When the disease process in rheumatoid arthritis patients has been brought under control with medication, the medication is usually tapered. Unfortunately, when tapering medication, there is always a risk that rheumatic symptoms will re-appear. Such a flare can usually be treated by increasing the dose of medication. To reduce the risk of a flare, UMC Utrecht has developed a predictive computer model based on artificial intelligence that is now being investigated in a clinical study.
Due to, amongst other things, the use of biologicals (these are drugs that inhibit the action of inflammatory proteins or immune cells in the body and consist entirely or partly of animal or human protein) many patients with rheumatoid arthritis (RA) have few or no disease-related complaints. However, biologics also have drawbacks such as an increased risk of infection, high costs and local skin reactions. Clinical epidemiologist and principal investigator of the clinical study Paco Welsing PhD (Department of Rheumatology and Clinical Immunology, UMC Utrecht) explains: “Several studies have shown that in many RA patients with low disease activity, symptoms can remain well suppressed even with a lower dose of biologicals. Sometimes medication can even be stopped completely. Unfortunately, there is always a risk that the RA symptoms will increase again, a so-called flare. Although these are usually easy to treat by increasing the medication dose, this can still temporarily cause considerable symptoms. It remains difficult for rheumatologists to predict in whom and when such a flare will occur. As a result, physicians (and patients) are often reluctant to discontinue or taper treatment with a biological. A better assessment of the risk of (further) tapering would therefore help ”
In order to reduce the risk of a flare, a multidisciplinary team at UMC Utrecht has developed a predictive computer model. Marianne Messelink MD is a research physician and PhD student at the Department of Rheumatology and Clinical Immunology and performs research on tapering of biologicals: “Our computer model predicts whether it is possible to responsibly taper a biological without causing a disease flare-up. The computer model uses information from the patient such as previous blood results, medication use and disease activity as digitally available in the electronic health record. With the support of a grant from ZonMw, we want to test whether the use of this computer model in clinical practice can actually reduce the number of flare-ups during the tapering of biologicals. Therefore, together with the Sint Maartenskliniek in Nijmegen, we have set up the PATIO trial”.
Participants in the PATIO trial (Prediction Aided Tapering In RA patients treated with biOlogicals) – in which 160 patients with RA will be followed for 18 months – visit the rheumatology outpatient clinic every 3 months, during which the possibility of further tapering of the biological is assessed. Prior to this visit, blood is taken. Such blood tests are standard in the treatment of RA. In addition, specifically for this study questionnaires are regularly administered. There are two groups in this study. The first group will gradually taper the biological agent in the usual manner. The second group will also use the computer model to predict whether it is safe to taper. If the model predicts that there is a high risk of a flare, then no further tapering will be performed. In both groups, of course, if a flare occurs, tapering will be stopped. Eventually (probably in 2024) the results of the study should provide insight into whether the predictive model has added value (and actually showing that fewer flares occurred in the group in which the model was used).
Large data sets (“big data”), artificial intelligence and digital technology are driving innovation in science and healthcare. They make it possible to provide personalized care, both medically and in terms of location. Data scientist O’Jay Medina (Department of Digital Health) summarizes the process: “By collecting large amounts of data, which is interpreted by complex mathematical models (algorithms), our physicians can arrive at well-founded advices. In the case of the PATIO study: can we safely taper the biological agent used in a particular patient? And if so, to what dosage? The model’s predictions, the patient’s medication and disease activity are displayed in an integrated ‘dashboard’, which we developed together with the Dutch data analytics and technology company ORTEC.”
“It’s great to see that we are getting better at developing digital applications and actually applying them in healthcare. The PATIO study is an excellent example where this has been achieved, which we can be really proud of. It is a good example of how we can use artificial intelligence. In this case, in developing customized treatments for patients with RA. The result is fewer side effects of medication and therefore better quality of care at lower costs. A great example of how 1+1 can be 3,” says Remco van Lunteren, member of the board of directors. “Developments like these have enormous potential, therefore we at UMC Utrecht will continue to invest in them.”
Rheumatoid arthritis (RA) is a chronic autoimmune disease in which the immune system turns against a patient’s own body, resulting in joint inflammation. The course of RA is often variable: periods in which patients suffer from joint inflammation alternate with periods in which they experience little or no symptoms. Although, partly due to new treatment possibilities, many patients are well controlled, about 20 percent of patients repeatedly fail to respond adequately to medication. UMC Utrecht is an expert center for treatment of these difficult-to-treat forms of RA and conducts scientific research to identify causes and new treatment options for these patients. The goal is to provide each patient with a personalized treatment plan. In the Netherlands, approximately 260,000 people have RA. The UMC Utrecht sees 1200 patients with RA every year, of which about 100-200 with a difficult to treat form.
