About

Martina Morris is an Emeritus faculty member at the University of Washington, affiliated with the Department of Sociology and the Department of Statistics. Her research focuses on methodology for the social sciences, including distributional comparison, inequality, social networks, and epidemiology of HIV. She has contributed to the development of social science research methods and has a particular interest in understanding social structures and health-related issues through quantitative analysis.

Research topics

• Machine Learning
• Computer Science
• Data Mining
• Artificial Intelligence
• Programming language
• Data science
• Mathematics
• Statistics
• Theoretical computer science

Selected publications

• High Resolution Computed Tomography Ventilation-Perfusion Analysis in Healthy, Active Individuals With Post-COVID-19 Dyspnea

  American Journal of Respiratory and Critical Care Medicine · 2025-05-01

  articleSenior author

  Abstract Rationale: High-resolution Computed Tomography (HRCT) is a valuable imaging technique in the assessment of diffuse lung diseases, offering insight into dynamic changes in lung attenuation. Combining HRCT with pulmonary function tests (PFTs) provides a comprehensive approach to understanding lung pathophysiology. This study aimed to assess lung function in healthy, active individuals who have recovered after COVID-19. We hypothesize that there exists a significant association between specific HRCT ventilation-perfusion analysis parameters and traditional PFT measurements in a cohort of healthy, active individuals who have recovered from COVID-19. Methods: A cross-sectional study was conducted on 37 healthy, active individuals who had tested positive for COVID-19 and had since recovered. Each participant underwent HRCT and PFTs in a single session. HRCT images were processed using CT lung ventilation analysis software (4DMedical, Australia) to generate functional ventilation metrics: Change in volume between inspiration and expiration (Delta Vol), Ventilation Defect Percent (VDP), Mean Specific Ventilation (MSV), and Ventilation Heterogeneity (VH). A cohort of 10 participants had CT lung perfusion analysis performed with measurement of perfusion metrics: Perfusion Defect Percent (QDP), and Perfusion Heterogeneity (QH). PFT measures included FEV1, FVC, TLC, RV, and DLCO. Statistical analyses were performed to evaluate associations between HRCT ventilation parameters and PFT measurements while controlling for potential confounding factors. Results: The mean FEV1 % predicted was 92% and FVC % predicted of 96%. Strong correlations were found between TLC and HRCT-derived inspiratory volume (r=0.8, p<0.001) and RV and HRCT-derived expiratory volume (r=0.69, p=0.002). There was a significant correlation between FVC and delta vol (r=0.85, p<0.001). Average perfusion heterogeneity was 30.9 ± 9.5 and perfusion defect percentage was 3.7% ± 2.7%. Among patients who were hospitalized both PFTs and HRCT including ventilation analysis detected presence of air trapping. There were no significant abnormalities identified on perfusion analysis even amongst those admitted for COVID-19 pneumonia. Conclusions: In healthy, active individuals who have recovered from COVID-19, we reported significant correlations between HRCT-derived functional metrics and those obtained via standard tests (PFTs). No perfusion defects were identified to suggest macro or microvascular circulatory dysfunction. The use of HRCT ventilation-perfusion analysis and PFTs enhances our understanding of post-COVID-19 lung pathology and provides a comprehensive approach for assessing lung health.

  PublisherDOI

• The value of implementing a nursing and midwifery clinical accreditation programme at two NHS trusts

  British Journal of Nursing · 2024-02-22 · 1 citations

  article1st authorCorresponding

  Unit/ward accreditation programmes have been widely implemented by nursing and midwifery teams across healthcare providers in the UK over the recent years and have many associated benefits. These include promoting quality improvement on a wider scale across the organisation, strengthening oversight and accountability of quality and safety from ward to board and vice versa, promoting shared learning, and providing opportunities for sharing and celebrating excellence. The Royal Wolverhampton NHS Trust and Walsall Healthcare NHS Trust have recognised the value of this approach, launching a clinical accreditation programme in April 2023. This initially focused on nursing and midwifery, with plans to widen the approach to other disciplines and specialist teams. Up to the time of writing, 56 visits had been undertaken with 30 clinical areas accredited. The remaining visited areas are awaiting their accreditation outcome. The approach has positively contributed to improvements in patient outcomes, such as more patient observations being completed on time, a reduction in patient falls and improvements in pressure ulcers. Colleagues participating in the programme and frontline staff working in the clinical areas assessed have reported how positive the approach has been, providing opportunities for shared learning and celebrating excellence.

  PublisherDOI

• <b>ergm</b> 4: New Features for Analyzing Exponential-Family Random Graph Models

  Journal of Statistical Software · 2023 · 61 citations

  • Computer Science
  • Computer Science
  • Data Mining

  The ergm package supports the statistical analysis and simulation of network data. It anchors the statnet suite of packages for network analysis in R introduced in a special issue in Journal of Statistical Software in 2008. This article provides an overview of the new functionality in the 2021 release of ergm version 4. These include more flexible handling of nodal covariates, term operators that extend and simplify model specification, new models for networks with valued edges, improved handling of constraints on the sample space of networks, and estimation with missing edge data. We also identify the new packages in the statnet suite that extend ergm's functionality to other network data types and structural features and the robust set of online resources that support the statnet development process and applications.

  PublisherOA PDFDOI

• Partnership types and coital frequency as predictors of gonorrhea and chlamydia among young MSM and young transgender women

  International Journal of STD & AIDS · 2023-05-05 · 7 citations

  articleOpen access

  BACKGROUND: Sexually transmitted infections pose a major public health challenge in the United States and this burden is especially acute in subpopulations like young men who have sex with men (YMSM) and young transgender women (YTW). Yet, the direct behavioral antecedents of these infections are not well understood making it difficult to identify the cause of recent increases in incidence. This study examines how variations in partnership rates and the number of condomless sex acts are associated with STI infections among YMSM-YTW. METHOD: This study leveraged 3 years of data from a large longitudinal cohort of YMSM-YTW. A series of generalized linear mixed models examined the association between the number of condomless anal sex acts, number of one-time partners, number of casual partners, and number of main partners and chlamydia, gonorrhea, or any STI. RESULTS: Results indicated the number of casual partners was associated with gonorrhea [aOR = 1.17 (95% CI: 1.08, 1.26)], chlamydia [aOR = 1.12 (95% CI: 1.05, 1.20)], and any STI [aOR = 1.14 (95% CI: 1.08, 1.21)] while the number of one-time partners was only associated with gonorrhea [aOR = 1.13 (95% CI: 1.02, 1.26)]. The number of condomless anal sex acts was not associated with any outcome. CONCLUSION: These findings suggest the number of casual partners is a consistent predictor of STI infection among YMSM-YTW. This may reflect the quick saturation of risk within partnerships making the number of partners, rather than the number of acts, the more relevant factor for STI risk.

  PublisherOA PDFDOI

• Improving and Extending STERGM Approximations Based on Cross-Sectional Data and Tie Durations

  Figshare · 2023-01-01

  datasetOpen access

  Temporal exponential-family random graph models (TERGMs) are a flexible class of models for network ties that change over time. Separable TERGMs (STERGMs) are a subclass of TERGMs in which the dynamics of tie formation and dissolution can be separated within each discrete time step and may depend on different factors. The Carnegie et al. (2015) approximation improves estimation efficiency for a subclass of STERGMs, allowing them to be reliably estimated from inexpensive cross-sectional study designs. This approximation adapts to cross-sectional data by attempting to construct a STERGM with two specific properties: a cross-sectional equilibrium distribution defined by an exponential-family random graph model (ERGM) for the network structure, and geometric tie duration distributions defined by constant hazards for tie dissolution. In this paper we focus on approaches for improving the behavior of the Carnegie et al. approximation and increasing its scope of application. We begin with Carnegie et al.’s observation that the exact result is tractable when the ERGM is dyad-independent, and then show that taking the sparse limit of the exact result leads to a different approximation than the one they presented. We show that the new approximation outperforms theirs for sparse, dyad-independent models, and observe that the errors tend to increase with the strength of dependence for dyad-dependent models. We then develop theoretical results in the dyad-dependent case, showing that when the ERGM is allowed to have arbitrary dyad-dependent terms and some dyad-dependent constraints, both the old and new approximations are asymptotically exact as the size of the STERGM time step goes to zero. We note that the continuous-time limit of the discrete-time approximations has the desired cross-sectional equilibrium distribution and exponential tie duration distributions with the desired means. We show that our results extend to hypergraphs, and we propose an extension to dissolution hazards that depend on tie age.

  PublisherDOI

• Improving and Extending STERGM Approximations Based on Cross-Sectional Data and Tie Durations

  Figshare · 2023-01-01

  datasetOpen access

  Temporal exponential-family random graph models (TERGMs) are a flexible class of models for network ties that change over time. Separable TERGMs (STERGMs) are a subclass of TERGMs in which the dynamics of tie formation and dissolution can be separated within each discrete time step and may depend on different factors. The Carnegie et al. approximation improves estimation efficiency for a subclass of STERGMs, allowing them to be reliably estimated from inexpensive cross-sectional study designs. This approximation adapts to cross-sectional data by attempting to construct a STERGM with two specific properties: a cross-sectional equilibrium distribution defined by an exponential-family random graph model (ERGM) for the network structure, and geometric tie duration distributions defined by constant hazards for tie dissolution. In this article we focus on approaches for improving the behavior of the Carnegie et al. approximation and increasing its scope of application. We begin with Carnegie et al.’s observation that the exact result is tractable when the ERGM is dyad-independent, and then show that taking the sparse limit of the exact result leads to a different approximation than the one they presented. We show that the new approximation outperforms theirs for sparse, dyad-independent models, and observe that the errors tend to increase with the strength of dependence for dyad-dependent models. We then develop theoretical results in the dyad-dependent case, showing that when the ERGM is allowed to have arbitrary dyad-dependent terms and some dyad-dependent constraints, both the old and new approximations are asymptotically exact as the size of the STERGM time step goes to zero. We note that the continuous-time limit of the discrete-time approximations has the desired cross-sectional equilibrium distribution and exponential tie duration distributions with the desired means. We show that our results extend to hypergraphs, and we propose an extension of the Carnegie et al. framework to dissolution hazards that depend on tie age. Supplementary materials for this article are available online.

  PublisherDOI

• Improving and Extending STERGM Approximations Based on Cross-Sectional Data and Tie Durations

  Journal of Computational and Graphical Statistics · 2023-07-07

  articleOpen access

  Temporal exponential-family random graph models (TERGMs) are a flexible class of models for network ties that change over time. Separable TERGMs (STERGMs) are a subclass of TERGMs in which the dynamics of tie formation and dissolution can be separated within each discrete time step and may depend on different factors. The Carnegie et al. (2015) approximation improves estimation efficiency for a subclass of STERGMs, allowing them to be reliably estimated from inexpensive cross-sectional study designs. This approximation adapts to cross-sectional data by attempting to construct a STERGM with two specific properties: a cross-sectional equilibrium distribution defined by an exponential-family random graph model (ERGM) for the network structure, and geometric tie duration distributions defined by constant hazards for tie dissolution. In this paper we focus on approaches for improving the behavior of the Carnegie et al. approximation and increasing its scope of application. We begin with Carnegie et al.'s observation that the exact result is tractable when the ERGM is dyad-independent, and then show that taking the sparse limit of the exact result leads to a different approximation than the one they presented. We show that the new approximation outperforms theirs for sparse, dyad-independent models, and observe that the errors tend to increase with the strength of dependence for dyad-dependent models. We then develop theoretical results in the dyad-dependent case, showing that when the ERGM is allowed to have arbitrary dyad-dependent terms and some dyad-dependent constraints, both the old and new approximations are asymptotically exact as the size of the STERGM time step goes to zero. We note that the continuous-time limit of the discrete-time approximations has the desired cross-sectional equilibrium distribution and exponential tie duration distributions with the desired means. We show that our results extend to hypergraphs, and we propose an extension of the Carnegie et al. framework to dissolution hazards that depend on tie age.

  PublisherOA PDFDOI

• Introducing the Back to the Floor concept at the Royal Wolverhampton and Walsall Healthcare NHS Trusts

  British Journal of Nursing · 2023-08-17

  article1st authorCorresponding
  PublisherDOI

• networkLite: An Simplified Implementation of the 'network' Package Functionality

  2023-01-24

  datasetOpen access

  An implementation of some of the core 'network' package functionality based on a simplified data structure that is faster in many research applications. This package is designed for back-end use in the 'statnet' family of packages, including 'EpiModel'. Support is provided for binary and weighted, directed and undirected, bipartite and unipartite networks; no current support for multigraphs, hypergraphs, or loops.

  PublisherOA PDFDOI

• Comparing Sexual Network Mean Active Degree Measurement Metrics among Men who have Sex with Men

  medRxiv · 2022-02-13 · 2 citations

  preprintOpen access

  ABSTRACT Background Mean active degree is an important proxy measure of cross-sectional network connectivity commonly used in HIV/STI epidemiology research. No current studies have compared measurement methods of mean degree using cross-sectional surveys for men who have sex with men (MSM) in the United States. Methods We compared mean degree estimates based on reported ongoing main and casual sexual partnerships ( current method ) against dates of first and last sex ( retrospective method ) from 0–12 months prior to survey date in ARTnet, a cross-sectional survey of MSM in the U.S. (2017–2019). ARTnet collected data on the number of sexual partners in the past year but limited reporting on details used for calculating mean degree to the 5 most recent partners. We used linear regression to understand the impact of truncated partnership data on mean degree estimation. Results Retrospective method mean degree systematically decreased as the month at which it was calculated increased from 0–12 months prior to survey date. Among participants with &gt;5 partners in the past year compared to those with ≤5, the average change in main degree between 12 and 0 months prior to survey date was −0.05 (95% CI: −0.08, −0.03) after adjusting for race/ethnicity, age, and education. The adjusted average change in casual degree was −0.40 (95% CI: −0.45, −0.35). Conclusions The retrospective method underestimates mean degree for MSM in surveys with truncated partnership data, especially for casual partnerships. The current method is less prone to bias from partner truncation when the target population experiences higher cumulative partners per year. Summary Survey designs can lead to potential bias, such as underestimation, in the measurement of mean active degree in sexual networks of men who have sex with men.

  PublisherOA PDFDOI

Recent grants

• Collaborative Research: Modeling Interaction Between Individual Behavior, Social Networks And Public Policy To Support Public Health Epidemiology.

  NSF · $120k · 2007–2010

Frequent coauthors

• Mark S. Handcock

  Development Fund

  42 shared

• Steven M. Goodreau

  University of Washington

  32 shared

• Samuel M. Jenness

  Emory University

  27 shared

• James Moody

  Duke University

  19 shared

• Pavel N. Krivitsky

  UNSW Sydney

  19 shared

• Maria J. Wawer

  Johns Hopkins University

  18 shared

• David R. Hunter

  17 shared

• Annette Bernhardt

  13 shared