George Kephart
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
George Kephart has a total of 189 co-authors. Here are the top co-authors by number of shared publications:
- Ingrid Sketris
College of Pharmacy Dalhousie University Halifax Nova Scotia Canada
15 shared publications - Tanya Leigh Packer
Dalhousie University
14 shared publications - Åsa Audulv
Department of Nursing Umeå University Umeå Sweden
10 shared publications - Grace Warner
Faculty of Health, Dalhousie University, Halifax, NS B3H 4R2, Canada
9 shared publications - Yukiko Asada
Department of Bioethics, Clinical Center, National Institutes of Health, Bethesda, USA
6 shared publications - Ingrid Olsson
Department of Nursing, Umea University, Umea, Sweden
4 shared publications
Latest Publications
- Open Access
- 6 January 2026
- Published by Springer Nature in Quality of Life Research
- Vol. 35 (1) , 24
The Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC): testing for bias across patient characteristics and languages
Purpose: Developed simultaneously in English and Swedish, the Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) assesses patients’ perceived difficulty managing life with long-term health conditions. This study assessed the comparability of the PRISM-CC across sociodemographic groups, number of health conditions and language (English and Swedish). Methods: Differential item functioning (DIF) and differential test functioning (DTF) were analysed by age, gender, education level, and number of conditions using independent English and Swedish datasets. Language-based DIF and DTF were examined using pooled data. An iterative hybrid ordinal logistic regression approach was applied to identify potential DIF across the PRISM-CC’s seven domains. The impact of flagged items on total scores (DTF) was evaluated by comparing test characteristic curves. Results: Few items were flagged for potential DIF in the English, Swedish or pooled data, and only at low cutoff values. The impact of items with potential DIF on DTF was negligible. Conclusion: The absence of meaningful DIF and DTF in either the English or Swedish version of the PRISM-CC or between English and Swedish versions provides further support for the PRISM-CC as a valuable tool for assessing self-management ease and difficulty. These results also demonstrate the value of simultaneous development of instruments in two languages. Further evaluation of DIF is necessary in populations with greater self-management challenges, such as among people with severe disease burden.
- 6 January 2026
- Published by Springer Nature in Quality of Life Research
- Vol. 35 (1) , 24
Open AccessThe Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC): testing for bias across patient characteristics and languages
Purpose: Developed simultaneously in English and Swedish, the Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) assesses patients’ perceived difficulty managing life with long-term health conditions. This study assessed the comparability of the PRISM-CC across sociodemographic groups, number of health conditions and language (English and Swedish). Methods: Differential item functioning (DIF) and differential test functioning (DTF) were analysed by age, gender, education level, and number of conditions using independent English and Swedish datasets. Language-based DIF and DTF were examined using pooled data. An iterative hybrid ordinal logistic regression approach was applied to identify potential DIF across the PRISM-CC’s seven domains. The impact of flagged items on total scores (DTF) was evaluated by comparing test characteristic curves. Results: Few items were flagged for potential DIF in the English, Swedish or pooled data, and only at low cutoff values. The impact of items with potential DIF on DTF was negligible. Conclusion: The absence of meaningful DIF and DTF in either the English or Swedish version of the PRISM-CC or between English and Swedish versions provides further support for the PRISM-CC as a valuable tool for assessing self-management ease and difficulty. These results also demonstrate the value of simultaneous development of instruments in two languages. Further evaluation of DIF is necessary in populations with greater self-management challenges, such as among people with severe disease burden.
- Open Access
- 1 June 2025
- Published by Taylor & Francis in Scandinavian Journal of Primary Health Care
- Vol. 44 (1) , 1-13
Self-management difficulties in Swedish older adults and associations with sociodemographic factors, number of conditions, depression and health status
This study describes patterns of self-management ease and difficulty among older adults with long-term health conditions and the associations with gender, level of education, number of conditions, depression and/or health status. Cross-sectional data were collected between 2021-2022 in a municipality in northern Sweden. The survey included demographic and health-related questions. To assess self-management ease or difficulty and symptoms of depression, the Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) and the Geriatric Depression Scale were used. 516 older adults between 72-73 years of age with long-term health conditions were included. Descriptive statistics and logistic regression were used to describe patterns of self-management ease and difficulty and to examine which factors were associated with self-management difficulty. Most older adults did not experience self-management difficulty. There were, however, differences between the seven PRISM-CC domains. The Internal domain (managing negative emotions and stress) had the highest percentage (25.39%) of older adults with self-management difficulty. In all domains, there was also a subgroup of individuals (n = 26) that had noticeably lower PRISM-CC scores (more difficulty). A strong association between having depressive symptoms or having poor health status and self-management difficulty was found. This study highlights the need for regular mental health screenings and individualized self-management support for older adults. Future research should explore intervention strategies that integrate mental health support into self-management programs for individuals with long-term health conditions.
- 1 June 2025
- Published by Taylor & Francis in Scandinavian Journal of Primary Health Care
- Vol. 44 (1) , 1-13
Open AccessSelf-management difficulties in Swedish older adults and associations with sociodemographic factors, number of conditions, depression and health status
This study describes patterns of self-management ease and difficulty among older adults with long-term health conditions and the associations with gender, level of education, number of conditions, depression and/or health status. Cross-sectional data were collected between 2021-2022 in a municipality in northern Sweden. The survey included demographic and health-related questions. To assess self-management ease or difficulty and symptoms of depression, the Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) and the Geriatric Depression Scale were used. 516 older adults between 72-73 years of age with long-term health conditions were included. Descriptive statistics and logistic regression were used to describe patterns of self-management ease and difficulty and to examine which factors were associated with self-management difficulty. Most older adults did not experience self-management difficulty. There were, however, differences between the seven PRISM-CC domains. The Internal domain (managing negative emotions and stress) had the highest percentage (25.39%) of older adults with self-management difficulty. In all domains, there was also a subgroup of individuals (n = 26) that had noticeably lower PRISM-CC scores (more difficulty). A strong association between having depressive symptoms or having poor health status and self-management difficulty was found. This study highlights the need for regular mental health screenings and individualized self-management support for older adults. Future research should explore intervention strategies that integrate mental health support into self-management programs for individuals with long-term health conditions.
- Open Access
- 28 May 2025
- Published by Springer Nature in Journal of Patient-Reported Outcomes
- Vol. 9 (1) , 1-16
Structural validity and test-retest reliability of the Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) in a Swedish population of seventy-year-olds with long-term health conditions
Background: Self-management is internationally recognized as important to maintain independence, quality of life and to minimize the risk of poor health outcomes, especially among persons with multi-morbidity. Self-management can be especially challenging for older adults, who have higher rates of multi-morbidity and experience diverse impacts of long-term health conditions on everyday life. Good measures of self-management are currently lacking. The Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) is a new, generic, multidimensional measure of self-perceived ease or difficulty with self-management, that overcomes many of the limitations of existing measures. Objectives: To test the structural validity and test-retest reliability of the Swedish version of the PRISM-CC among seventy-year-olds with long-term health conditions. Methods: Translation of PRISM-CC items into Swedish followed the Patient-Reported Outcome (PRO) Consortium process. Survey data (n = 516 Swedish seventy-year-olds with ≥1 long-term health condition) was used to assess structural validity of the 36-item PRISM-CC using multidimensional item response theory (IRT) models. Test-retest reliability was assessed on a subsample of 58 individuals using intra-class correlation coefficient (ICC) and Bland-Altman Plots. Results: The Swedish PRISM-CC demonstrated good internal consistency with Cronbach’s alpha >0.8 for all domains, and good fit to a graded response IRT model (RMSEA 0.034, SRMSR 0.050, CFI 0.952 and TLI 0.945). All 36 items had standardized loadings >0.7. ICC showed moderate to good test-retest reliability for all seven domains. The Bland-Altman plots showed minimal bias and good test-retest agreement for all domains. Conclusion: The Swedish PRISM-CC showed good structural validity and test-retest reliability in this sample of relatively healthy seventy-year-olds with long-term health condition(s). Further validation in a population with more severe health issues is needed.
- 28 May 2025
- Published by Springer Nature in Journal of Patient-Reported Outcomes
- Vol. 9 (1) , 1-16
Open AccessStructural validity and test-retest reliability of the Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) in a Swedish population of seventy-year-olds with long-term health conditions
Background: Self-management is internationally recognized as important to maintain independence, quality of life and to minimize the risk of poor health outcomes, especially among persons with multi-morbidity. Self-management can be especially challenging for older adults, who have higher rates of multi-morbidity and experience diverse impacts of long-term health conditions on everyday life. Good measures of self-management are currently lacking. The Patient Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) is a new, generic, multidimensional measure of self-perceived ease or difficulty with self-management, that overcomes many of the limitations of existing measures. Objectives: To test the structural validity and test-retest reliability of the Swedish version of the PRISM-CC among seventy-year-olds with long-term health conditions. Methods: Translation of PRISM-CC items into Swedish followed the Patient-Reported Outcome (PRO) Consortium process. Survey data (n = 516 Swedish seventy-year-olds with ≥1 long-term health condition) was used to assess structural validity of the 36-item PRISM-CC using multidimensional item response theory (IRT) models. Test-retest reliability was assessed on a subsample of 58 individuals using intra-class correlation coefficient (ICC) and Bland-Altman Plots. Results: The Swedish PRISM-CC demonstrated good internal consistency with Cronbach’s alpha >0.8 for all domains, and good fit to a graded response IRT model (RMSEA 0.034, SRMSR 0.050, CFI 0.952 and TLI 0.945). All 36 items had standardized loadings >0.7. ICC showed moderate to good test-retest reliability for all seven domains. The Bland-Altman plots showed minimal bias and good test-retest agreement for all domains. Conclusion: The Swedish PRISM-CC showed good structural validity and test-retest reliability in this sample of relatively healthy seventy-year-olds with long-term health condition(s). Further validation in a population with more severe health issues is needed.
- Open Access
- 6 March 2025
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 22 (3) , 381
Creating an Inclusive Definition for High Users of Inpatient Hospital Systems That Considers Different Levels of Rurality
Multiple definitions have been used to identify individuals who are high system users (HSUs), through economic costs, frequency of use, or length of stay for inpatient care users. However, no definition has been validated to be representative of those residing in rural communities, who face unique service accessibility. This paper identifies an HSU definition for rural Canada that is inclusive of various levels of rurality, longitudinal patient experiences, and types of hospitalizations experienced. This study utilized the 2011 Canadian Census Health and Environment Cohort (CanCHEC) linkage profile to assess hospitalization experiences between 1 January 2009 and 31 December 2013. A range of common HSU indicators were compared using Cox proportional hazards modelling for multiple periods of assessment and types of admissions. The preferred definition for rural HSUs was individuals who are in the 90th percentile of unplanned hospitalization episodes for 2 of 3 consecutive years. This approach is innovative in that it includes longitudinal hospital experiences and multiple types of hospitalizations and assesses an individual’s rurality as a point of context for analysis, rather than a characteristic. These differences provide an opportunity for community characteristic needs assessment and subsequent adjustments to policy development and resource allocation to meet each rural community’s specific needs.
- 6 March 2025
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 22 (3) , 381
Open AccessCreating an Inclusive Definition for High Users of Inpatient Hospital Systems That Considers Different Levels of Rurality
Multiple definitions have been used to identify individuals who are high system users (HSUs), through economic costs, frequency of use, or length of stay for inpatient care users. However, no definition has been validated to be representative of those residing in rural communities, who face unique service accessibility. This paper identifies an HSU definition for rural Canada that is inclusive of various levels of rurality, longitudinal patient experiences, and types of hospitalizations experienced. This study utilized the 2011 Canadian Census Health and Environment Cohort (CanCHEC) linkage profile to assess hospitalization experiences between 1 January 2009 and 31 December 2013. A range of common HSU indicators were compared using Cox proportional hazards modelling for multiple periods of assessment and types of admissions. The preferred definition for rural HSUs was individuals who are in the 90th percentile of unplanned hospitalization episodes for 2 of 3 consecutive years. This approach is innovative in that it includes longitudinal hospital experiences and multiple types of hospitalizations and assesses an individual’s rurality as a point of context for analysis, rather than a characteristic. These differences provide an opportunity for community characteristic needs assessment and subsequent adjustments to policy development and resource allocation to meet each rural community’s specific needs.
- Open Access
- 19 February 2025
- Published by JMIR Publications Inc. in JMIR Public Health and Surveillance
- Vol. 11 (1) , e66056
Development and Validation of Prediction Models for Perceived and Unmet Mental Health Needs in the Canadian General Population: Model-Based Synthetic Estimation Study
Background: Research has shown that perceptions of a mental health need are closely associated with service demands and are an important dimension in needs assessment. Perceived and unmet mental health needs are important factors in the decision-making process regarding mental health services planning and resources allocation. However, few prediction tools are available to be used by policy and decision makers to forecast perceived and unmet mental health needs at the population level. Objective: We aim to develop prediction models to forecast perceived and unmet mental health needs at the provincial and health regional levels in Canada. Methods: Data from 2018, 2019, and 2020 Canadian Community Health Survey and Canadian Urban Environment were used (n=65,000 each year). Perceived and unmet mental health needs were measured by the Perceived Needs for Care Questionnaire. Using the 2018 dataset, we developed the prediction models through the application of regression synthetic estimation for the Atlantic, Central, and Western regions. The models were validated in the 2019 and 2020 datasets at the provincial level and in 10 randomly selected health regions by comparing the observed and predicted proportions of the outcomes. Results: In 2018, a total of 17.82% of the participants reported perceived mental health need and 3.81% reported unmet mental health need. The proportions were similar in 2019 (18.04% and 3.91%) and in 2020 (18.1% and 3.92%). Sex, age, self-reported mental health, physician diagnosed mood and anxiety disorders, self-reported life stress and life satisfaction were the predictors in the 3 regional models. The individual based models had good discriminative power with C statistics over 0.83 and good calibration. Applying the synthetic models in 2019 and 2020 data, the models had the best performance in Ontario, Quebec, and British Columbia; the absolute differences between observed and predicted proportions were less than 1%. The absolute differences between the predicted and observed proportion of perceived mental health needs in Newfoundland and Labrador (−4.16% in 2020) and Prince Edward Island (4.58% in 2019) were larger than those in other provinces. When applying the models in the 10 selected health regions, the models calibrated well in the health regions in Ontario and in Quebec; the absolute differences in perceived mental health needs ranged from 0.23% to 2.34%. Conclusions: Predicting perceived and unmet mental health at the population level is feasible. There are common factors that contribute to perceived and unmet mental health needs across regions, at different magnitudes, due to different population characteristics. Therefore, predicting perceived and unmet mental health needs should be region specific. The performance of the models at the provincial and health regional levels may be affected by population size.
- 19 February 2025
- Published by JMIR Publications Inc. in JMIR Public Health and Surveillance
- Vol. 11 (1) , e66056
Open AccessDevelopment and Validation of Prediction Models for Perceived and Unmet Mental Health Needs in the Canadian General Population: Model-Based Synthetic Estimation Study
Background: Research has shown that perceptions of a mental health need are closely associated with service demands and are an important dimension in needs assessment. Perceived and unmet mental health needs are important factors in the decision-making process regarding mental health services planning and resources allocation. However, few prediction tools are available to be used by policy and decision makers to forecast perceived and unmet mental health needs at the population level. Objective: We aim to develop prediction models to forecast perceived and unmet mental health needs at the provincial and health regional levels in Canada. Methods: Data from 2018, 2019, and 2020 Canadian Community Health Survey and Canadian Urban Environment were used (n=65,000 each year). Perceived and unmet mental health needs were measured by the Perceived Needs for Care Questionnaire. Using the 2018 dataset, we developed the prediction models through the application of regression synthetic estimation for the Atlantic, Central, and Western regions. The models were validated in the 2019 and 2020 datasets at the provincial level and in 10 randomly selected health regions by comparing the observed and predicted proportions of the outcomes. Results: In 2018, a total of 17.82% of the participants reported perceived mental health need and 3.81% reported unmet mental health need. The proportions were similar in 2019 (18.04% and 3.91%) and in 2020 (18.1% and 3.92%). Sex, age, self-reported mental health, physician diagnosed mood and anxiety disorders, self-reported life stress and life satisfaction were the predictors in the 3 regional models. The individual based models had good discriminative power with C statistics over 0.83 and good calibration. Applying the synthetic models in 2019 and 2020 data, the models had the best performance in Ontario, Quebec, and British Columbia; the absolute differences between observed and predicted proportions were less than 1%. The absolute differences between the predicted and observed proportion of perceived mental health needs in Newfoundland and Labrador (−4.16% in 2020) and Prince Edward Island (4.58% in 2019) were larger than those in other provinces. When applying the models in the 10 selected health regions, the models calibrated well in the health regions in Ontario and in Quebec; the absolute differences in perceived mental health needs ranged from 0.23% to 2.34%. Conclusions: Predicting perceived and unmet mental health at the population level is feasible. There are common factors that contribute to perceived and unmet mental health needs across regions, at different magnitudes, due to different population characteristics. Therefore, predicting perceived and unmet mental health needs should be region specific. The performance of the models at the provincial and health regional levels may be affected by population size.
- Open Access
- 27 August 2024
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 21 (9) , 1128
Time-Use Sequences: A Mixed-Methods Study Exploring How, When, and Where Spatiotemporal Patterns of Everyday Routines Can Strengthen Public Health Interventions
Background: Behavior change interventions are critical for the secondary prevention of cardiovascular disease and for reducing the risk of a repeat event or mortality. However, the effectiveness of behavior change interventions is challenged by a lack of spatiotemporal contexts, limiting our understanding of factors that influence the timing and location in which day-to-day activities occur and the maintenance of behavior change. This study explored how behavior change interventions could incorporate spatiotemporal contexts of patient activities for modifying behaviors. Methods: A mixed-methods approach with adapted geo-ethnography techniques was used to solicit detailed descriptions of patients’ day-to-day routines, including where, when, and how patients spend time. Data were gathered from patients in one cardiac intervention program in Nova Scotia, Canada, from June to September 2021. Results: A total of 29 individuals (19 men and 10 women) between the ages of 45 and 81 and referred to the program after a cardiac event participated. The results show three key findings: (1) most patients exceeded the minimum guidelines of 30 min of daily physical activity but were sedentary for long periods of time, (2) patient time-use patterns are heterogenous and unique to contexts of individual space-time activity paths, and (3) time-use patterns reveal when, where, and how patients spend significant portions of time and opportunities for adapting patients’ day-to-day health activities. Conclusions: This study demonstrates the potential for interventions to integrate tools for collecting and communicating spatial and temporal contexts of patient routines, such as the types of activities that characterize how patients spend significant portions of time and identification of when, where, and how to encourage health-promoting changes in routine activities. Time-use patterns provide insight for tailoring behavior change interventions so that clinic-based settings are generalizable to the contexts of where, when, and how patient routines could be adapted to mitigate cardiovascular risk factors.
- 27 August 2024
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 21 (9) , 1128
Open AccessTime-Use Sequences: A Mixed-Methods Study Exploring How, When, and Where Spatiotemporal Patterns of Everyday Routines Can Strengthen Public Health Interventions
Background: Behavior change interventions are critical for the secondary prevention of cardiovascular disease and for reducing the risk of a repeat event or mortality. However, the effectiveness of behavior change interventions is challenged by a lack of spatiotemporal contexts, limiting our understanding of factors that influence the timing and location in which day-to-day activities occur and the maintenance of behavior change. This study explored how behavior change interventions could incorporate spatiotemporal contexts of patient activities for modifying behaviors. Methods: A mixed-methods approach with adapted geo-ethnography techniques was used to solicit detailed descriptions of patients’ day-to-day routines, including where, when, and how patients spend time. Data were gathered from patients in one cardiac intervention program in Nova Scotia, Canada, from June to September 2021. Results: A total of 29 individuals (19 men and 10 women) between the ages of 45 and 81 and referred to the program after a cardiac event participated. The results show three key findings: (1) most patients exceeded the minimum guidelines of 30 min of daily physical activity but were sedentary for long periods of time, (2) patient time-use patterns are heterogenous and unique to contexts of individual space-time activity paths, and (3) time-use patterns reveal when, where, and how patients spend significant portions of time and opportunities for adapting patients’ day-to-day health activities. Conclusions: This study demonstrates the potential for interventions to integrate tools for collecting and communicating spatial and temporal contexts of patient routines, such as the types of activities that characterize how patients spend significant portions of time and identification of when, where, and how to encourage health-promoting changes in routine activities. Time-use patterns provide insight for tailoring behavior change interventions so that clinic-based settings are generalizable to the contexts of where, when, and how patient routines could be adapted to mitigate cardiovascular risk factors.
- Open Access
- 24 October 2023
- Published by Oxford University Press (OUP) in Translational Behavioral Medicine
- Vol. 14 (5) , 257-272
Integrating health geography and behavioral economic principles to strengthen context-specific behavior change interventions
The long-term economic viability of modern health care systems is uncertain, in part due to costs of health care at the end of life and increasing health care utilization associated with an increasing population prevalence of multiple chronic diseases. Control of health care spending and sustaining delivery of health care services will require strategic investments in prevention to reduce the risk of disease and its complications over an individual’s life course. Behavior change interventions aimed at reducing a range of harmful and risky health-related behaviors including smoking, physical inactivity, excess alcohol consumption, and excess weight, are one approach that has proven effective at reducing risk and preventing chronic disease. However, large-scale efforts to reduce population-level chronic diseases are challenging and have not been very successful at reducing the burden of chronic diseases. A new approach is required to identify when, where, and how to intervene to disrupt patterns of behavior associated with high-risk factors using context-specific interventions that can be scaled. This paper introduces the need to integrate theoretical and methodological principles of health geography and behavioral economics as opportunities to strengthen behavior change interventions for the prevention of chronic diseases. We discuss how health geography and behavioral economics can be applied to expand existing behavior change frameworks and how behavior change interventions can be strengthened by characterizing contexts of time and activity space.
- 24 October 2023
- Published by Oxford University Press (OUP) in Translational Behavioral Medicine
- Vol. 14 (5) , 257-272
Open AccessIntegrating health geography and behavioral economic principles to strengthen context-specific behavior change interventions
The long-term economic viability of modern health care systems is uncertain, in part due to costs of health care at the end of life and increasing health care utilization associated with an increasing population prevalence of multiple chronic diseases. Control of health care spending and sustaining delivery of health care services will require strategic investments in prevention to reduce the risk of disease and its complications over an individual’s life course. Behavior change interventions aimed at reducing a range of harmful and risky health-related behaviors including smoking, physical inactivity, excess alcohol consumption, and excess weight, are one approach that has proven effective at reducing risk and preventing chronic disease. However, large-scale efforts to reduce population-level chronic diseases are challenging and have not been very successful at reducing the burden of chronic diseases. A new approach is required to identify when, where, and how to intervene to disrupt patterns of behavior associated with high-risk factors using context-specific interventions that can be scaled. This paper introduces the need to integrate theoretical and methodological principles of health geography and behavioral economics as opportunities to strengthen behavior change interventions for the prevention of chronic diseases. We discuss how health geography and behavioral economics can be applied to expand existing behavior change frameworks and how behavior change interventions can be strengthened by characterizing contexts of time and activity space.
- Open Access
- 27 June 2022
- Published by Springer Nature in Quality of Life Research
- Vol. 31 (9) , 2867-2880
Item selection, scaling and construct validation of the Patient-Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) measurement tool in adults
Purpose: To select and scale items for the seven domains of the Patient-Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) and assess its construct validity. Methods: Using an online survey, data on 100 potential items, and other variables for assessing construct validity, were collected from 1055 adults with one or more chronic health conditions. Based on a validated conceptual model, confirmatory factor analysis (CFA) and item response models (IRT) were used to select and scale potential items and assess the internal consistency and structural validity of the PRISM-CC. To further assess construct validity, hypothesis testing of known relationships was conducted using structural equation models. Results: Of 100 potential items, 36 (4–8 per domain) were selected, providing excellent fit to our hypothesized correlated factors model and demonstrating internal consistency and structural validity of the PRISM-CC. Hypothesized associations between PRISM-CC domains and other measures and variables were confirmed, providing further evidence of construct validity. Conclusion: The PRISM-CC overcomes limitations of assessment tools currently available to measure patient self-management of chronic health conditions. This study provides strong evidence for the internal consistency and construct validity of the PRISM-CC as an instrument to assess patient-reported difficulty in self-managing different aspects of daily life with one or more chronic conditions. Further research is needed to assess its measurement equivalence across patient attributes, ability to measure clinically important change, and utility to inform self-management support.
- 27 June 2022
- Published by Springer Nature in Quality of Life Research
- Vol. 31 (9) , 2867-2880
Open AccessItem selection, scaling and construct validation of the Patient-Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) measurement tool in adults
Purpose: To select and scale items for the seven domains of the Patient-Reported Inventory of Self-Management of Chronic Conditions (PRISM-CC) and assess its construct validity. Methods: Using an online survey, data on 100 potential items, and other variables for assessing construct validity, were collected from 1055 adults with one or more chronic health conditions. Based on a validated conceptual model, confirmatory factor analysis (CFA) and item response models (IRT) were used to select and scale potential items and assess the internal consistency and structural validity of the PRISM-CC. To further assess construct validity, hypothesis testing of known relationships was conducted using structural equation models. Results: Of 100 potential items, 36 (4–8 per domain) were selected, providing excellent fit to our hypothesized correlated factors model and demonstrating internal consistency and structural validity of the PRISM-CC. Hypothesized associations between PRISM-CC domains and other measures and variables were confirmed, providing further evidence of construct validity. Conclusion: The PRISM-CC overcomes limitations of assessment tools currently available to measure patient self-management of chronic health conditions. This study provides strong evidence for the internal consistency and construct validity of the PRISM-CC as an instrument to assess patient-reported difficulty in self-managing different aspects of daily life with one or more chronic conditions. Further research is needed to assess its measurement equivalence across patient attributes, ability to measure clinically important change, and utility to inform self-management support.
- Open Access
- 27 November 2021
- Published by Springer Nature in BMC Medical Research Methodology
- Vol. 21 (1) , 1-14
Predicting COVID-19 mortality risk in Toronto, Canada: a comparison of tree-based and regression-based machine learning methods
Background: Coronavirus disease (COVID-19) presents an unprecedented threat to global health worldwide. Accurately predicting the mortality risk among the infected individuals is crucial for prioritizing medical care and mitigating the healthcare system’s burden. The present study aimed to assess the predictive accuracy of machine learning methods to predict the COVID-19 mortality risk.Methods: We compared the performance of classification tree, random forest (RF), extreme gradient boosting (XGBoost), logistic regression, generalized additive model (GAM) and linear discriminant analysis (LDA) to predict the mortality risk among 49,216 COVID-19 positive cases in Toronto, Canada, reported from March 1 to December 10, 2020. We used repeated split-sample validation andk-steps-ahead forecasting validation. Predictive models were estimated using training samples, and predictive accuracy of the methods for the testing samples was assessed using the area under the receiver operating characteristic curve, Brier’s score, calibration intercept and calibration slope.Results: We found XGBoost is highly discriminative, with an AUC of 0.9669 and has superior performance over conventional tree-based methods, i.e., classification tree or RF methods for predicting COVID-19 mortality risk. Regression-based methods (logistic, GAM and LASSO) had comparable performance to the XGBoost with slightly lower AUCs and higher Brier’s scores.Conclusions: XGBoost offers superior performance over conventional tree-based methods and minor improvement over regression-based methods for predicting COVID-19 mortality risk in the study population.
- 27 November 2021
- Published by Springer Nature in BMC Medical Research Methodology
- Vol. 21 (1) , 1-14
Open AccessPredicting COVID-19 mortality risk in Toronto, Canada: a comparison of tree-based and regression-based machine learning methods
Background: Coronavirus disease (COVID-19) presents an unprecedented threat to global health worldwide. Accurately predicting the mortality risk among the infected individuals is crucial for prioritizing medical care and mitigating the healthcare system’s burden. The present study aimed to assess the predictive accuracy of machine learning methods to predict the COVID-19 mortality risk.Methods: We compared the performance of classification tree, random forest (RF), extreme gradient boosting (XGBoost), logistic regression, generalized additive model (GAM) and linear discriminant analysis (LDA) to predict the mortality risk among 49,216 COVID-19 positive cases in Toronto, Canada, reported from March 1 to December 10, 2020. We used repeated split-sample validation andk-steps-ahead forecasting validation. Predictive models were estimated using training samples, and predictive accuracy of the methods for the testing samples was assessed using the area under the receiver operating characteristic curve, Brier’s score, calibration intercept and calibration slope.Results: We found XGBoost is highly discriminative, with an AUC of 0.9669 and has superior performance over conventional tree-based methods, i.e., classification tree or RF methods for predicting COVID-19 mortality risk. Regression-based methods (logistic, GAM and LASSO) had comparable performance to the XGBoost with slightly lower AUCs and higher Brier’s scores.Conclusions: XGBoost offers superior performance over conventional tree-based methods and minor improvement over regression-based methods for predicting COVID-19 mortality risk in the study population.
- Open Access
- 5 June 2021
- Published by MDPI AG in Current Oncology
- Vol. 28 (3) , 2107-2114
Nurse Navigators’ Views on Patient and System Factors Associated with Navigation Needs among Women with Breast Cancer
Coordinating breast cancer treatment is a complex task that can overwhelm patients and their support networks. Though the Cancer Patient Navigator (CPN) program in Nova Scotia (NS) provides professional assistance to patients, certain groups of patients may still face barriers to accessing its services. Employing interviews and a modified Delphi approach with CPN participants, this study sought to identify factors associated with the need for navigation to help better target CPN program referrals among breast cancer patients. Six CPNs were recruited directly through the CPN program manager for interviews and surveys. The CPNs identified 27 different factors, which were divided into 4 categories: sociodemographic, psychological, clinical and health systems. While these patient factors (particularly sociodemographic) are not directly modifiable, awareness of their association with the need for navigation could be used to better target patients with a high need for navigation for referral to CPN services.
- 5 June 2021
- Published by MDPI AG in Current Oncology
- Vol. 28 (3) , 2107-2114
Open AccessNurse Navigators’ Views on Patient and System Factors Associated with Navigation Needs among Women with Breast Cancer
Coordinating breast cancer treatment is a complex task that can overwhelm patients and their support networks. Though the Cancer Patient Navigator (CPN) program in Nova Scotia (NS) provides professional assistance to patients, certain groups of patients may still face barriers to accessing its services. Employing interviews and a modified Delphi approach with CPN participants, this study sought to identify factors associated with the need for navigation to help better target CPN program referrals among breast cancer patients. Six CPNs were recruited directly through the CPN program manager for interviews and surveys. The CPNs identified 27 different factors, which were divided into 4 categories: sociodemographic, psychological, clinical and health systems. While these patient factors (particularly sociodemographic) are not directly modifiable, awareness of their association with the need for navigation could be used to better target patients with a high need for navigation for referral to CPN services.