ABSTRACT
OBJECTIVES: To determine whether a multicomponent intervention improves care in hospitalized older adults with cognitive impairment. DESIGN: One-year retrospective chart review with propensity score matching on critical demographic and clinical variables was used to compare individauls with cognitive impairmenet on intervention and nonintervention units. SETTING: Large tertiary medical center. PARTICIPANTS: All hospitalized individuals age 65 and older with cognitive impairment admitted to medicine who required constant or enhanced observation for behavioral and psychological symptoms. INTERVENTION: Multicomponent intervention (geographic unit cohorting, multidisciplinary approach, patient engagement specialists (PES), staff education) or usual care. MEASUREMENTS: In-hospital mortality, length of stay, readmission, management of behavioral disturbances. RESULTS: After propensity score matching, 476 of the 712 intervention visits were pair-matched with 476 of the 558 usual care visits. Matching was successful in balancing baseline covariates between intervention and usual care units. Individuals admitted to the intervention unit had lower in-hospital mortality (1.1% vs 2.9%, p=0.05) and shorter stays (5.0 vs 5.8 days, p=0.04). There was no difference in discharge home (p=0.90) or 30-day readmission rates (p=0.44). Individuals on the intervention unit were less likely than those receivng usual care to have an order for constant (12.0% vs 45.8%, p<0.01) or enhanced (22.1% vs 79.6%, p<0.01) observation, to be taking benzodiazepines (26.3% vs 38.0%, p<0.01), to be taking nothing by mouth (29.6% vs 40.8%, p=0.01), to be on bedrest (17.0% vs 25.8%, p=0.01), to be taking antipsychotics (41.2% vs 54.0%, p<0.01), or to have restraints (3.2% vs 6.9%, p=.01). CONCLUSION: A multicomponent intervention of geographic cohorting, multidisciplinary approach, PES, and staff education may offer a new paradigm in the management of hospitalized older adults with cognitive impairment.
Subject(s)
Cognitive Dysfunction/therapy , Delivery of Health Care/methods , Patient Care Team , Aged , Aged, 80 and over , Cognitive Dysfunction/mortality , Female , Hospital Mortality , Humans , Male , Outcome Assessment, Health Care , Patient Discharge/statistics & numerical data , Patient Participation , Patient Readmission/statistics & numerical data , Propensity Score , Retrospective StudiesABSTRACT
Clinical prediction rules can help clinicians make personalized and complex decisions at the point of care. They have the potential to improve patient care outcomes and reduce healthcare costs, but remain underutilized. There are a number of well-derived and validated clinical prediction rules. Few, however, have been studied by means of an impact analysis or successfully integrated into provider workflow. A heavily identified area of opportunity for integration is the electronic health record. There are, however, a number of barriers to adoption at both the infrastructure and organizational levels. Research efforts should focus on impact analysis and how to successfully implement existing, well-validated clinical prediction rules into daily practice. Recommendations include emphasis on a collaborative framework, using existing technologies, and utilization of usability and workflow integration methodology.
