Norah L Crossnohere1,2, Sarah Janse3, Ellen Janssen4, John F P Bridges3,5. 1. Department of Biomedical Informatics, The Ohio State University College of Medicine, 1800 Cannon Drive, Columbus, OH, 43210, USA. norah.crossnohere@osumc.edu. 2. Department of Health, Behavior, and Society, Johns Hopkins Bloomberg School of Public Health, 624 N. Broadway, Baltimore, MD, 21205, USA. norah.crossnohere@osumc.edu. 3. Department of Biomedical Informatics, The Ohio State University College of Medicine, 1800 Cannon Drive, Columbus, OH, 43210, USA. 4. Center for Medical Technology Policy, World Trade Center Baltimore, 401 East Pratt Street, Suite 631, Baltimore, MD, 21202, USA. 5. Department of Health, Behavior, and Society, Johns Hopkins Bloomberg School of Public Health, 624 N. Broadway, Baltimore, MD, 21205, USA.
Abstract
BACKGROUND: Healthcare treatments and interventions are traditionally evaluated from the societal perspective, but a more patient-centric perspective has been proposed in recent years. We sought to compare preferences of patients and the general public for treatment outcomes of type 2 diabetes using both best-worst scaling (BWS) and rating approaches. METHODS: A survey evaluating the treatment priorities for type 2 diabetes was conducted in the United States. Members of the general public and patients with type 2 diabetes were recruited from a nationally sampled panel. Participants indicated the importance of seven potential treatment outcomes (hypoglycemic events, glycated hemoglobin [A1c], weight loss, mental health, functioning, glycemic stability, and cardiovascular health) using (1) BWS case 1 and (2) a rating task. Preference differences from BWS prioritizations were explored using mixed logistic regression (BWS preference weights were probability re-scaled so that the weightings of the seven items collectively summed to 100). The consistency of scale between samples was explored using heteroskedastic conditional logistic regression of BWS data. Spearman rank correlation was used to compare standardized BWS preference weights and rating scores for each group. Both groups evaluated the BWS and rating activities using debriefing questions. RESULTS: The public and patient samples included 314 and 313 respondents, respectively. The public was on average 16 years younger than patients (48 vs 64 years, P < 0.001). In BWS, patients and the public both ranked A1c, glycemic stability, and cardiovascular health within their top three outcomes. Patients valued the outcome A1c most highly and found it twice as important as did the public (41.0 vs 20.2, P < 0.001). The public valued cardiovascular health most highly, and found it to be twice as important than did patients (31.3 vs 17.4, P < 0.001). Patients were more consistent in their preferences than the public (λ = 1.66, P = 0.01). Preferences elicited using BWS and rating approaches were highly correlated for both patients (ρ = 0.96) and the public (ρ = 0.92). Patients were more likely than the public to endorse the BWS as easy to answer (P < 0.001), easy to understand (P < 0.001), consistent with preferences (P < 0.001), and relevant (P < 0.001). Both patients and the public found the rating activity easier to answer and understand, and more consistent with their preferences, than the BWS (P < 0.001). CONCLUSIONS: We provide some of the first evidence demonstrating a difference in patient and public treatment priorities for diabetes. That patients were more consistent in their preferences than the public and found the BWS and Likert rating instruments more relevant suggests that patient priorities may be more appropriate than those of the general public in some medical decision-making contexts.
BACKGROUND: Healthcare treatments and interventions are traditionally evaluated from the societal perspective, but a more patient-centric perspective has been proposed in recent years. We sought to compare preferences of patients and the general public for treatment outcomes of type 2 diabetes using both best-worst scaling (BWS) and rating approaches. METHODS: A survey evaluating the treatment priorities for type 2 diabetes was conducted in the United States. Members of the general public and patients with type 2 diabetes were recruited from a nationally sampled panel. Participants indicated the importance of seven potential treatment outcomes (hypoglycemic events, glycated hemoglobin [A1c], weight loss, mental health, functioning, glycemic stability, and cardiovascular health) using (1) BWS case 1 and (2) a rating task. Preference differences from BWS prioritizations were explored using mixed logistic regression (BWS preference weights were probability re-scaled so that the weightings of the seven items collectively summed to 100). The consistency of scale between samples was explored using heteroskedastic conditional logistic regression of BWS data. Spearman rank correlation was used to compare standardized BWS preference weights and rating scores for each group. Both groups evaluated the BWS and rating activities using debriefing questions. RESULTS: The public and patient samples included 314 and 313 respondents, respectively. The public was on average 16 years younger than patients (48 vs 64 years, P < 0.001). In BWS, patients and the public both ranked A1c, glycemic stability, and cardiovascular health within their top three outcomes. Patients valued the outcome A1c most highly and found it twice as important as did the public (41.0 vs 20.2, P < 0.001). The public valued cardiovascular health most highly, and found it to be twice as important than did patients (31.3 vs 17.4, P < 0.001). Patients were more consistent in their preferences than the public (λ = 1.66, P = 0.01). Preferences elicited using BWS and rating approaches were highly correlated for both patients (ρ = 0.96) and the public (ρ = 0.92). Patients were more likely than the public to endorse the BWS as easy to answer (P < 0.001), easy to understand (P < 0.001), consistent with preferences (P < 0.001), and relevant (P < 0.001). Both patients and the public found the rating activity easier to answer and understand, and more consistent with their preferences, than the BWS (P < 0.001). CONCLUSIONS: We provide some of the first evidence demonstrating a difference in patient and public treatment priorities for diabetes. That patients were more consistent in their preferences than the public and found the BWS and Likert rating instruments more relevant suggests that patient priorities may be more appropriate than those of the general public in some medical decision-making contexts.
Authors: Mihir Gandhi; Ru San Tan; Raymond Ng; Su Pin Choo; Whay Kuang Chia; Chee Keong Toh; Carolyn Lam; Phong Teck Lee; Nang Khaing Zar Latt; Kim Rand-Hendriksen; Yin Bun Cheung; Nan Luo Journal: Qual Life Res Date: 2017-08-14 Impact factor: 4.147
Authors: Eline van Overbeeke; Rosanne Janssens; Chiara Whichello; Karin Schölin Bywall; Jenny Sharpe; Nikoletta Nikolenko; Berkeley S Phillips; Paolo Guiddi; Gabriella Pravettoni; Laura Vergani; Giulia Marton; Irina Cleemput; Steven Simoens; Jürgen Kübler; Juhaeri Juhaeri; Bennett Levitan; Esther W de Bekker-Grob; Jorien Veldwijk; Isabelle Huys Journal: Front Pharmacol Date: 2019-12-03 Impact factor: 5.810