Developing and Implementing Asthma-Guidance and Prediction System (a-GPS) for Better Asthma Management

Developing and Implementing Asthma-Guidance and Prediction System (a-GPS) for Better Asthma Management

Enhancing Asthma Care and Outcome Through the Implementation of Asthma-Guidance and Prediction System (a-GPS) on Asthma Management Program: A Randomized Block Design

Asthma is the most common chronic condition in children and one of the five most burdensome diseases in the United States. Despite this, research and care for childhood asthma are limited by inefficient utilization of electronic medical records (EMRs) to facilitate large-scale studies and care. The primary goal of this clinical trial is to implement the asthma-Guidance and Prediction System (a-GPS) on the Asthma Management Program (AMP, a current care coordination program for asthma care of children aged 5-17 years at Mayo Clinic). Primary hypothesis: The implementation of a-GPS in the current care is logistically feasible.

Despite the availability of evidence-based guidelines for asthma management and effective asthma therapies, asthma continues to cause a significant morbidity and burden to our society. Growing deployments of Electronic Health Records (EHRs) systems have established large practice-based longitudinal datasets, which allow for the identification of patient cohorts for epidemiological investigations and population-based management. Natural Language Processing (NLP; automated chart review using computer program) has received great attention and has played a critical role in secondary use of EHRs for clinical care and translational research. For example, we recently developed an NLP algorithm for the Predetermined Asthma Criteria (PAC) that can ascertain asthma status without manual chart review. The primary goals of this proposed clinical trial are 1) to implement the asthma-Guidance and Prediction System (a-GPS) on Asthma Management Program (AMP, a current care coordination program for asthma care of children aged 5-17 years at Mayo Clinic) and 2) assess the impact of a-GPS on the primary and secondary end points for a one-year study period. These goals will be accomplished by conducting a randomized clinical trial with block design for three groups of children as the groups (blocks) of children are significantly heterogeneous in terms of receiving asthma care. The a-GPS program includes 1) natural language processing (NLP) capabilities (i.e., automated EHR review to identify asthma status (yes vs. no) and monitor asthma activity (onset, remission, and relapse) in real time), 2) temporal and geospatial trends analysis of asthma outcome and care, and 3) asthma care optimization through predictive analytics. The primary end points include asthma outcome using quarterly measured age-appropriate asthma control questionnaire (ie, Asthma Control Test (ACT; validated for children aged ≥ 4 years) scores for children ≥ 4 years: a total duration of ACT scores > 19, or Test for Respiratory and Asthma Control in Kids (TRACK; validated for children under 5 years) scores for children <4 years: a total duration of TRACK scores < 80), care quality (timely care in response to asthma-related events), and costs (total costs per member). For those in Block 3, the rate of a physician diagnosis of asthma during the study will be also compared between the intervention and control groups as a measure for quality care.

Decision Support Techniques, Delayed Diagnosis, Natural Language Processing, Geographic Information Systems

Clinicians will be provided a-GPS data on a regular basis for intervention group, but not control group such as their risk factors for asthma, quality of care, and asthma outcomes.

Asthma exacerbation will be defined as the number of emergency department visits/ hospitalization for asthma (asthma symptoms) or unscheduled visit for asthma (asthma symptoms) requiring oral corticosteroid. This outcome will be retrieved from the electronic health record for the subjects.

Clinicians' workload in duration of time to collect and review clinical data from EHRs for making a clinical decision

A total cost of health care (regardless of asthma) per study subject for the 1 year before the study starts and during the study period will be calculated and assessed.

A quarterly asthma control status will be measured by administering Asthma Control Test (ACT) or Test for Respiratory and Asthma Control in Kids (TRACK) by an asthma care coordinator, care team or study coordinator over the phone or online ACT questionnaire with a reminding system. Good asthma control will be defined as >ACT score of 19 for children ≥ 4 years or <TRACK score of 80 for children <4 years. Patient''s asthma control status will be determined as good vs. bad control.

Documented any asthma care either via clinic visit or by asthma care coordinator's contact after asthma-related adverse events (ie, ER/Hospitalization for asthma, or asthma exacerbation requiring oral corticosteroid use) and time gap (in days) will be retrieved and assessed.

Inclusion Criteria for Children in Block 1: - Must be enrolled in AMP at the time of enrollment. Inclusion Criteria for Children in Block 2: Physician diagnosis of persistent asthma by NLP program for the list of physician diagnoses referring to persistent asthma, and/or Persistent asthma equivalent condition by either the Healthcare Effectiveness Data and Information Set (HEDIS); (e.g., ER visit or hospitalization for asthma during the past 12 months) or the National Asthma Education and Prevention Program (NAEPP); (e.g., ≥2 exacerbations requiring oral systemic corticosteroids in the past 6 months for children aged 0-4 years and 12 months for those aged ≥5 years), and/or Physician diagnosis of asthma with controller medication (e.g., inhaled corticosteroid) documented in the past 12 months, but they were not enrolled in AMP at the time of enrollment or during run-in period. Inclusion Criteria for Children in Block 3: - Children must meet the criteria for asthma delineated in Table 1 in protocol for asthma and recurrent asthma-like symptoms, but do not have a documentation of a diagnosis of asthma in medical records aged 0-17 years. Exclusion Criteria (All Blocks): Non-Olmsted County residents Children who are not enrolled in Mayo Clinic downtown pediatric practice No research authorization for using medical records for research Immunosuppressive therapy Conditions making asthma ascertainment difficult for Block 3 (pulmonary function tests that showed forced expiratory volume at one second (FEV1) to be consistently below 50% predicted or diminished diffusion capacity, tracheobronchial foreign body at or about the incidence date of asthma, wheezing occurring only in response to anesthesia or medications, bullous emphysema or pulmonary fibrosis on chest radiograph, homozygous alpha 1-protease inhibitor deficiency (PiZZ) alpha1-antitrypsin, cystic fibrosis, other major chest disease such as severe kyphoscoliosis or bronchiectasis) Children and their caregivers who decline to participate in the study

Seol HY, Shrestha P, Muth JF, Wi CI, Sohn S, Ryu E, Park M, Ihrke K, Moon S, King K, Wheeler P, Borah B, Moriarty J, Rosedahl J, Liu H, McWilliams DB, Juhn YJ. Artificial intelligence-assisted clinical decision support for childhood asthma management: A randomized clinical trial. PLoS One. 2021 Aug 2;16(8):e0255261. doi: 10.1371/journal.pone.0255261. eCollection 2021.