Publications
Dual mode scheduling in volunteer management
Escallon-Barrios, Mariana, Reut Noham, Karen Smilowitz. (2023), Socio-Economic Planning Sciences.
Explainable machine learning analysis of immunology assay patterns among hospitalized patients undergoing immunology testing
Liat Shabtay, David Hagin, Daphna Paran, Yifat Alcalay, Reut Noham. (2026), Scientific Reports.
Overuse of immunology laboratory testing remains a challenge in acute care, increasing workload, downstream investigations, and diagnostic complexity. We applied explainable machine learning to evaluate real-world immunology test utilization among hospitalized patients who underwent immunology testing. We analyzed electronic health records from a tertiary-care hospital between 2018 and 2023, including 7,402 adults and 75 model predictors. Discharge diagnoses were clinically reviewed and mapped to a retrospective immune-related versus non-immune outcome. Multiple machine-learning models were evaluated, with XGBoost selected for detailed interpretation using SHapley Additive exPlanations (SHAP). Immunology assays were frequently ordered in both groups, with only modest differences in ordering frequency. XGBoost achieved an ROC-AUC of 0.749 and weighted F1 of 0.75. SHAP analyses showed that routine clinical and laboratory variables generally contributed more strongly to discrimination than most immunology assays, while complement-related markers showed higher contribution and assay effects varied across clinical contexts. Subgroup analyses demonstrated context-dependent variability, and outcome-independent sensitivity analyses showed largely stable performance when information from individual assays was randomly masked. These findings show how explainable machine learning can support evaluation of diagnostic test-utilization patterns and provide a practical framework for identifying potential misalignment between ordering frequency and model-derived predictive contribution within immunology-tested populations.
Dynamic Scheduling of Recurring Multisession Appointments with Heterogeneous Clients
Reut Noham, Karen Smilowitz. (2026), Service Science.
Clients seeking paramedical and rehabilitation services require recurring treatment sessions over an extended period. Unlike single-visit problems, these services must assign each accepted client to a fixed, recurring day–time slot that remains occupied throughout the treatment program. This structure creates long-termcapacity commitments that limit future scheduling flexibility and complicate acceptance decisions, particularly when clients differ in availability and required program durations. Motivated by an early intervention programfor infants and toddlers with developmental delays, we study scheduling policies designed to address this combination of heterogeneity, uncertainty, and recurrence constraints. We model the multisession appointment scheduling problem as a Markov decision process in which requests arrive sequentially and decisionsmust consider both immediate feasibility and the long-term implications of blocking a slot across many periods. Our analysis identifies key structural elements of the scheduling decision, including a slot-selection guideline that assigns accepted clients to the least popular feasible slot and a duration-based threshold that characterizes acceptance behavior. These insights highlight the value of preserving flexibility and anticipating demand when scheduling recurring appointments under uncertainty. Building on these results, we develop a heuristic that groups schedule states into occupancy categories and applies simplified acceptance thresholds. Computational experiments show that this anticipatory approach outperforms first come, first served benchmarks, particularly when slot popularity is uneven or program durations vary widely. Whereas we do not model health outcomes directly, prior research links improved access, timely initiation, and continuity of care with better therapeutic results, underscoring the broader potential impact of more efficient scheduling.
Using a Large Language Model to Support Thematic Analysis of Patient Experiences in Chronic Illness Management: Comparative Qualitative Study
Sara Kivity, Yechiel Michael Barilan, Reut Noham, Mor Saban. (2026), Journal of Medical Internet Research.
Background: Qualitative health research often focuses on how patients experience and manage chronic illnesses, a topic that has been extensively studied in the literature. With the emergence of large language models (LLMs), such as Claude (Anthropic PBC) and ChatGPT (OpenAI), new opportunities are arising to support and scale the thematic analysis of narrative health data. However, their role and added value compared to traditional human-led approaches remain underexplored, particularly in complex clinical contexts such as multimorbidity. Objective: We aim to evaluate the methodological contribution of LLM-assisted analysis by examining its ability to replicate and extend established qualitative insights, in comparison with traditional thematic analysis. Methods: Semistructured interviews were conducted with 30 individuals living with two or more chronic illnesses. Transcripts were analyzed using both manual thematic coding and Claude 3.5 Sonnet. A structured comparison was conducted to identify shared and unique themes across the two approaches. The analysis examined thematic overlap, differences in subtheme identification, and variation in the level of detail between the methods. Results: Both approaches identified similar core themes related to the patient experience, including health care navigation and challenges, support systems and family dynamics, and emotional challenges and coping. Manual analysis produced more contextually detailed interpretations, while the LLM approach identified a larger number of subthemes. Each method also revealed distinct themes: the manual analysis included themes such as faith, caregiving roles, and a proactive mindset, whereas the LLM identified themes such as future planning and multiple health conditions. The findings show both similarities and differences between the two approaches. The LLM analysis also demonstrated efficiency in processing large volumes of qualitative data. Conclusions: A hybrid approach that integrates artificial intelligence–assisted and human-led thematic analysis can enhance both analytical depth and scalability. These findings support the use of LLMs as a complementary tool in qualitative research, while highlighting the importance of combining automated pattern detection with human interpretation.
Balancing effectiveness and equity in the face of uncertainty: The case of humanitarian lateral transshipments
Reut Noham, Ohad Eisenhandler. (2026), Socio-Economic Planning Sciences.
In humanitarian operations, decision makers must allocate limited resources efficiently while ensuring fair outcomes for disaster-affected communities. This paper addresses the Humanitarian Transshipment Problem (HTP), which involves the joint optimization of pre-positioned inventory and post-disaster lateral transshipments. The central challenge lies in balancing effectiveness, namely, the extent to which needs are met, and equity, namely, the fairness of aid distribution. This trade-off becomes even more complex under uncertainty, when multiple disaster scenarios must be anticipated. Existing approaches to this balance in stochastic settings often rely on ad-hoc formulations, offering little justification for how social welfare should be measured across scenarios. We propose an alternative framework that establishes an axiomatic foundation for evaluating the effectiveness–equity trade-off under uncertainty, filling a significant gap in the literature. Our formulation preserves desirable properties across scenarios while remaining tractable and interpretable, making it suitable for humanitarian decision-making. Using both real-world and synthetic data, we demonstrate that our model enhances system performance and supports equitable decision-making under uncertainty. Our results highlight the strategic value of lateral transshipments, especially under tight budgets and high uncertainty, and provide guidance for organizations seeking to improve fairness, effectiveness, and efficiency in disaster preparedness.
Dual mode scheduling in volunteer management
Escallon-Barrios, Mariana, Reut Noham, Karen Smilowitz. (2024), Socio-Economic Planning Sciences.
Nonprofit organizations have adopted online scheduling platforms that give autonomy to volunteers in the scheduling process. However, this strategy can create imbalances in task coverage, often requiring staff to fill the gaps. The aim of this study is to develop scheduling strategies to create a balanced schedule that effectively combines workforce types (paid staff and volunteers) while keeping volunteers engaged. This is achieved by accounting for volunteers’ responses to changes in scheduling options. We develop an optimization model that recognizes volunteers’ scheduling responses and utilizes these responses to design policies aimed at achieving a balanced coverage across time slots. This involves reducing over-covered and under-covered time slots over the planning horizon. By understanding the preferences of volunteers, organizations can modify their current policies to better match supply with demand keeping their volunteers engaged. We provide an implementable scheduling strategy combining staff assignment and volunteers’ autonomy in scheduling choices. Case study results show an improvement compared to current scheduling policies. Volunteers’ satisfaction increases, resulting in a long-term impact on the organizations and the communities they serve.
An indirect prioritization approach to optimizing sample referral networks for HIV early infant diagnosis
Reut Noham, Michal Tzur, Dan Yamin (2022), IISE Transactions.
Early diagnosis and treatment of newborns with Human Immunodeficiency Virus (HIV) can substantially reduce mortality rates. Polymerase chain reduction technology is desirable for diagnosing HIV-exposed infants and for monitoring the disease progression in older patients. In low- and middle-income countries (LMICs), processing both types of tests requires the use of scarce resources. In this article, we present a supply chain network model for referring/assigning HIV test samples from clinics to labs. These assignments aim to minimize the expected infant mortality from AIDS due to delays in the return of test results. Using queuing theory, we present an analytical framework to evaluate the distribution of the sample waiting times at the testing labs and incorporate it into a mathematical model. The suggested framework takes into consideration the non-stationarity in the availability of reagents and technical staff. Hence, our model provides a method to find an assignment strategy that involves an indirect prioritization of samples that are more likely than others to be positive. We also develop a heuristic to simplify the implementation of an assignment strategy and provide general managerial insights for operating sample referral networks in LMICs with limited resources. Using a case study from Tanzania, we show that the potential improvement is substantial, especially when some labs are utilized almost to their full capacity. Our results apply to other settings in which expensive equipment with volatile availability is used to perform crucial operations, for example, the recent COVID-19 pandemic.
Design and incentive decisions to increase cooperation in humanitarian relief networks
Reut Noham, Michal Tzur (2020), IISE Transactions.
During humanitarian relief operations, designated facilities are established to assist the affected population and distribute relief goods. In settings where the authorities manage the operations, they instruct the population regarding which facility they should visit. However, in times of crises and uncertainty, these instructions are often not followed. In this work, we investigate how the authorities should invest in incentivizing the population to follow their instructions. These decisions need to be combined with those concerning the relief network design. The population’s behavior and level of cooperation are key factors in deciding on the incentive investments. We present a new mathematical model that incorporates decisions regarding which populations to incentivize to follow the local authorities’ instructions. Then, we develop properties that can help the authorities decide on the level of investment in incentives. A numerical study demonstrates that incentives can improve the system’s performance and enable an equitable supply allocation. Furthermore, an investment in a small number of communities is typically sufficient to significantly improve the system’s performance. We also demonstrate that incentives affect relief-network design decisions.
Designing humanitarian supply chains by incorporating actual post-disaster decisions
Reut Noham, Michal Tzur (2018), European Journal of Operational Research.
Existing models for disaster preparedness and response address network design and resource allocation challenges. However, these models typically adopt a global optimization point of view, which may not be attainable since they do not consider the actual decision-making process after a disaster occurs. This process is based mostly on practitioners' knowledge and experience, rules of thumb and the population behavior. In this paper, we develop a new mathematical model that incorporates such practical considerations. The model includes actual post-disaster decisions through a set of “humanitarian constraints”. We then present an efficient optimal solution method to solve small/medium-size instances of the resulting problem and a heuristic algorithm based on the Tabu-search method for large instances of the problem. We test our methods on problems with randomly generated data, as well as real data obtained from the Geophysical Institute of Israel. The results demonstrate that our heuristic performs exceptionally well, and optimal solutions are obtained in almost all cases. More importantly, we show that ignoring the actual decision-making process that occurs at the post-disaster stage results in inferior actual overall solutions. Using the humanitarian constraints improves the entire supply chain performance. Therefore, it is critical to accurately incorporate post-disaster decisions during the pre-disaster planning phase.
The single and multi‐item transshipment problem with fixed transshipment costs
Reut Noham, Michal Tzur (2014), Naval Research Logistics.
This article deals with supply chain systems in which lateral transshipments are allowed. For a system with two retailers facing stochastic demand, we relax the assumption of negligible fixed transshipment costs, thus, extending existing results for the single-item case and introducing a new model with multiple items. The goal is to determine optimal transshipment and replenishment policies, such that the total centralized expected profit of both retailers is maximized. For the single-item problem with fixed transshipment costs, we develop optimality conditions, analyze the expected profit function, and identify the optimal solution. We extend our analysis to multiple items with joint fixed transshipment costs, a problem that has not been investigated previously in the literature, and show how the optimality conditions may be extended for any number of items. Due to the complexity involved in solving these conditions, we suggest a simple heuristic based on the single-item results. Finally, we conduct a numerical study that provides managerial insights on the solutions obtained in various settings and demonstrates that the suggested heuristic performs very well.