose control is to initiate treatment with an additional glucose-lowering medication that is not part of a patient’s current treatment. However, questions remain about the appropriate hemoglobin A1c level to start treatment and if it would reduce the risk for and severity of COVID-19 among patients with diabetes.
Objective
To compare the effectiveness of initiating intensified glucose-lowering medication at lower versus higher blood glucose levels on the risk of COVID-19 and COVID-19-related outcomes in adult patients with type 2 diabetes
Study Design
| Design Element | Description |
|---|---|
| Design | Retrospective cohort study |
| Population | EHR data for 7,199 adult patients with type 2 diabetes who were taking 2 or more oral glucose-lowering medications or basal insulin, and who had elevated A1c levels between 7% and 8.5% following a period of controlled glycemia (A1c<7%) under the baseline therapy |
| Interventions/ Comparators |
|
| Outcomes |
Infection with COVID-19; hospitalizations due to COVID-19 infection; use of ventilator while hospitalized due to COVID-19; deaths due to COVID-19 while hospitalized |
| Timeframe | Patient EHR data from January 2018 to May 2021 |
This retrospective cohort study compared the effect of initiating intensified glucose-lowering medication at lower versus higher blood glucose thresholds on the risk of developing COVID-19 and on other COVID-19-related outcomes, including hospitalization, ventilator use, and death among patients with type 2 diabetes.
Researchers reviewed electronic health record (EHR) data for patients who received care for type 2 diabetes between January 1, 2018, and May 31, 2021. Researchers compared outcomes among patients who started additional glucose-lowering medications based on one of four blood glucose thresholds versus patients who did not initiate additional glucose-lowering medications.
To estimate the effect of intensifying blood glucose therapy on COVID-19 outcomes, researchers employed two methods used in another PCORI-funded study: inverse probability weighting (IPW) and targeted learning (TL). Using EHR data, these two methods imitate the results of randomized controlled trials and control for confounding and selection bias. In this study, researchers controlled for patient characteristics that could influence COVID-19 outcomes, including vaccination status, comorbidities, test results for cholesterol and kidney function, and neighborhood income levels.
The study included EHR records for 7,199 adult patients with type 2 diabetes from one health system in California. Of these patients, 41% were White, 25% were Asian, 5% were Black, and 5% were another race; 24% were Hispanic or Latino. The average age was 59, and 59% were male.
Patients with type 2 diabetes, clinicians, pharmacists, and health system administrators helped plan the study.
Results
Overall, intensifying medication at lower A1c levels reduced the risk for COVID-19 infection compared with intensifying medication at higher A1c levels or not intensifying medication. For example, using TL, compared with intensifying medication at A1c≥7.5%, intensifying medication at A1c≥8%, at A1c≥8.5% and not intensifying medication progressively increased the risk of COVID-19 infection by 8.1% (95% confidence interval [CI]: 5.4%, 10.8%); 9.9% (95% CI: 6.8%, 13.0%); and 16.1% (95% CI: 11.6%, 20.6%), respectively. IPW results showed similar effects but were not statistically significant.
Intensifying medication did not significantly affect risk for hospitalization due to COVID-19 using both TL and IPW.
Limitations
Because few patients required ventilators or died from COVID-19, researchers could not assess the effect of intensifying medication on these outcomes.
Conclusions and Relevance
This study provided evidence for initiating intensified glucose-lowering therapy at lower A1C levels to prevent COVID-19 infection.
Peer Review Summary
The Peer-Review Summary for this COVID-19 study will be posted here soon.
Final Enhancement Report
This COVID-19 study's final enhancement report is expected to be available by August 2023.