Other Titles
Using technologies to influence care
Abstract
Session presented on Saturday, July 25, 2015: Background: The popularity of 'big data' along with an increasing capacity for real-time predictive analytics to augment clinical decision support systems (CDSS) within electronic health records has led to rapid innovations. Hundreds of complex prediction models have been developed for healthcare-focused outcomes over the last few years, and many of these are able to incorporate data from dozens of variables in real-time while providing a probability of a particular event. While these models can be highly accurate, the ability of these systems to influence patient outcomes is relatively unknown. Furthermore, most of the outcomes for which models are developed target the workflow of physicians and other advanced practice providers even though nurses are the largest profession within the healthcare workforce. Even in the broader realm of CDSS, few studies have examined the impact of CDSS on nurses' decisions and the patient outcomes associated with them. Objective: A literature review was performed to summarize the state of the science for the impact of predictive analytic-enhanced CDSS on nursing-sensitive patient outcomes. Method: A scoping literature review explored the impact predictive analytic-enhanced CDSS have on 4 nursing-sensitive patient outcomes (pressure ulcers, failure to rescue [including sepsis and cardiopulmonary arrests of all causes], falls, and infections). These outcomes were chosen due to their high incidence and cost along with their ability to be predicted in real-time with current technology and modeling strategies. Reviews and primary research studies were sought in MEDLINE and the Cumulative Index to Nursing and Allied Health Literature (CINAHL) by including concepts and phrases surrounding CDSS, predictive analytics, nursing, and each outcome. Topical and keyword searches were performed in the Science Citation Index and the Social Sciences Citation Index as well as the Virginia Henderson Global Nursing e-Repository. Studies were included in the critique if they measured the impact of predictive analytics on patient outcomes. Due to the expected paucity of literature, no additional a priori exclusion criteria were defined. One additional study was published during the review process and is also included in the critique. Results: A total of 306 studies were reviewed following removal of duplicates, and only 4 studies met criteria for inclusion in the critique. The oldest article was published in 2011, highlighting the relatively novel nature of this technology. None of the studies explored falls or nosocomial infections; only one study explored pressure ulcers. Studies exploring failure to rescue used a randomized control trial design at either the individual or unit/ward level while the study exploring pressure ulcers used a pre-/post-intervention design. Although statistically significant results were reported for at least one outcome in 3 of the 4 studies, several methodological limitations were present. Discussion: While many of the articles retrieved during the search discussed variable selection and predictive model development/validation, only 4 articles examined the impact on patient outcomes. The novelty of predictive analytics and the inherent methodological challenges in studying CDSS impact are likely responsible for this paucity of literature. These challenges included: (a) multilevel nature of the intervention [i.e., determining whether the patient or the nurse is the targeted level of treatment and analysis], (b) treatment fidelity [i.e., assessing whether or not nurses changed their behaviors following new information from the CDSS], and (c) adequacy of clinicians' subsequent behavior [i.e., uncertainty in the sufficiency of biomedical evidence to recommend a particular intervention for the patient outcome]. Conclusions: Insufficient evidence currently exists to demonstrate efficacy of predictive analytic-enhanced CDSS on nursing-sensitive patient outcomes. In addition to the need for innovative research methods to study this phenomenon, a greater emphasis on examining its potential within nursing is recommended before practice can be influenced.
Sigma Membership
Unknown
Lead Author Affiliation
Vanderbilt University, Nashville, Tennessee, USA
Type
Presentation
Format Type
Text-based Document
Study Design/Type
N/A
Research Approach
N/A
Keywords:
Clinical Decision Support Systems, Big Data, Predictive Analytics
Recommended Citation
Jeffery, Alvin Dean, "Impact of real-time prediction model-enhanced clinical decision support systems on nursing sensitive patient outcomes: A review of the literature" (2016). INRC (Congress). 280.
https://www.sigmarepository.org/inrc/2015/presentations_2015/280
Conference Name
26th International Nursing Research Congress
Conference Host
Sigma Theta Tau International
Conference Location
San Juan, Puerto Rico
Conference Year
2015
Rights Holder
All rights reserved by the author(s) and/or publisher(s) listed in this item record unless relinquished in whole or part by a rights notation or a Creative Commons License present in this item record.
All permission requests should be directed accordingly and not to the Sigma Repository.
All submitting authors or publishers have affirmed that when using material in their work where they do not own copyright, they have obtained permission of the copyright holder prior to submission and the rights holder has been acknowledged as necessary.
Acquisition
Proxy-submission
Impact of real-time prediction model-enhanced clinical decision support systems on nursing sensitive patient outcomes: A review of the literature
San Juan, Puerto Rico
Session presented on Saturday, July 25, 2015: Background: The popularity of 'big data' along with an increasing capacity for real-time predictive analytics to augment clinical decision support systems (CDSS) within electronic health records has led to rapid innovations. Hundreds of complex prediction models have been developed for healthcare-focused outcomes over the last few years, and many of these are able to incorporate data from dozens of variables in real-time while providing a probability of a particular event. While these models can be highly accurate, the ability of these systems to influence patient outcomes is relatively unknown. Furthermore, most of the outcomes for which models are developed target the workflow of physicians and other advanced practice providers even though nurses are the largest profession within the healthcare workforce. Even in the broader realm of CDSS, few studies have examined the impact of CDSS on nurses' decisions and the patient outcomes associated with them. Objective: A literature review was performed to summarize the state of the science for the impact of predictive analytic-enhanced CDSS on nursing-sensitive patient outcomes. Method: A scoping literature review explored the impact predictive analytic-enhanced CDSS have on 4 nursing-sensitive patient outcomes (pressure ulcers, failure to rescue [including sepsis and cardiopulmonary arrests of all causes], falls, and infections). These outcomes were chosen due to their high incidence and cost along with their ability to be predicted in real-time with current technology and modeling strategies. Reviews and primary research studies were sought in MEDLINE and the Cumulative Index to Nursing and Allied Health Literature (CINAHL) by including concepts and phrases surrounding CDSS, predictive analytics, nursing, and each outcome. Topical and keyword searches were performed in the Science Citation Index and the Social Sciences Citation Index as well as the Virginia Henderson Global Nursing e-Repository. Studies were included in the critique if they measured the impact of predictive analytics on patient outcomes. Due to the expected paucity of literature, no additional a priori exclusion criteria were defined. One additional study was published during the review process and is also included in the critique. Results: A total of 306 studies were reviewed following removal of duplicates, and only 4 studies met criteria for inclusion in the critique. The oldest article was published in 2011, highlighting the relatively novel nature of this technology. None of the studies explored falls or nosocomial infections; only one study explored pressure ulcers. Studies exploring failure to rescue used a randomized control trial design at either the individual or unit/ward level while the study exploring pressure ulcers used a pre-/post-intervention design. Although statistically significant results were reported for at least one outcome in 3 of the 4 studies, several methodological limitations were present. Discussion: While many of the articles retrieved during the search discussed variable selection and predictive model development/validation, only 4 articles examined the impact on patient outcomes. The novelty of predictive analytics and the inherent methodological challenges in studying CDSS impact are likely responsible for this paucity of literature. These challenges included: (a) multilevel nature of the intervention [i.e., determining whether the patient or the nurse is the targeted level of treatment and analysis], (b) treatment fidelity [i.e., assessing whether or not nurses changed their behaviors following new information from the CDSS], and (c) adequacy of clinicians' subsequent behavior [i.e., uncertainty in the sufficiency of biomedical evidence to recommend a particular intervention for the patient outcome]. Conclusions: Insufficient evidence currently exists to demonstrate efficacy of predictive analytic-enhanced CDSS on nursing-sensitive patient outcomes. In addition to the need for innovative research methods to study this phenomenon, a greater emphasis on examining its potential within nursing is recommended before practice can be influenced.