Application of interRAI Assessments in Disaster Management: Identifying Vulnerable Persons in the Community
MetadataShow full item record
Background: Several studies have shown the increased vulnerability and disproportionate mortality rate among frail community dwelling older adults as a result of disasters. Parallel to an escalating number of disasters, Canada is faced with an aging demographic and a policy shift emphasizing aging at home. This results in a greater vulnerability of this group of high needs community dwelling individuals to the effects of events that lead to interruption of home health care services and/or displacement. Despite the growing vulnerability it has proven to be difficult to identify those most vulnerable older adults and their characteristics. This makes it challenging for emergency managers, first responders and health care providers to develop targeted preparedness, response and recovery strategies aimed at the most vulnerable older adults living at home. Relatively recent developments in electronic health records provide an unprecedented opportunity to use comprehensive assessment information collected as part of routine clinical practice in the home care sector to identify vulnerable community dwelling older adults. In Ontario, the Resident Assessment Instrument for Home Care (RAI-HC) is the mandated primary assessment tool for long-stay home care clients. Objective: The three specific objectives of this dissertation are to examine: 1. The application of the New Zealand Priority Algorithm used during the Christchurch earthquake to the Ontario Home Care Client database. 2. Determinants of Emergency Response Level (ERL) designation within CCACs. 3. The person-level factors that contribute to increased vulnerability of home care clients to power interruptions through examining the health effects of the power outage that occurred as a result of the December 2013 Ice Storm including emergency department (ED) visits, hospitalization and service utilization. Conceptual Framework: The person-environment fit model is used as the conceptual framework for this dissertation. This model views individual vulnerability as a product of the interaction between individual competence, adaptive behavior and the strength of the environmental stress (the emergency or disaster). Where the demands of an emergency or disaster exceed the ability of the older adult to cope, a person- environment misfit may lead to negative health outcomes. Methodology: All research questions were addressed using RAI-HC datasets in combination with other datasets. Chapter three used the RAI-HC database by selecting unique home care clients with assessments closest to December 31st 2014 (N=275,797). For chapter four Emergency Response Level (ERL) codes were provided by the Hamilton Niagara Haldimand Brant (HNHB) and Toronto Central (TC) Community Care Access Centre (CCAC) and matched with a RAI-HC assessment in both CCACs (N=70,292 and N=8,996 respectively). In addition, linkages were made with data regarding death, hospitalization and long term care (LTC) admission. Lastly, chapter five uses information on Toronto Hydro power outages and an estimation of outage areas based on outage mapping in addition to the HC database. The exposure group (N=10,748) was compared to two comparison groups. Group one included clients with HC assessments in the same period and receiving services during the same week but were unaffected by the hydro outage (N=12,072). The second comparison group was comprised of clients residing in the same area as the hydro outage one year prior to the storm (N=10,886). Service utilization was collected from the Client Health Related Information System (CHRIS). Statistical analyses were done using SAS version 9.4 and methods used include frequency tabulation, bivariate logistic regression, multivariate logistic regression as well as Kaplan-Meier survival plotting and Cox proportional hazards ratios calculations. Results: When comparing four decision support algorithms (University of Waterloo, Canterbury, Vulnerable Persons at Risk (VPR) and VPR Plus) to identify high priority clients, the VPR and VPR Plus were most predictive of mortality, LTC admission and hospitalization. The high priority groups were significantly more impaired than lower priority clients with both the VPR and VPR Plus. They had higher levels of health instability, experienced more falls, required more assistance with Activities of Daily Living (ADL), were more cognitively impaired and had higher levels of depression ratings. When comparing the chosen algorithms, the VPR and VPR Plus, with ERL levels assigned by care coordinators the analysis showed considerable overlap in predictive variables. The ERL was highly predictive of mortality and LTC admission, but less predictive of hospitalization. C-stats of logistic regression modeling with ERL and VPR/VPR Plus in predicting mortality showed that the VPR and VPR Plus models were a better or equal fit as models with the ERL. Finally, when examining the characteristics of clients that were affected by the 2013 power outage with the two comparison groups, a significant difference was found for the non-exposed group in the year of the outage in relation to numbers of nursing and personal support worker (PSW) visits, hospital admission and emergency department (ED) visits as well as mortality, LTC admission and hospitalization rates. The analysis showed that clients in the non-affected areas in the year of the outage were more likely to decline in Depression Rating Scale (DRS), Changes in Health, End-Stage Disease, Signs and Symptoms Scale (CHESS) and Instrumental Activities of Daily Living (IADL). This is consistent with the higher rates of LTC admission and hospitalization within six months after the outage for non-exposed clients as well as higher frequency of nursing and PSW visits during and 30 days after the outage. In contrast to the expectation that exposed clients would do worse during and after the outage, the analysis showed that exposed clients showed in fact less health decline than non-exposed clients. However, when looking at those clients that would have been considered high and medium risk clients based on the VPR and VPR Plus, the analysis showed that those clients in areas with hydro outages were more likely to die and to be admitted to long term care (LTC) than the high and medium risk clients living in unaffected areas. Conclusions: The analyses in this dissertation have shown the usefulness of information collected as routine clinical practice using interRAI assessment tools. The current system of designating Emergency Response Levels (ERL) by care coordinators is highly dependent on consistent updating of the ERLs in the system whenever a new home care assessment is completed. The analyses showed that this is not consistently done, and may render the ERL code obsolete overtime. The VPR and VPR Plus have been shown to be valid and reliable alternatives to ERL codes and they are kept up to date as new assessments are completed on home care clients. Incorporating these decision support algorithms into the RAI-HC assessment system software enables an automatic and up to date vulnerability assessment of clients. This can make it possible for emergency managers, first responders and health care providers to use a comprehensive priority system before, during and after emergency, ultimately preventing unnecessary death or health deterioration.
Cite this version of the work
Alexandra Van Solm (2016). Application of interRAI Assessments in Disaster Management: Identifying Vulnerable Persons in the Community. UWSpace. http://hdl.handle.net/10012/10795