Haofei Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Haofei Li has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Weidong Huang
School of Health Management, Harbin Medical University, Harbin, China
2 shared publications - Yiyin Cao
School of Health Management, Harbin Medical University, Harbin, China
2 shared publications David Cella
Feinberg School of Medicine, Northwestern University, Chicago, USA
1 shared publicationDongfu Qian
School of Health Policy and Management, Nanjing Medical University, Nanjing, People’s Republic of China
1 shared publicationGeorg Kemmler
Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Innsbruck, Austria
1 shared publication- Hongjuan Yu
Department of Hematology, The First Affiliated Hospital of Harbin Medical University, Harbin, China
1 shared publication
Latest Publications
- Open Access
- 1 September 2026
- Published by Taylor & Francis in Patient Preference and Adherence
- Vol. ume 20, 1-17
Pro-Rich Inequality in Health-Related Quality of Life Among Primary Care Diabetes Patients: A Cross-Sectional Study
Objective: This cross-sectional study aimed to assess HRQoL and to examine income-related inequalities among diabetes patients in Harbin, a cold-region city in China, where such evidence remains limited. Methods: In July 2025, 433 adults with diabetes were recruited from two community health centers in Harbin using random sampling from electronic health records. HRQoL was measured using the EuroQol 5-Dimension 5-Level (EQ-5D-5L) utility index and the EuroQol Visual Analogue Scale (EQ-VAS). Income-related inequality was quantified using the concentration index (CI) and the Erreygers-corrected CI. Wagstaff decomposition was applied to identify contributing factors, and horizontal inequity (HI) indices were computed after adjusting for need variables. Sensitivity analyses included Tobit model. Results: The mean EQ-5D-5L utility index was 0.9303 (SD = 0.1294), and the mean EQ-VAS score was 77.60 (SD = 12.88). Significant pro-rich inequality was observed for both outcomes (standard CI: 0.0093 for EQ-5D-5L and 0.0131 for EQ-VAS; Erreygers index: 0.0249 and 0.0405, respectively). Household income was the largest contributor to inequality, accounting for 73.30% of the inequality in EQ-5D-5L utility index and 53.77% of the inequality in EQ-VAS score. Other notable contributors included blood glucose control, exercise habits, marital status, and digital health literacy for EQ-5D-5L, and adherence to chronic disease follow-up and exercise habits for EQ-VAS. After adjusting for need variables, the HI remained positive for both measures (0.0095 for EQ-5D-5L and 0.0135 for EQ-VAS). Conclusion: Income-related inequalities in HRQoL were evident among primary care diabetes patients in this cold-region setting. Socioeconomic disparities and modifiable health behaviors are important correlates of HRQoL inequality, warranting further investigation in longitudinal studies.
- 1 September 2026
- Published by Taylor & Francis in Patient Preference and Adherence
- Vol. ume 20, 1-17
Open AccessPro-Rich Inequality in Health-Related Quality of Life Among Primary Care Diabetes Patients: A Cross-Sectional Study
Objective: This cross-sectional study aimed to assess HRQoL and to examine income-related inequalities among diabetes patients in Harbin, a cold-region city in China, where such evidence remains limited. Methods: In July 2025, 433 adults with diabetes were recruited from two community health centers in Harbin using random sampling from electronic health records. HRQoL was measured using the EuroQol 5-Dimension 5-Level (EQ-5D-5L) utility index and the EuroQol Visual Analogue Scale (EQ-VAS). Income-related inequality was quantified using the concentration index (CI) and the Erreygers-corrected CI. Wagstaff decomposition was applied to identify contributing factors, and horizontal inequity (HI) indices were computed after adjusting for need variables. Sensitivity analyses included Tobit model. Results: The mean EQ-5D-5L utility index was 0.9303 (SD = 0.1294), and the mean EQ-VAS score was 77.60 (SD = 12.88). Significant pro-rich inequality was observed for both outcomes (standard CI: 0.0093 for EQ-5D-5L and 0.0131 for EQ-VAS; Erreygers index: 0.0249 and 0.0405, respectively). Household income was the largest contributor to inequality, accounting for 73.30% of the inequality in EQ-5D-5L utility index and 53.77% of the inequality in EQ-VAS score. Other notable contributors included blood glucose control, exercise habits, marital status, and digital health literacy for EQ-5D-5L, and adherence to chronic disease follow-up and exercise habits for EQ-VAS. After adjusting for need variables, the HI remained positive for both measures (0.0095 for EQ-5D-5L and 0.0135 for EQ-VAS). Conclusion: Income-related inequalities in HRQoL were evident among primary care diabetes patients in this cold-region setting. Socioeconomic disparities and modifiable health behaviors are important correlates of HRQoL inequality, warranting further investigation in longitudinal studies.
- Open Access
- 1 October 2024
- Published by Springer Nature in Health Economics Review
- Vol. 14 (1) , 1-10
A comparison of measurement properties between EORTC QLU-C10D and FACT-8D in patients with hematological malignances
To perform a comparison of the measurement properties of two cancer-specific Multi-Attribute Utility Instruments (MAUIs), EORTC QLU-C10D and FACT-8D, in Chinese patients with hematologic malignancies (HM). We conducted a longitudinal study on patients with HM in China, using QLU-C10D and FACT-8D at baseline and follow-up (3–4 months from baseline). We assessed: (i) convergent validity using Spearman’s rank correlation test (r) with EQ-5D-5L; (ii) clinical-groups validity by differentiating cancer stages, overall health assessment (OHA), Eastern Cancer Oncology Group (ECOG) performance status, and mental health status. We also examined clinical validity with effect size (ES) and relative efficiency (RE); (iii) responsiveness to changes in patient self-perception using receiver operating characteristics (ROC) curves and area under the curves (AUC); and (iv) agreement using intraclass correlation coefficients (ICC) and visualized with Bland-Altman plot. Among the 308 patients with HM at baseline, 131 completed the follow-up survey. Agreement between the two measures was high (ICC = 0.76). Both measures were highly correlated with EQ-5D-5 L and significantly differentiated (p < 0.001) among groups categorized by cancer stage, OHA performance status, and mental health. ESs for QLU-C10D were numerically higher for cancer stage, OHA, and performance status (ES = 0.53–1.49), whereas ES was higher for FACT-8D and mental health status (ES = 1.35). Responsiveness was higher for QLU-C10D (AUC = 0.84) compared to FACT-8D (AUC = 0.78). Both QLU-C10D and FACT-8D are valid cancer-specific MAUIs for evaluating patients with HM. However, scholars should consider their slight differences in focus when choosing between the two measures.
- 1 October 2024
- Published by Springer Nature in Health Economics Review
- Vol. 14 (1) , 1-10
Open AccessA comparison of measurement properties between EORTC QLU-C10D and FACT-8D in patients with hematological malignances
To perform a comparison of the measurement properties of two cancer-specific Multi-Attribute Utility Instruments (MAUIs), EORTC QLU-C10D and FACT-8D, in Chinese patients with hematologic malignancies (HM). We conducted a longitudinal study on patients with HM in China, using QLU-C10D and FACT-8D at baseline and follow-up (3–4 months from baseline). We assessed: (i) convergent validity using Spearman’s rank correlation test (r) with EQ-5D-5L; (ii) clinical-groups validity by differentiating cancer stages, overall health assessment (OHA), Eastern Cancer Oncology Group (ECOG) performance status, and mental health status. We also examined clinical validity with effect size (ES) and relative efficiency (RE); (iii) responsiveness to changes in patient self-perception using receiver operating characteristics (ROC) curves and area under the curves (AUC); and (iv) agreement using intraclass correlation coefficients (ICC) and visualized with Bland-Altman plot. Among the 308 patients with HM at baseline, 131 completed the follow-up survey. Agreement between the two measures was high (ICC = 0.76). Both measures were highly correlated with EQ-5D-5 L and significantly differentiated (p < 0.001) among groups categorized by cancer stage, OHA performance status, and mental health. ESs for QLU-C10D were numerically higher for cancer stage, OHA, and performance status (ES = 0.53–1.49), whereas ES was higher for FACT-8D and mental health status (ES = 1.35). Responsiveness was higher for QLU-C10D (AUC = 0.84) compared to FACT-8D (AUC = 0.78). Both QLU-C10D and FACT-8D are valid cancer-specific MAUIs for evaluating patients with HM. However, scholars should consider their slight differences in focus when choosing between the two measures.