PURPOSE/ OBJECTIVES: To determine predictors of fall events in hospitalized patients with cancer and develop a scoring system to predict fall events. DESIGN: Retrospective medical record review. SETTING: A 1,200-bed tertiary care hospital in northeastern Ohio. SAMPLE: 145 patients with cancer who did not have a fall event were randomly selected from all oncology admissions from February 2006-January 2007 and compared to 143 hospitalized patients with cancer who had a fall event during the same period. METHODS: Multivariable logistic regression models predicting falls were fit. Risk score analysis was completed using bootstrap samples to evaluate discrimination between patients who did or did not fall and agreement between predicted and actual fall status. A nomogram of risk scores was created. MAIN RESEARCH VARIABLES: Fall episodes during hospitalization and patient characteristics that predict falls. FINDINGS: While patients were hospitalized for cancer care, their predictors of a fall episode were low pain level, abnormal gait, cancer type, presence of metastasis, antidepressant and antipsychotic medication use, and blood product use (all p < 0.02); risk model c-statistic was 0.89. CONCLUSIONS: For hospitalized patients with cancer, predictors reflecting greater fall episode risk can be assessed easily by nursing staff and acted on when the risk is sufficiently high. IMPLICATIONS FOR NURSING: Understanding specific risk factors of falls in an adult oncology population may lead to interventions that reduce fall risk.
PURPOSE/ OBJECTIVES: To determine predictors of fall events in hospitalized patients with cancer and develop a scoring system to predict fall events. DESIGN: Retrospective medical record review. SETTING: A 1,200-bed tertiary care hospital in northeastern Ohio. SAMPLE: 145 patients with cancer who did not have a fall event were randomly selected from all oncology admissions from February 2006-January 2007 and compared to 143 hospitalized patients with cancer who had a fall event during the same period. METHODS: Multivariable logistic regression models predicting falls were fit. Risk score analysis was completed using bootstrap samples to evaluate discrimination between patients who did or did not fall and agreement between predicted and actual fall status. A nomogram of risk scores was created. MAIN RESEARCH VARIABLES: Fall episodes during hospitalization and patient characteristics that predict falls. FINDINGS: While patients were hospitalized for cancer care, their predictors of a fall episode were low pain level, abnormal gait, cancer type, presence of metastasis, antidepressant and antipsychotic medication use, and blood product use (all p < 0.02); risk model c-statistic was 0.89. CONCLUSIONS: For hospitalized patients with cancer, predictors reflecting greater fall episode risk can be assessed easily by nursing staff and acted on when the risk is sufficiently high. IMPLICATIONS FOR NURSING: Understanding specific risk factors of falls in an adult oncology population may lead to interventions that reduce fall risk.
Authors: Tanya M Wildes; Priya Dua; Susan A Fowler; J Philip Miller; Christopher R Carpenter; Michael S Avidan; Susan Stark Journal: J Geriatr Oncol Date: 2014-10-30 Impact factor: 3.599
Authors: Schroder Sattar; Shabbir M H Alibhai; Sandra L Spoelstra; Rouhi Fazelzad; Martine T E Puts Journal: Support Care Cancer Date: 2016-07-22 Impact factor: 3.603
Authors: Elizabeth A Fehlberg; Robert J Lucero; Michael T Weaver; Anna M McDaniel; A Michelle Chandler; Phyllis A Richey; Lorraine C Mion; Ronald I Shorr Journal: BMJ Open Date: 2017-08-07 Impact factor: 2.692
Authors: Natalie B V Riblet; Evelyn M Schlosser; Jennifer A Snide; Lara Ronan; Katherine Thorley; Melissa Davis; Jennifer Hong; Linda P Mason; Tobi J Cooney; Lanelle Jalowiec; Nancy L Kennedy; Sabrina Richie; David Nalepinski; Camilo E Fadul Journal: Neurooncol Pract Date: 2015-10-23