{
  "abstract": "Background Predictive models in healthcare use historical data to forecast future trends and events for patients and hospitals. Integrating predictive models with healthcare improvement methodology has great potential to focus change concepts on patients most at risk of a particular outcome. This targeted approach can lead to improvement in outcomes while simultaneously efficiently utilizing resources, as interventions are tested on and applied only to the patients most likely to benefit from them.Objectives The primary objective was to detail how two quality improvement (QI) projects leveraged predictive models to efficiently drive outcomes using the IHI Model for Improvement. Secondary objectives included describing how predictive modeling identifies high-risk patients and informs targeted interventions and demonstrating examples of both clinical and operational improvements achieved.Methods The two detailed quality improvement projects utilized the Institute for Healthcare Improvement (IHI) Model for Improvement. In both cases, predictive models were applied to identify and target high-risk patient populations.· Clinical Improvement: One project targeted orthopedic trauma patients. Interventions included evidence based chemoprophylaxis (Lovenox) and early continuous daily physical and occupational therapy (PT/OT) (figure 1). Plan-Do-Study-Act (PDSA) cycles were used, specifically testing early PT/OT in all patients versus testing it only in high-risk patients (figure 2). The outcome measured was the rate of perioperative pulmonary embolism (PE) and deep venous thrombosis (DVT), also known as Patient Safety Indicator 12 (PSI 12).· Operational Improvement: The second project focused on reducing same-day case cancellations for patients having elective ear, nose, throat (ENT) surgery. This improvement utilized a Preoperative Evaluation Triage System. This system incorporated a Machine-Learning American Society of Anesthesiologist (ASA) Score and Patient Questionnaire to determine if an Anesthesia televisit was indicated (figure 4).Results The project targeting orthopedic trauma patients resulted in a 50% reduction in perioperative pulmonary embolism and deep venous thrombosis (PE/DVT). The rate, measured by PSI 12, was reduced from a baseline of 5.2 to 2.6 ( figure 3). For the project focused on elective ENT surgery achieved a 13% reduction in same-day case cancellations. Furthermore, the preoperative evaluation triage system reduced unnecessary primary care physician (PCP) Preoperative Visits from 12% to 4% (figure 5), thereby increasing access for other patients.Conclusions Integrating predictive models with the IHI Model for Improvement is a successful strategy to efficiently drive outcomes. This approach demonstrates effectiveness in achieving both clinical improvements (PE/DVT reduction) and operational efficiencies (cancellation reduction). Crucially, by applying interventions only to those patients most likely to benefit, the integration of predictive modeling facilitates the conservation of resources.Abstract 43 Figure 1Intervention for high-risk patientsAbstract 43 Figure 2Physical and occupational therapy run chartAbstract 43 Figure 3Outcome measure-PE/DVT rateAbstract 43 Figure 4Preoperative evaluation triage systemAbstract 43 Figure 5Otorhinolaryngology (ORL) cancellation rate",
  "authors": [
    {
      "affiliations": [
        "Montefiore Medical Center"
      ],
      "name": "Mark Wnorowski"
    },
    {
      "affiliations": [
        "Montefiore Medical Center"
      ],
      "name": "Christopher Tam"
    },
    {
      "affiliations": [
        "Montefiore Medical Center"
      ],
      "name": "Erin Andrews"
    },
    {
      "affiliations": [
        "Montefiore Medical Center"
      ],
      "name": "Karuna Wongtangman"
    },
    {
      "affiliations": [
        "Montefiore Medical Center"
      ],
      "name": "Fran Ganz-Lord"
    }
  ],
  "title": "43 Maximizing resources: predictive modeling meets the model for improvement",
  "uid": "04805b9b-7851-5d12-afee-c9252da8f928"
}