Andres joined the MGH Institute for Technology Assessment (ITA) in August 2021 as a postdoctoral fellow. He received his PhD in operations management from INSEAD, France, and his BS in applied mathematics and physics from the New Jersey Institute of Technology.
Andres’s research uses stochastic simulation and optimization to support decision-making in healthcare settings. In the past, he has worked on clinical trial design, ICU patient-flow, and resource allocation of mobile healthcare units. At the ITA, he is working with Dr. Carrie Cunningham on the optimization of thyroid nodule treatment.
Selected Publications
2021
Forster, Martin; Brealey, Stephen; Chick, Stephen; Keding, Ada; Corbacho, Belen; Alban, Andres; Pertile, Paolo; Rangan, Amar
Cost-effective clinical trial design: Application of a Bayesian sequential model to the ProFHER pragmatic trial. Journal Article
In: Clinical trials (London, England), pp. 17407745211032909, 2021, ISSN: 1740-7753, ().
@article{Forster2021,
title = {Cost-effective clinical trial design: Application of a Bayesian sequential model to the ProFHER pragmatic trial.},
author = {Martin Forster and Stephen Brealey and Stephen Chick and Ada Keding and Belen Corbacho and Andres Alban and Paolo Pertile and Amar Rangan},
url = {https://pubmed.ncbi.nlm.nih.gov/34407641/},
doi = {10.1177/17407745211032909},
issn = {1740-7753},
year = {2021},
date = {2021-08-01},
journal = {Clinical trials (London, England)},
pages = {17407745211032909},
abstract = {There is growing interest in the use of adaptive designs to improve the efficiency of clinical trials. We apply a Bayesian decision-theoretic model of a sequential experiment using cost and outcome data from the ProFHER pragmatic trial. We assess the model's potential for delivering value-based research. Using parameter values estimated from the ProFHER pragmatic trial, including the costs of carrying out the trial, we establish when the trial could have stopped, had the model's value-based stopping rule been used. We use a bootstrap analysis and simulation study to assess a range of operating characteristics, which we compare with a fixed sample size design which does not allow for early stopping. We estimate that application of the model could have stopped the ProFHER trial early, reducing the sample size by about 14%, saving about 5% of the research budget and resulting in a technology recommendation which was the same as that of the trial. The bootstrap analysis suggests that the expected sample size would have been 38% lower, saving around 13% of the research budget, with a probability of 0.92 of making the same technology recommendation decision. It also shows a large degree of variability in the trial's sample size. Benefits to trial cost stewardship may be achieved by monitoring trial data as they accumulate and using a stopping rule which balances the benefit of obtaining more information through continued recruitment with the cost of obtaining that information. We present recommendations for further research investigating the application of value-based sequential designs.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
There is growing interest in the use of adaptive designs to improve the efficiency of clinical trials. We apply a Bayesian decision-theoretic model of a sequential experiment using cost and outcome data from the ProFHER pragmatic trial. We assess the model's potential for delivering value-based research. Using parameter values estimated from the ProFHER pragmatic trial, including the costs of carrying out the trial, we establish when the trial could have stopped, had the model's value-based stopping rule been used. We use a bootstrap analysis and simulation study to assess a range of operating characteristics, which we compare with a fixed sample size design which does not allow for early stopping. We estimate that application of the model could have stopped the ProFHER trial early, reducing the sample size by about 14%, saving about 5% of the research budget and resulting in a technology recommendation which was the same as that of the trial. The bootstrap analysis suggests that the expected sample size would have been 38% lower, saving around 13% of the research budget, with a probability of 0.92 of making the same technology recommendation decision. It also shows a large degree of variability in the trial's sample size. Benefits to trial cost stewardship may be achieved by monitoring trial data as they accumulate and using a stopping rule which balances the benefit of obtaining more information through continued recruitment with the cost of obtaining that information. We present recommendations for further research investigating the application of value-based sequential designs.
Blaettchen, Philippe; Vries, Harwin; Wassenhove, Luk N. Van; Alban, Andres
Resource Allocation with Sigmoidal Demands: Mobile Healthcare Units and Service Adoption Journal Article
In: Manufacturing & Service Operations Management, 2021, ().
@article{Alban2021,
title = {Resource Allocation with Sigmoidal Demands: Mobile Healthcare Units and Service Adoption},
author = {Philippe Blaettchen and Harwin Vries and Luk N. Van Wassenhove and Andres Alban},
url = {https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3906146},
year = {2021},
date = {2021-01-01},
journal = {Manufacturing \& Service Operations Management},
abstract = {Achieving broad access to health services (a target within the sustainable development goals) requires reaching rural populations. Mobile healthcare units (MHUs) visit remote sites to offer health services to these populations. However, limited exposure, health literacy, and trust can lead to sigmoidal (S-shaped) adoption dynamics, presenting a difficult obstacle in allocating limited MHU resources. It is tempting to allocate resources in line with current demand, as seen in practice. However, to maximize access in the long term, this may be far from optimal, and insights into allocation decisions are limited.
We present a formal model of the allocation of MHU resources, i.e., the frequency of visits to each site, to maximize long-term uptake of preventative health services. We formulate the problem as the optimization of a sum of sigmoidal functions. While the problem is NP-hard, we provide closed-form solutions to particular cases of the model that elucidate insights into the optimal allocation. For example, more visits should generally be allocated to sites where the cumulative demand potential is higher and, counterintuitively, often those where demand is currently lower. To apply our insights in practice, we propose a practical method for estimating our model's parameters from pre-existing data. Our estimation approach achieves better predictions than standard methods. Finally, we demonstrate the potential of our approach by applying our methods to family planning MHUs in Uganda. In particular, we show that operationalizable heuristic allocations, grounded in our insights, outperform allocations based on current demand.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Achieving broad access to health services (a target within the sustainable development goals) requires reaching rural populations. Mobile healthcare units (MHUs) visit remote sites to offer health services to these populations. However, limited exposure, health literacy, and trust can lead to sigmoidal (S-shaped) adoption dynamics, presenting a difficult obstacle in allocating limited MHU resources. It is tempting to allocate resources in line with current demand, as seen in practice. However, to maximize access in the long term, this may be far from optimal, and insights into allocation decisions are limited.
We present a formal model of the allocation of MHU resources, i.e., the frequency of visits to each site, to maximize long-term uptake of preventative health services. We formulate the problem as the optimization of a sum of sigmoidal functions. While the problem is NP-hard, we provide closed-form solutions to particular cases of the model that elucidate insights into the optimal allocation. For example, more visits should generally be allocated to sites where the cumulative demand potential is higher and, counterintuitively, often those where demand is currently lower. To apply our insights in practice, we propose a practical method for estimating our model's parameters from pre-existing data. Our estimation approach achieves better predictions than standard methods. Finally, we demonstrate the potential of our approach by applying our methods to family planning MHUs in Uganda. In particular, we show that operationalizable heuristic allocations, grounded in our insights, outperform allocations based on current demand.
We present a formal model of the allocation of MHU resources, i.e., the frequency of visits to each site, to maximize long-term uptake of preventative health services. We formulate the problem as the optimization of a sum of sigmoidal functions. While the problem is NP-hard, we provide closed-form solutions to particular cases of the model that elucidate insights into the optimal allocation. For example, more visits should generally be allocated to sites where the cumulative demand potential is higher and, counterintuitively, often those where demand is currently lower. To apply our insights in practice, we propose a practical method for estimating our model's parameters from pre-existing data. Our estimation approach achieves better predictions than standard methods. Finally, we demonstrate the potential of our approach by applying our methods to family planning MHUs in Uganda. In particular, we show that operationalizable heuristic allocations, grounded in our insights, outperform allocations based on current demand.
2020
Chick, Stephen E.; Dongelmans, Dave A.; Vlaar, Alexander P. J.; Sent, Danielle; Group, Study; Alban, Andres
ICU capacity management during the COVID-19 pandemic using a process simulation. Journal Article
In: Intensive care medicine, vol. 46, pp. 1624–1626, 2020, ISSN: 1432-1238, ().
@article{Alban2020,
title = {ICU capacity management during the COVID-19 pandemic using a process simulation.},
author = {Stephen E. Chick and Dave A. Dongelmans and Alexander P. J. Vlaar and Danielle Sent and Study Group and Andres Alban},
url = {https://pubmed.ncbi.nlm.nih.gov/32383060/},
doi = {10.1007/s00134-020-06066-7},
issn = {1432-1238},
year = {2020},
date = {2020-08-01},
journal = {Intensive care medicine},
volume = {46},
pages = {1624--1626},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2017
Darji, Hardik A.; Imamura, Atsuki; Nakayama, Marvin K.; Alban, Andres
Efficient Monte Carlo methods for estimating failure probabilities Journal Article
In: Reliability Engineering & System Safety, vol. 165, pp. 376-394, 2017, ISSN: 0951-8320, ().
@article{Alban2017,
title = {Efficient Monte Carlo methods for estimating failure probabilities},
author = {Hardik A. Darji and Atsuki Imamura and Marvin K. Nakayama and Andres Alban},
url = {https://www.sciencedirect.com/science/article/pii/S0951832017304325},
doi = {https://doi.org/10.1016/j.ress.2017.04.001},
issn = {0951-8320},
year = {2017},
date = {2017-01-01},
urldate = {2017-01-01},
journal = {Reliability Engineering \& System Safety},
volume = {165},
pages = {376-394},
abstract = {We develop efficient Monte Carlo methods for estimating the failure probability of a system. An example of the problem comes from an approach for probabilistic safety assessment of nuclear power plants known as risk-informed safety-margin characterization, but it also arises in other contexts, e.g., structural reliability, catastrophe modeling, and finance. We estimate the failure probability using different combinations of simulation methodologies, including stratified sampling (SS), (replicated) Latin hypercube sampling (LHS), and conditional Monte Carlo (CMC). We prove theorems establishing that the combination SS+LHS (resp., SS+CMC+LHS) has smaller asymptotic variance than SS (resp., SS+LHS). We also devise asymptotically valid (as the overall sample size grows large) upper confidence bounds for the failure probability for the methods considered. The confidence bounds may be employed to perform an asymptotically valid probabilistic safety assessment. We present numerical results demonstrating that the combination SS+CMC+LHS can result in substantial variance reductions compared to stratified sampling alone.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
We develop efficient Monte Carlo methods for estimating the failure probability of a system. An example of the problem comes from an approach for probabilistic safety assessment of nuclear power plants known as risk-informed safety-margin characterization, but it also arises in other contexts, e.g., structural reliability, catastrophe modeling, and finance. We estimate the failure probability using different combinations of simulation methodologies, including stratified sampling (SS), (replicated) Latin hypercube sampling (LHS), and conditional Monte Carlo (CMC). We prove theorems establishing that the combination SS+LHS (resp., SS+CMC+LHS) has smaller asymptotic variance than SS (resp., SS+LHS). We also devise asymptotically valid (as the overall sample size grows large) upper confidence bounds for the failure probability for the methods considered. The confidence bounds may be employed to perform an asymptotically valid probabilistic safety assessment. We present numerical results demonstrating that the combination SS+CMC+LHS can result in substantial variance reductions compared to stratified sampling alone.