Lili Tian
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Lili Tian has a total of 68 co-authors. Here are the top co-authors by number of shared publications:
- Deborah O. Erwin
Roswell Park Comprehensive Cancer Center, Cancer Prevention and Control, Buffalo, USA
14 shared publications - Furrina F. Lee
Bureau of Cancer Epidemiology, Division of Chronic Disease Prevention, New York State Department of Health, Albany, USA
14 shared publications - Heather M. Ochs-Balcom
Department of Epidemiology and Environmental Health, School of Public Health and Health Professions, University at Buffalo, Buffalo, USA
14 shared publications - Helen C. S. Meier
Survey Research Center, Institute for Social Research, University of Michigan, Ann Arbor, USA
14 shared publications - Henry Louis Taylor
University at Buffalo
14 shared publications - Sarah M. Lima
Department of Epidemiology and Environmental Health, School of Public Health and Health Professions, University at Buffalo, Buffalo, USA
14 shared publications
Latest Publications
- 11 November 2026
- Published by Wiley in Pharmaceutical Statistics
- Vol. 25 (6)
A New Optimal Cut‐Points Selection Method for Diseases With Multiple Ordinal Stages and Its Confidence Interval Estimation
Optimal cut‐points selection plays a fundamental role in multi‐stage classification. The popular Youden index method primarily emphasizes overall diagnostic accuracy while overlooking imbalance among classification rates across disease stages, whereas other methods improve balance at the cost of substantially reduced overall accuracy. Such trade‐offs are undesirable in clinical decision‐making. This article aims to develop a flexible optimal cut‐points selection method that simultaneously preserves overall diagnostic accuracy and balance among stage‐specific classification rates. By introducing a tuning parameter, the proposed method integrates the arithmetic mean and the geometric mean of stage‐specific classification rates into a unified objective function. Estimation procedures for confidence intervals are developed under both parametric and nonparametric settings. Simulation studies are conducted to evaluate the performance of the proposed cut‐points selection method and the confidence interval estimation approaches. The proposed method is applied to a real pancreatic cancer dataset to illustrate its practical utility.
- 11 November 2026
- Published by Wiley in Pharmaceutical Statistics
- Vol. 25 (6)
A New Optimal Cut‐Points Selection Method for Diseases With Multiple Ordinal Stages and Its Confidence Interval Estimation
Optimal cut‐points selection plays a fundamental role in multi‐stage classification. The popular Youden index method primarily emphasizes overall diagnostic accuracy while overlooking imbalance among classification rates across disease stages, whereas other methods improve balance at the cost of substantially reduced overall accuracy. Such trade‐offs are undesirable in clinical decision‐making. This article aims to develop a flexible optimal cut‐points selection method that simultaneously preserves overall diagnostic accuracy and balance among stage‐specific classification rates. By introducing a tuning parameter, the proposed method integrates the arithmetic mean and the geometric mean of stage‐specific classification rates into a unified objective function. Estimation procedures for confidence intervals are developed under both parametric and nonparametric settings. Simulation studies are conducted to evaluate the performance of the proposed cut‐points selection method and the confidence interval estimation approaches. The proposed method is applied to a real pancreatic cancer dataset to illustrate its practical utility.
- 14 October 2026
- Published by Springer Nature in Journal of Urban Health
- p. 1-16
After the Maps: Assessing Longitudinal Neighborhood Change After Historical Redlining with Present-Day Breast Cancer Survival
Historical redlining is a 1930s residential segregation policy that graded neighborhoods according to race, class, and land-use and has been linked to worse breast cancer outcomes today. Since redlining grades were determined by socioeconomic and environmental attributes, evaluating neighborhood change presents a unique opportunity to study the relative contributions of socioeconomic and physical environment to breast cancer. We evaluated whether longitudinal neighborhood change, specifically gentrification and disinvestment, is associated with survival among breast cancer cases today. We used historical redlining grade and labor class census data from 1940 to 2010 to create neighborhood change typologies: stable high neighborhood socioeconomic status (NSES), disinvested, gentrified, and stable low NSES; additional criteria were used for a second, stricter definition of stable types (“truly stable”). Typologies were linked to 43,023 breast cancer cases diagnosed 2008–2018 from New York State. Cox models estimated hazard ratios (HR) associated with neighborhood change types, using both stable high NSES and stable low NSES as referents in separate models. We found significant survival differences according to neighborhood change type. All neighborhood change typologies had higher mortality risks compared to stable high NSES, regardless of how stable NSES was defined. This was consistent in models using stable low NSES as the referent when stable NSES was defined broadly, where only stable high NSES was associated with higher survival (HR = 0.73, 95% CI 0.67–0.80) and gentrification and disinvestment did not show differential mortality risks. However, when stable NSES is defined conservatively (truly stable NSES), we found evidence of survival benefits for disinvestment (HR = 0.76, 95% CI 0.65, 0.89) and gentrification (HR = 0.76, 95% CI 0.67, 0.87) compared to truly stable low NSES neighborhoods, and the survival benefit for truly stable high NSES was strengthened (HR = 0.43, 95% CI 0.35, 0.52), though sample size is reduced in the truly stable NSES models. Our results suggest survival is always highest among consistently high NSES neighborhoods but that gentrification and disinvestment may offer buffers compared to redlined neighborhoods with consistently low NSES. This study demonstrates the complexity of neighborhood change and suggests contemporary socioeconomic conditions and underlying physical environment may both be significant domains for breast cancer outcomes.
- 14 October 2026
- Published by Springer Nature in Journal of Urban Health
- p. 1-16
After the Maps: Assessing Longitudinal Neighborhood Change After Historical Redlining with Present-Day Breast Cancer Survival
Historical redlining is a 1930s residential segregation policy that graded neighborhoods according to race, class, and land-use and has been linked to worse breast cancer outcomes today. Since redlining grades were determined by socioeconomic and environmental attributes, evaluating neighborhood change presents a unique opportunity to study the relative contributions of socioeconomic and physical environment to breast cancer. We evaluated whether longitudinal neighborhood change, specifically gentrification and disinvestment, is associated with survival among breast cancer cases today. We used historical redlining grade and labor class census data from 1940 to 2010 to create neighborhood change typologies: stable high neighborhood socioeconomic status (NSES), disinvested, gentrified, and stable low NSES; additional criteria were used for a second, stricter definition of stable types (“truly stable”). Typologies were linked to 43,023 breast cancer cases diagnosed 2008–2018 from New York State. Cox models estimated hazard ratios (HR) associated with neighborhood change types, using both stable high NSES and stable low NSES as referents in separate models. We found significant survival differences according to neighborhood change type. All neighborhood change typologies had higher mortality risks compared to stable high NSES, regardless of how stable NSES was defined. This was consistent in models using stable low NSES as the referent when stable NSES was defined broadly, where only stable high NSES was associated with higher survival (HR = 0.73, 95% CI 0.67–0.80) and gentrification and disinvestment did not show differential mortality risks. However, when stable NSES is defined conservatively (truly stable NSES), we found evidence of survival benefits for disinvestment (HR = 0.76, 95% CI 0.65, 0.89) and gentrification (HR = 0.76, 95% CI 0.67, 0.87) compared to truly stable low NSES neighborhoods, and the survival benefit for truly stable high NSES was strengthened (HR = 0.43, 95% CI 0.35, 0.52), though sample size is reduced in the truly stable NSES models. Our results suggest survival is always highest among consistently high NSES neighborhoods but that gentrification and disinvestment may offer buffers compared to redlined neighborhoods with consistently low NSES. This study demonstrates the complexity of neighborhood change and suggests contemporary socioeconomic conditions and underlying physical environment may both be significant domains for breast cancer outcomes.
- 15 September 2026
- Published by Elsevier in Annals of Epidemiology
- Vol. 122
Long-term ambient temperature and temperature variability in relation to incident hypertension among postmenopausal women
- 15 September 2026
- Published by Elsevier in Annals of Epidemiology
- Vol. 122
Long-term ambient temperature and temperature variability in relation to incident hypertension among postmenopausal women
- Open Access
The effect of time on associations between historical redlining and breast cancer survival
Background: Historical redlining, a 1930s residential segregation policy, has been associated with shorter breast cancer survival. It is unknown whether this disparity has changed over time, given the time since redlining increases and subsequent breast cancer advances have occurred. This study evaluated whether associations between historical redlining and all‐cause survival of breast cancer have changed from 1995 to 2019. Methods: This New York State Cancer Registry–based cohort included 135,827 patients with breast cancer who were diagnosed between 1995 and 2019 and resided in a historically graded census tract at diagnosis. Cases were assigned a historical redlining grade (range, A–D) and divided into 5‐year diagnostic time periods. Cox models estimated associations with survival. Multiplicative interaction between redlining grade and period tested temporal effects. Models were stratified by stage, hormone receptor status, and race/ethnicity. Results: A significant time interaction with historical redlining for survival was found ( p < .0001). Survival disparities for D grade versus A grade were largest in 1995–1999 (hazard ratio [HR], 1.75; 95% CI, 1.55–1.98) and lowest in 2005–2009 (HR, 1.48; 95% CI, 1.32–1.66). Redlining associations for local and regional tumors were largest in 1995–1999 and attenuated over time. Redlining was generally not associated with survival among distant tumors. D‐grade versus A‐grade survival disparities grew over time among patients with hormone receptor–positive tumors. Significant temporal effects were detected among all race/ethnicity groups. Conclusions: Associations between historical redlining and breast cancer survival have changed over time, with disparities generally narrowing. Results stratified by stage and hormone receptor status suggest that such disparities reflect differential access to advances in breast cancer treatment across time.
Open AccessThe effect of time on associations between historical redlining and breast cancer survival
Background: Historical redlining, a 1930s residential segregation policy, has been associated with shorter breast cancer survival. It is unknown whether this disparity has changed over time, given the time since redlining increases and subsequent breast cancer advances have occurred. This study evaluated whether associations between historical redlining and all‐cause survival of breast cancer have changed from 1995 to 2019. Methods: This New York State Cancer Registry–based cohort included 135,827 patients with breast cancer who were diagnosed between 1995 and 2019 and resided in a historically graded census tract at diagnosis. Cases were assigned a historical redlining grade (range, A–D) and divided into 5‐year diagnostic time periods. Cox models estimated associations with survival. Multiplicative interaction between redlining grade and period tested temporal effects. Models were stratified by stage, hormone receptor status, and race/ethnicity. Results: A significant time interaction with historical redlining for survival was found ( p < .0001). Survival disparities for D grade versus A grade were largest in 1995–1999 (hazard ratio [HR], 1.75; 95% CI, 1.55–1.98) and lowest in 2005–2009 (HR, 1.48; 95% CI, 1.32–1.66). Redlining associations for local and regional tumors were largest in 1995–1999 and attenuated over time. Redlining was generally not associated with survival among distant tumors. D‐grade versus A‐grade survival disparities grew over time among patients with hormone receptor–positive tumors. Significant temporal effects were detected among all race/ethnicity groups. Conclusions: Associations between historical redlining and breast cancer survival have changed over time, with disparities generally narrowing. Results stratified by stage and hormone receptor status suggest that such disparities reflect differential access to advances in breast cancer treatment across time.
- Open Access
- 19 December 2025
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 195 (4) , 1015-1026
Historical redlining, breast cancer survival, and the mediating and modifying role of contemporary neighborhood socioeconomic conditions
Historical redlining, a 1930s residential segregation policy, has been associated with contemporary breast cancer survival, but the role of contemporary neighborhood socioeconomic condition is unclear. We investigated mediating and modifying effects of neighborhood socioeconomic condition. This New York State Cancer Registry-based cohort included 60 773 invasive breast cancer cases. Cases were assigned a historical redlining grade (A-D) through linkage to residential census tract at diagnosis. We used labor class index of concentration at the extremes to proxy neighborhood socioeconomic condition. Four-way decomposition evaluated mediation and modification of neighborhood socioeconomic condition on redlining and breast cancer survival. The total effect risk ratio (RR) on 5-year mortality from D-grade vs A-grade = 1.20 (95% CI, 1.09-1.31), which decomposed to a controlled direct effect excess relative risk (ERR) = 0.10 (95% CI, −0.01 to 0.21) and a pure indirect effect ERR = 0.09 (95% CI, 0.05-0.13); significant interaction was not detected. Results were consistent among hormone receptor+, local-stage, and regional-stage tumors, but not hormone receptor− or distant-stage tumors. Decomposition differs by race/ethnicity. Contemporary neighborhood socioeconomic condition mediates approximately half the association between historical redlining and all-cause breast cancer survival, while the other half is attributed to historical redlining. Interventions addressing neighborhood socioeconomic condition may attenuate redlining-based breast cancer survival disparities.
- 19 December 2025
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 195 (4) , 1015-1026
Open AccessHistorical redlining, breast cancer survival, and the mediating and modifying role of contemporary neighborhood socioeconomic conditions
Historical redlining, a 1930s residential segregation policy, has been associated with contemporary breast cancer survival, but the role of contemporary neighborhood socioeconomic condition is unclear. We investigated mediating and modifying effects of neighborhood socioeconomic condition. This New York State Cancer Registry-based cohort included 60 773 invasive breast cancer cases. Cases were assigned a historical redlining grade (A-D) through linkage to residential census tract at diagnosis. We used labor class index of concentration at the extremes to proxy neighborhood socioeconomic condition. Four-way decomposition evaluated mediation and modification of neighborhood socioeconomic condition on redlining and breast cancer survival. The total effect risk ratio (RR) on 5-year mortality from D-grade vs A-grade = 1.20 (95% CI, 1.09-1.31), which decomposed to a controlled direct effect excess relative risk (ERR) = 0.10 (95% CI, −0.01 to 0.21) and a pure indirect effect ERR = 0.09 (95% CI, 0.05-0.13); significant interaction was not detected. Results were consistent among hormone receptor+, local-stage, and regional-stage tumors, but not hormone receptor− or distant-stage tumors. Decomposition differs by race/ethnicity. Contemporary neighborhood socioeconomic condition mediates approximately half the association between historical redlining and all-cause breast cancer survival, while the other half is attributed to historical redlining. Interventions addressing neighborhood socioeconomic condition may attenuate redlining-based breast cancer survival disparities.
- 5 June 2025
- Published by SAGE Publications in Statistical Methods in Medical Research
Assessing accuracy for multi-class classification when subclasses are involved
Classifications that involve subclasses are common in many applied fields. “Compound multi-class classification” refers to the settings which involve three or more main classes and at least one of the main classes has multiple subclasses. In this paper, we propose an accuracy metric proper for “compound M -class classification,” namely “hypervolume under compound R O C manifold ( H U M C , M ) .” The proposed H U M C , M evaluates the overall accuracy of a biomarker measured on continuous scale correctly identifying M main classes without requiring specification of an ordering in terms of marker values for subclasses relative to each other within each main class. The probabilistic interpretation of H U M C , M is analytically derived. A network-based computing algorithm which enables efficient computation of the empirical estimate of H U M C , M is developed. Non-parametric bootstrap percentile confidence intervals of H U M C , M are assessed through extensive simulation studies. Lastly, a real data example is included to illustrate the usage of our proposed method.
- 5 June 2025
- Published by SAGE Publications in Statistical Methods in Medical Research
Assessing accuracy for multi-class classification when subclasses are involved
Classifications that involve subclasses are common in many applied fields. “Compound multi-class classification” refers to the settings which involve three or more main classes and at least one of the main classes has multiple subclasses. In this paper, we propose an accuracy metric proper for “compound M -class classification,” namely “hypervolume under compound R O C manifold ( H U M C , M ) .” The proposed H U M C , M evaluates the overall accuracy of a biomarker measured on continuous scale correctly identifying M main classes without requiring specification of an ordering in terms of marker values for subclasses relative to each other within each main class. The probabilistic interpretation of H U M C , M is analytically derived. A network-based computing algorithm which enables efficient computation of the empirical estimate of H U M C , M is developed. Non-parametric bootstrap percentile confidence intervals of H U M C , M are assessed through extensive simulation studies. Lastly, a real data example is included to illustrate the usage of our proposed method.
- Open Access
- 3 June 2025
- Published by American Association for Cancer Research (AACR)
Supplemental Table S1 from Historical Redlining and All-Cause Survival after Breast Cancer Diagnosis
Supplemental Table S1: 2010 to 2014 and 2015 to 2019 ACS variables used to calculate labor class ICE.
- 3 June 2025
- Published by American Association for Cancer Research (AACR)
Open AccessSupplemental Table S1 from Historical Redlining and All-Cause Survival after Breast Cancer Diagnosis
Supplemental Table S1: 2010 to 2014 and 2015 to 2019 ACS variables used to calculate labor class ICE.
- Open Access
- 3 June 2025
- Published by American Association for Cancer Research (AACR)
Supplemental Figure S2 from Historical Redlining and All-Cause Survival after Breast Cancer Diagnosis
Supplemental Figure S2: Cases by HOLC grade and race/ethnicity. Figure shows the distribution of cases by HOLC grade and race/ethnicity (N = 60,773).
- 3 June 2025
- Published by American Association for Cancer Research (AACR)
Open AccessSupplemental Figure S2 from Historical Redlining and All-Cause Survival after Breast Cancer Diagnosis
Supplemental Figure S2: Cases by HOLC grade and race/ethnicity. Figure shows the distribution of cases by HOLC grade and race/ethnicity (N = 60,773).
- Open Access
- 3 June 2025
- Published by American Association for Cancer Research (AACR)
Supplemental Figure S3 from Historical Redlining and All-Cause Survival after Breast Cancer Diagnosis
Supplemental Figure S3: Kaplan-Meier curve for survival estimates among 2008 to 2018 breast cancer cases by historical redlining grade with 10 years follow-up. Supplemental Figure S3 shows significant differences in survival curves according to historical redlining grade over 10 years of follow-up, with D-grade having the lowest survival probability and A-grade having the highest survival probability.
- 3 June 2025
- Published by American Association for Cancer Research (AACR)
Open AccessSupplemental Figure S3 from Historical Redlining and All-Cause Survival after Breast Cancer Diagnosis
Supplemental Figure S3: Kaplan-Meier curve for survival estimates among 2008 to 2018 breast cancer cases by historical redlining grade with 10 years follow-up. Supplemental Figure S3 shows significant differences in survival curves according to historical redlining grade over 10 years of follow-up, with D-grade having the lowest survival probability and A-grade having the highest survival probability.
- Open Access
- 3 June 2025
- Published by American Association for Cancer Research (AACR)
Supplemental Table S5 from Historical Redlining and All-Cause Survival after Breast Cancer Diagnosis
Supplemental Table S5: Complete-case analysis of historical redlining and 5-year all-cause mortality.
- 3 June 2025
- Published by American Association for Cancer Research (AACR)
Open AccessSupplemental Table S5 from Historical Redlining and All-Cause Survival after Breast Cancer Diagnosis
Supplemental Table S5: Complete-case analysis of historical redlining and 5-year all-cause mortality.