New research may prevent hospital readmission after heart failure

Heart failure is serious and causes many deaths. It is also a disease that affects many and for which much of the resources in healthcare are used. New studies at Halmstad University show that one in three patients who have been hospitalized with heart failure is admitted again within 100 days. The researchers can also contribute with knowledge about how many of these unplanned admissions can be predicted and prevented.

Woman in white, could be a medical doctor, talks to a woman in blue. Photo.

"The idea is that patients will receive better care at a lower cost."

Farzaneh Kobra Etminani

In heart failure, the risk of further cardiovascular problems is greatest within 100 days, such as new heart failure, angina, heart attack, stroke, and irregular heartbeat. A new study at Halmstad University shows which factors could be considered already at discharge, as these can be linked to an increased risk of new, unplanned hospital admission within 100 days. Being a woman reduces the risk, as does elevated blood pressure. Other factors that can be linked to an increased risk of readmission are age, certain hormone levels and other diseases such as Chronic obstructive pulmonary disease (COPD).

"These factors are examples of clear indicators for readmission," says Farzaneh Kobra Etminani, Associate Professor of Information Technology at Halmstad University and one of the researchers behind the current research.

Health data

The studies is an example of Halmstad University's research on information-driven care, where AI and data analysis are used to draw conclusions from health data.

There is a difference between planned and unplanned hospitalizations, Farzaneh Etminani points out. The researchers have therefore begun to develop an algorithm to predict which people are statistically at greater risk of being hospitalized.

Woman in black hair and brown jacket. Photo.

Farzaneh Kobra Etminani, Associate Professor of Information Technology at Halmstad University.

"This knowledge contributes to a clinical decision support system that can support healthcare professionals while they still have the patient with them. The system can "flag" a high risk of readmission, and then it is up to the doctor not to discharge the patient at all or to discharge the patient with, for example, more support measures and planned follow-up visits.”

The knowledge becomes a decision support that is part of the doctor's overall knowledge of the patient, combined with test results, the patient's history, experience and so on. In addition to the emotional aspects for the patient linked to being admitted to hospital, the decision support is about saving resources and directing resources correctly.

"The idea is that patients will receive better care at a lower cost. Unplanned admissions cost more than planned ones. It's not about avoiding patients being re-admitted, but about being able to focus on those who really need to be admitted. Who needs the care and resources the most?”

The decision support system is currently being implemented in healthcare together with a company that makes medical record systems.

"It's important that we find ways that the clinics think are best, so that new knowledge can be integrated into the care chain. At the end of the day, it's all for the patient’s benefit," says Farzaneh Kobra Etminani.

Work preventively

Markus Lingman, Chief Physician and Strategist at Region Halland and Adjunct Professor of Medicine at Halmstad University, believes that it is extremely important for healthcare to understand what will happen to a patient to be able to work preventively and take preventive measures in time.

"In addition, understanding what increases the risk, in this case of rehospitalization, means that we can target the preventive measures and become more accurate," says Markus Lingman and continues:

"Some factors can be influenced, while others, such as age, cannot be done much about. Knowing which factors are most important and possible to influence can make planning for patients much better so that the need for hospitalization might not be necessary."

More about the research

The study is based on 7500 patients over the age of 40 who were admitted with heart failure during the years 2017-2020 in Region Halland. Of the patients, 12 percent died and one in three patients was readmitted to hospital within 100 days of discharge.

Research on information-driven care at Halmstad University is often conducted in close collaboration with Region Halland.

Read the article Clinical characteristics at hospital discharge that predict cardiovascular readmission within 100 days in heart failure patients – An observational study External link, opens in new International Journal of Cardiology: Cardiovascular Risk and Prevention (IJCCRP)

The researchers behind the study are Jason Davidge, PhD student in Family Medicine, Lund University, Anders Halling, Professor of Family Medicine, Lund University, Awais Ashfaq, Postdoctoral Researcher in Signal and Systems Engineering, Halmstad University, Farzanhe Kobra Etminani, Associate Professor of Information Technology, Halmstad University, Björn Agvall, Specialist in Family Medicine and R&D supervisor, Region Halland.

Read the article The Price of Explainabitlity in Machine Learning Models for 100-Day Readmission Predicition in Heart Failure: Retrospective, Comparative, Machince Learning Study External link, opens in new window. in Journal of Medical Internet Research.

The researchers behind the study are Amira Soliman, Senior Lecturer, Halmstad University, Björn Agvall, Specialist in General Practice and R&D Supervisor, Region Halland, Kobra Farazaneh Etminani, Associate Professor in Information Technology, Halmstad University, Omar Hamed, Research Engineer, Halmstad University, Markus Lingman, Chief Physician and Strategist at Region Halland.