Using Research Networks to Compare Medicines for Autoimmune and Inflammatory Diseases -- The Comparative Health Outcomes in Immune-mediated diseases CollaborativE (CHOICE) Study

• Download Summary
• Español (pdf)
• Audio Recording (mp3)

What was the research about?

In autoimmune and inflammatory illnesses, the immune system attacks healthy cells by mistake. Examples are rheumatoid arthritis, ankylosing spondylitis, and vasculitis. Treatments include two types of medicines that limit inflammation and reduce damage to joints and tissues. Questions remain about how safe and effective these medicines are.

In this study, the research team wanted to see if they could use PCORnet® data to look at the safety and effectiveness of these medicines. PCORnet is a research network. It merges health records, prescription, and health insurance data from health systems, including hospitals and clinics. The team used the data to compare rates of infection among patients with different autoimmune illnesses taking the two types of medicines. They looked at rates of shingles and hospital stays due to infection.

What were the results?

The research team had problems getting consistent and complete data. For example, prescription data were missing for patients in some health systems.

Among patients with different autoimmune illnesses, the research team found that patients taking the two types of medicines had similar rates of shingles and hospital stays due to infection. But the rates varied among patients with different autoimmune illnesses. For example, among patients with illness related to blood vessels and bones:

• Patients with vasculitis had the highest rate of shingles. Patients with ankylosing spondylitis had the lowest rate.
• Patients with vasculitis had the highest rate of hospital stays due to infection. Patients with ankylosing spondylitis had the lowest rate.

What did the research team do?

The research team used PCORnet data. The data were from patients taking medicines to treat autoimmune illnesses between 2016 and 2020.

Patients with autoimmune illnesses gave input throughout the study.

What were the limits of the study?

Many patients in the study didn’t have prescription or refill data recorded. Results may have differed with complete data.

Future research could look at ways to reduce missing prescription data.

How can people use the results?

Researchers can consider these results when using data from research networks. Doctors and patients can use these results when considering medicines for autoimmune illnesses.

Objective

To evaluate the feasibility of using clinical data from research networks to compare the effect of biologic and non-biologic medications on incidence of infections requiring hospitalizations and on incidence of herpes zoster associated with treatments among patients with autoimmune diseases

Study Design

This retrospective cohort study examined the feasibility of using clinical data obtained from PCORnet® to compare the effect of biologic and non-biologic medications among patients with autoimmune disease.

Researchers used PCORnet data for people treated for autoimmune diseases, including rheumatoid arthritis, juvenile idiopathic arthritis, psoriatic arthritis, psoriasis, ulcerative colitis, Crohn’s disease, ankylosing spondylitis, and vasculitis. The study included data from 204,247 patients with an autoimmune diagnosis participating in three PCORnet Clinical Research Networks. Clinical Research Networks consist of two or more healthcare systems, including hospitals, integrated delivery systems, and federally qualified health centers. Researchers identified patients with autoimmune diseases receiving biologic and non-biologic medications between 2016 and 2020. 

Researchers then examined the comparative risks of infections requiring hospitalization and the incidence of herpes zoster associated with treatments using biologic and non-biologic medications. They compared outcomes by disease cohort, such as patients with inflammatory bowel disease.

Patients with autoimmune diseases gave input on the study design and outcomes.

Results

Researchers had problems getting consistent and complete data from the health systems.

Across disease cohorts, patients treated with biologic and non-biologic medications had similar rates of hospitalization due to infection and similar rates of herpes zoster. Infection rates varied across disease cohorts. For example, researchers found that, among patients with musculoskeletal autoimmune diseases treated with biologic and non-biologic medications:

• Rates of hospitalization due to infection were highest among patients with vasculitis (9.7–11.5 events per 100 person years) and lowest among patients with ankylosing spondylitis (0.55–0.9 events per 100 person years).
• Herpes zoster rates were highest among patients with vasculitis (1.24–2.94 events per 100 person years) and lowest among patients with ankylosing spondylitis (0.21–0.33 events per 100 person years).

Among patients with inflammatory bowel disease treated with biologic and non-biologic medications:

• Rates of hospitalization due to infection were highest among patients with ulcerative colitis (6.1–7.2 events per 100 person years).
• Herpes zoster rates were similar among patients with ulcerative colitis and Crohn’s disease (0.41–0.5 events and 0.33–0.47 events per 100 patient years, respectively).

Limitations

Most prescription refill information was missing. Also, immunomodulatory prescription data were not uniformly available for patients with autoimmune diseases and varied between 14% and 68% across health systems. Results may have differed with more data.

Conclusions and Relevance

In this study, clinical data from PCORnet allowed researchers to study some elements of the comparative risks of biologic and non-biologic medications on rates of hospitalization due to infection and on rates of herpes zoster. Researchers found similar rates among patients with autoimmune diseases treated with biologic and non-biologic medications.

Future Research Needs

Future research could examine ways to mitigate issues with missing data related to biologic and non-biologic medications.