Jeanne A Teresi1, Katja Ocepek-Welikson2, Mildred Ramirez3, Marjorie Kleinman4, Katherine Ornstein5, Albert Siu6. 1. Research Division, The Hebrew Home at Riverdale, Riverdale, NY, USA Measurement and Data Management Core, Mount Sinai Pepper Older Americans Independence Center, Mount Sinai Medical Center, New York, NY, USA Columbia University Stroud Center, New York State Psychiatric Institute, New York, NY, USA. 2. Research Division, The Hebrew Home at Riverdale, Riverdale, NY, USA. 3. Research Division, The Hebrew Home at Riverdale, Riverdale, NY, USA Measurement and Data Management Core, Mount Sinai Pepper Older Americans Independence Center, Mount Sinai Medical Center, New York, NY, USA Division of Geriatrics and Palliative Care, Weill Cornell Medical Center, New York, NY, USA MilRamirez@aol.com. 4. Columbia University Stroud Center, New York State Psychiatric Institute, New York, NY, USA. 5. Department of Geriatrics and Palliative Medicine, Institute for Translational Epidemiology, Mount Sinai School of Medicine, New York, NY, USA. 6. Divisions of Geriatrics and Palliative Medicine, General Internal Medicine, Health Evidence and Policy, Mount Sinai Medical Center, New York, NY, USA.
Abstract
BACKGROUND: The Family Satisfaction with End-of-Life Care is an internationally used measure of satisfaction with cancer care. However, the Family Satisfaction with End-of-Life Care has not been studied for equivalence of item endorsement across different socio-demographic groups using differential item functioning. AIMS: The aims of this secondary data analysis were (1) to examine potential differential item functioning in the family satisfaction item set with respect to type of caregiver, race, and patient age, gender, and education and (2) to provide parameters and documentation of differential item functioning for an item bank. DESIGN: A mixed qualitative and quantitative analysis was conducted. A priori hypotheses regarding potential group differences in item response were established. Item response theory and Wald tests were used for the analyses of differential item functioning, accompanied by magnitude and impact measures. RESULTS: Very little significant differential item functioning was observed for patient's age and gender. For race, 13 items showed differential item functioning after multiple comparison adjustment, 10 with non-uniform differential item functioning. No items evidenced differential item functioning of high magnitude, and the impact was negligible. For education, 5 items evidenced uniform differential item functioning after adjustment, none of high magnitude. Differential item functioning impact was trivial. One item evidenced differential item functioning for the caregiver relationship variable. CONCLUSION: Differential item functioning was observed primarily for race and education. No differential item functioning of high magnitude was observed for any item, and the overall impact of differential item functioning was negligible. One item, satisfaction with "the patient's pain relief," might be singled out for further study, given that this item was both hypothesized and observed to show differential item functioning for race and education.
BACKGROUND: The Family Satisfaction with End-of-Life Care is an internationally used measure of satisfaction with cancer care. However, the Family Satisfaction with End-of-Life Care has not been studied for equivalence of item endorsement across different socio-demographic groups using differential item functioning. AIMS: The aims of this secondary data analysis were (1) to examine potential differential item functioning in the family satisfaction item set with respect to type of caregiver, race, and patient age, gender, and education and (2) to provide parameters and documentation of differential item functioning for an item bank. DESIGN: A mixed qualitative and quantitative analysis was conducted. A priori hypotheses regarding potential group differences in item response were established. Item response theory and Wald tests were used for the analyses of differential item functioning, accompanied by magnitude and impact measures. RESULTS: Very little significant differential item functioning was observed for patient's age and gender. For race, 13 items showed differential item functioning after multiple comparison adjustment, 10 with non-uniform differential item functioning. No items evidenced differential item functioning of high magnitude, and the impact was negligible. For education, 5 items evidenced uniform differential item functioning after adjustment, none of high magnitude. Differential item functioning impact was trivial. One item evidenced differential item functioning for the caregiver relationship variable. CONCLUSION: Differential item functioning was observed primarily for race and education. No differential item functioning of high magnitude was observed for any item, and the overall impact of differential item functioning was negligible. One item, satisfaction with "the patient's pain relief," might be singled out for further study, given that this item was both hypothesized and observed to show differential item functioning for race and education.
Authors: Katherine A Ornstein; Joan Penrod; Julie B Schnur; Cardinale B Smith; Jeanne A Teresi; Melissa M Garrido; Karen McKendrick; Albert L Siu; Diane E Meier; R Sean Morrison Journal: J Palliat Med Date: 2017-02-10 Impact factor: 2.947
Authors: Jeanne A Teresi; Katja Ocepek-Welikson; Mildred Ramirez; Marjorie Kleinman; Katherine Ornstein; Albert Siu; Jose Luchsinger Journal: Palliat Support Care Date: 2020-10