Improving the Delivery of Health Services at the Health Systems Level

AHRQ-funded research aims to improve the delivery of health services at the health systems or organizational level. Investment in research to improve the delivery of healthcare at the systems level was $36 million over the duration of projects that were ongoing in 2019. Efforts to share health information across technologies and healthcare environments and to leverage data and technology to strengthen the service quality are key aspects of research projects focused on health systems.

 

Telehealth and Telemedicine

Telehealth and Telemedicine facilitate long-distance care, education, and monitoring. The following are examples of AHRQ-funded 2019 telehealth and telemedicine research studies that ended in 2019:

 

Dr. Charles Ellis conducted an evidence-based speech therapy intervention that uses tele-rehabilitation approaches to improve communication for stroke patients with communication impairments (i.e., aphasia). The proof-of-concept evaluation revealed the feasibility, acceptability, and satisfaction of this technological approach to speech therapy.

 

Dr. Elizabeth Ferucci designed an observational study to evaluate disease activity and care quality among patients with rheumatoid arthritis. The study examined patient-reported symptoms in those who had experienced at least one telemedicine visit during the 12-month study period compared to those who had not. The findings demonstrated that patient outcomes were similar for in-person and telemedicine visits, suggesting that telemedicine is an acceptable method of followup.

Read Dr. Ferucci's Impact Story

 

Artificial Intelligence

Artificial Intelligence (AI) refers to the use of algorithms by computers to simulate human intelligence and improve performance. The use of AI in the healthcare industry provides considerable opportunity for technological advances. The following AHRQ-funded projects explore the capability of AI to enhance health system processes:

 

Dr. Jayant Pratap developed computerized models for predicting surgery cancellation. The aim of this last-minute surgery cancellation prediction system was to reduce healthcare costs and improve overall efficiency. The research found that the prediction system was able to predict two causes of cancellation (“no shows” and patients who cancelled because they had eaten) at higher rates than cancellations that occurred when patients were ill or because of patient or family refusal.

 

To enhance the clinical diagnosis of sepsis, Dr. Robert Sherwin developed Intelligent Sepsis Alert (ISA), an AI-enhanced version of an existing sepsis CDS tool. The tool integrates a real-time alert module into the EHR to identify patients with possible sepsis, resulting in enhanced performance from the existing sepsis CDS tool, further improving sepsis management.

 

Dr. Alexander Turchin investigated the outcome of combining natural language processing (NLP) and Dynamic Logic, a machine-learning algorithm to improve the identification of patients at high risk of death. Using data from EHRs for patient record analysis, the results indicated that Dynamic Logic had a consistent advantage in estimating the probability of death when compared to the standard statistical benchmark method.

 

Dr. Li Zhou developed an NLP system for allergy information, with the goal of improving patients' allergy lists in the EHR by identifying allergy references in free-text notes. The study found that the NLP system successfully identified 96 percent of allergy data in free-text notes.

 

Dr. Li Zhou was also funded to use NLP to improve the quality of medical documents created with speech recognition. While speech recognition is widely used, it has a significant 10 to 23 percent error rate. This research applied NLP and machine-learning methods to detect errors in speech-recognition notes. The study’s findings demonstrated that the use of language models for error detection is a promising tool for improving the accuracy of medical documents, but further work is needed to reduce false positives and the identification of errors in specific words versus sentences.

 

Interoperability

Interoperability is the ability of different information systems, devices, and applications (‘systems’) to access, exchange, integrate, and cooperatively use data in a coordinated manner, within and across organizational, regional, and national boundaries, to provide timely and seamless portability of information and optimize the health of individuals and populations globally. Health data exchange architectures, application interfaces, and standards enable data to be accessed and shared appropriately and securely across the complete spectrum of care, within all applicable settings and with relevant stakeholders, including by the individual . AHRQ funded the following research to develop interoperable systems in support of patient care and population health:

 

Dr. Mollie Rebecca Cummins developed a model process for poison control centers (PCC) to exchange data with EDs by using a standardized data exchange consult note template. In addition, Dr. Cummins and her team developed a user interface called SNOWHITE to populate the template and make it available to the State HIE. This study is a first step toward enabling data exchange between PCCs and EDs, and the resources developed in the study are now available to others as an open-source tool.

Read Dr. Cummins' Impact Story

 

Dr. Joshua Ryan Vest investigated the relationship between the health information exchange (HIE) concepts of “pull” (allowing providers to query community-wide patient records) and “push” (delivering key information to patient records). The research team found that “push” and “pull” HIE are complementary. “Pull” HIE usage by providers was higher for clinical encounters, while “push” usage was higher for imaging information and clinical documents.