Machine learning models for assessing risk factors affecting health care costs: 12-month exercise-based cardiac rehabilitation.

Introduction: Exercise-based cardiac rehabilitation (ECR) has proven to be effective and cost-effective dominant treatment option in health care. However, the contribution of well-known risk factors for prognosis of coronary artery disease (CAD) to predict health care costs is not well recognized. Since machine learning (ML) applications are rapidly giving new opportunities to assist health care professionals' work, we used selected ML tools to assess the predictive value of defined risk factors for health care costs during 12-month ECR in patients with CAD. Methods: The data for analysis was available from a total of 71 patients referred to Oulu University Hospital, Finland, due to an acute coronary syndrome (ACS) event (75% men, age 61 ± 12 years, BMI 27 ± 4 kg/m2, ejection fraction 62 ± 8, 89% have beta-blocker medication). Risk factors were assessed at the hospital immediately after the cardiac event, and health care costs for all reasons were collected from patient registers over a year. ECR was programmed in accordance with international guidelines. Risk analysis algorithms (cross-decomposition algorithms) were employed to rank risk factors based on variances in their effects. Regression analysis was used to determine the accounting value of risk factors by entering first the risk factor with the highest degree of explanation into the model. After that, the next most potent risk factor explaining costs was added to the model one by one (13 forecast models in total). Results: The ECR group used health care services during the year at an average of 1,624 ± 2,139€ per patient. Diabetes exhibited the strongest correlation with health care expenses (r = 0.406), accounting for 16% of the total costs (p < 0.001). When the next two ranked markers (body mass index; r = 0.171 and systolic blood pressure; r = - 0.162, respectively) were added to the model, the predictive value was 18% for the costs (p = 0.004). The depression scale had the weakest independent explanation rate of all 13 risk factors (explanation value 0.1%, r = 0.029, p = 0.811). Discussion: Presence of diabetes is the primary reason forecasting health care costs in 12-month ECR intervention among ACS patients. The ML tools may help decision-making when planning the optimal allocation of health care resources.

Machine learning models in predicting health care costs in patients with a recent acute coronary syndrome: A prospective pilot study.

Background: Health care budgets are limited, requiring the optimal use of resources. Machine learning (ML) methods may have an enormous potential for effective use of health care resources. Objective: We assessed the applicability of selected ML tools to evaluate the contribution of known risk markers for prognosis of coronary artery disease to predict health care costs for all reasons in patients with a recent acute coronary syndrome (n = 65, aged 65 ± 9 years) for 1-year follow-up. Methods: Risk markers were assessed at baseline, and health care costs were collected from electronic health registries. The Cross-decomposition algorithms were used to rank the considered risk markers based on their impacts on variances. Then regression analysis was performed to predict costs by entering the first top-ranking risk marker and adding the next-best markers, one by one, to build up altogether 13 predictive models. Results: The average annual health care costs were €2601 ± €5378 per patient. The Depression Scale showed the highest predictive value (r = 0.395), accounting for 16% of the costs (P = .001). When the next 2 ranked markers (LDL cholesterol, r = 0.230; and left ventricular ejection fraction, r = -0.227, respectively) were added to the model, the predictive value was 24% for the costs (P = .001). Conclusion: Higher depression score is the primary variable forecasting health care costs in 1-year follow-up among acute coronary syndrome patients. The ML tools may help decision-making when planning optimal utilization of treatment strategies.

Treed Gaussian Process Regression for Solving Offline Data-Driven Continuous Multiobjective Optimization Problems.

For offline data-driven multiobjective optimization problems (MOPs), no new data is available during the optimization process. Approximation models (or surrogates) are first built using the provided offline data, and an optimizer, for example, a multiobjective evolutionary algorithm, can then be utilized to find Pareto optimal solutions to the problem with surrogates as objective functions. In contrast to online data-driven MOPs, these surrogates cannot be updated with new data and, hence, the approximation accuracy cannot be improved by considering new data during the optimization process. Gaussian process regression (GPR) models are widely used as surrogates because of their ability to provide uncertainty information. However, building GPRs becomes computationally expensive when the size of the dataset is large. Using sparse GPRs reduces the computational cost of building the surrogates. However, sparse GPRs are not tailored to solve offline data-driven MOPs, where good accuracy of the surrogates is needed near Pareto optimal solutions. Treed GPR (TGPR-MO) surrogates for offline data-driven MOPs with continuous decision variables are proposed in this paper. The proposed surrogates first split the decision space into subregions using regression trees and build GPRs sequentially in regions close to Pareto optimal solutions in the decision space to accurately approximate tradeoffs between the objective functions. TGPR-MO surrogates are computationally inexpensive because GPRs are built only in a smaller region of the decision space utilizing a subset of the data. The TGPR-MO surrogates were tested on distance-based visualizable problems with various data sizes, sampling strategies, numbers of objective functions, and decision variables. Experimental results showed that the TGPR-MO surrogates are computationally cheaper and can handle datasets of large size. Furthermore, TGPR-MO surrogates produced solutions closer to Pareto optimal solutions compared to full GPRs and sparse GPRs.

Interactive multiobjective optimization for finding the most preferred exercise therapy modality in knee osteoarthritis.

BACKGROUND: There are no explicit guidelines or tools available to support clinicians in selecting exercise therapy modalities according to the characteristics of individual patients despite the apparent need. OBJECTIVE: This study develops a methodology based on a novel multiobjective optimization model and examines its feasibility as a decision support tool to support healthcare professionals in comparing different modalities and identifying the most preferred one based on a patient's needs. METHODS: Thirty-one exercise therapy modalities were considered from 21 randomized controlled trials. A novel interactive multiobjective optimization model was designed to characterize the efficacy of an exercise therapy modality based on five objectives: minimizing cost, maximizing pain reduction, maximizing disability improvement, minimizing the number of supervised sessions, and minimizing the length of the treatment period. An interactive model incorporates clinicians' preferences in finding the most preferred exercise therapy modality for each need. Multiobjective optimization methods are mathematical algorithms designed to identify the optimal balance between multiple conflicting objectives among available solutions/alternatives. They explicitly evaluate the conflicting objectives and support decision-makers in identifying the best balance. An experienced research-oriented physiotherapist was involved as a decision-maker in the interactive solution process testing the proposed decision support tool. RESULTS: The proposed methodology design and interactive process of the tool, including preference information, graphs, and exercise suggestions following the preferences, can help clinicians to find the most preferred exercise therapy modality based on a patient's needs and health status; paving the way to individualize recommendations. CONCLUSIONS: We examined the feasibility of our decision support tool using an interactive multiobjective optimization method designed to help clinicians balance between conflicting objectives to find the most preferred exercise therapy modality for patients with knee osteoarthritis. The proposed methodology is generic enough to be applied in any field of medical and healthcare settings, where several alternative treatment options exist.KEY MESSAGESWe demonstrate the potential of applying Interactive multiobjective optimization methods in a decision support tool to help clinicians compare different exercise therapy modalities and identify the most preferred one based on a patient's needs.The usability of the proposed decision support tool is tested and demonstrated in prescribing exercise therapy modalities to treat knee osteoarthritis patients.

Potential of interactive multiobjective optimization in supporting the design of a groundwater biodenitrification process.

The design of water treatment plants requires simultaneous analysis of technical, economic and environmental aspects, identified by multiple conflicting objectives. We demonstrated the advantages of an interactive multiobjective optimization (MOO) method over a posteriori methods in an unexplored field, namely the design of a biological treatment plant for drinking water production, that tackles the process drawbacks, contrarily to what happens in a traditional volumetric-load-driven design procedure. Specifically, we consider a groundwater denitrification biofilter, simulated by the Activated Sludge Model modified with two-stage denitrification kinetics. Three objectives were defined (nitrate removal efficiency, drawbacks on produced water, investment and management costs) and the interactive method NIMBUS applied to identify the best-suited design without any a priori evaluation, as for volumetric-load-driven design procedures. When compared to an evolutionary MOO algorithm, the interactive solution process was faster, more understandable and user-friendly and supported the decision maker well in identifying the most preferred solution (main design/operating parameters) to be implemented. Approach strength has been proved through both sensitivity analysis and positive experimental validation through a pilot scale biofilter operated for three months. In synthesis, without any "a priori" evaluation based on practical experience, the MOO design approach allowed obtaining a preferred Pareto optimal design, characterized by volumetric loading in the range 0.85-2.54 kgN m-3 d-1 (EBCTs: 5-15 min), a carbon dosage of 0.5-0.8 gC,dos/gC,stoich, with SRTs in the range 4-27 d.

Projections onto the Pareto surface in multicriteria radiation therapy optimization.

PURPOSE: To eliminate or reduce the error to Pareto optimality that arises in Pareto surface navigation when the Pareto surface is approximated by a small number of plans. METHODS: The authors propose to project the navigated plan onto the Pareto surface as a postprocessing step to the navigation. The projection attempts to find a Pareto optimal plan that is at least as good as or better than the initial navigated plan with respect to all objective functions. An augmented form of projection is also suggested where dose-volume histogram constraints are used to prevent that the projection causes a violation of some clinical goal. The projections were evaluated with respect to planning for intensity modulated radiation therapy delivered by step-and-shoot and sliding window and spot-scanned intensity modulated proton therapy. Retrospective plans were generated for a prostate and a head and neck case. RESULTS: The projections led to improved dose conformity and better sparing of organs at risk (OARs) for all three delivery techniques and both patient cases. The mean dose to OARs decreased by 3.1 Gy on average for the unconstrained form of the projection and by 2.0 Gy on average when dose-volume histogram constraints were used. No consistent improvements in target homogeneity were observed. CONCLUSIONS: There are situations when Pareto navigation leaves room for improvement in OAR sparing and dose conformity, for example, if the approximation of the Pareto surface is coarse or the problem formulation has too permissive constraints. A projection onto the Pareto surface can identify an inaccurate Pareto surface representation and, if necessary, improve the quality of the navigated plan.

Spatially dynamic forest management to sustain biodiversity and economic returns.

Production of marketed commodities and protection of biodiversity in natural systems often conflict and thus the continuously expanding human needs for more goods and benefits from global ecosystems urgently calls for strategies to resolve this conflict. In this paper, we addressed what is the potential of a forest landscape to simultaneously produce habitats for species and economic returns, and how the conflict between habitat availability and timber production varies among taxa. Secondly, we aimed at revealing an optimal combination of management regimes that maximizes habitat availability for given levels of economic returns. We used multi-objective optimization tools to analyze data from a boreal forest landscape consisting of about 30,000 forest stands simulated 50 years into future. We included seven alternative management regimes, spanning from the recommended intensive forest management regime to complete set-aside of stands (protection), and ten different taxa representing a wide variety of habitat associations and social values. Our results demonstrate it is possible to achieve large improvements in habitat availability with little loss in economic returns. In general, providing dead-wood associated species with more habitats tended to be more expensive than providing requirements for other species. No management regime alone maximized habitat availability for the species, and systematic use of any single management regime resulted in considerable reductions in economic returns. Compared with an optimal combination of management regimes, a consistent application of the recommended management regime would result in 5% reduction in economic returns and up to 270% reduction in habitat availability. Thus, for all taxa a combination of management regimes was required to achieve the optimum. Refraining from silvicultural thinnings on a proportion of stands should be considered as a cost-effective management in commercial forests to reconcile the conflict between economic returns and habitat required by species associated with dead-wood. In general, a viable strategy to maintain biodiversity in production landscapes would be to diversify management regimes. Our results emphasize the importance of careful landscape level forest management planning because optimal combinations of management regimes were taxon-specific. For cost-efficiency, the results call for balanced and correctly targeted strategies among habitat types.

Interactive multiobjective optimization for anatomy-based three-dimensional HDR brachytherapy.

In this paper, we present an anatomy-based three-dimensional dose optimization approach for HDR brachytherapy using interactive multiobjective optimization (IMOO). In brachytherapy, the goals are to irradiate a tumor without causing damage to healthy tissue. These goals are often conflicting, i.e. when one target is optimized the other will suffer, and the solution is a compromise between them. IMOO is capable of handling multiple and strongly conflicting objectives in a convenient way. With the IMOO approach, a treatment planner's knowledge is used to direct the optimization process. Thus, the weaknesses of widely used optimization techniques (e.g. defining weights, computational burden and trial-and-error planning) can be avoided, planning times can be shortened and the number of solutions to be calculated is small. Further, plan quality can be improved by finding advantageous trade-offs between the solutions. In addition, our approach offers an easy way to navigate among the obtained Pareto optimal solutions (i.e. different treatment plans). When considering a simulation model of clinical 3D HDR brachytherapy, the number of variables is significantly smaller compared to IMRT, for example. Thus, when solving the model, the CPU time is relatively short. This makes it possible to exploit IMOO to solve a 3D HDR brachytherapy optimization problem. To demonstrate the advantages of IMOO, two clinical examples of optimizing a gynecologic cervix cancer treatment plan are presented.

A preference-based evolutionary algorithm for multi-objective optimization.

In this paper, we discuss the idea of incorporating preference information into evolutionary multi-objective optimization and propose a preference-based evolutionary approach that can be used as an integral part of an interactive algorithm. One algorithm is proposed in the paper. At each iteration, the decision maker is asked to give preference information in terms of his or her reference point consisting of desirable aspiration levels for objective functions. The information is used in an evolutionary algorithm to generate a new population by combining the fitness function and an achievement scalarizing function. In multi-objective optimization, achievement scalarizing functions are widely used to project a given reference point into the Pareto optimal set. In our approach, the next population is thus more concentrated in the area where more preferred alternatives are assumed to lie and the whole Pareto optimal set does not have to be generated with equal accuracy. The approach is demonstrated by numerical examples.

Multiobjective optimization of an ultrasonic transducer using NIMBUS.

The optimal design of an ultrasonic transducer is a multiobjective optimization problem since the final outcome needs to satisfy several conflicting criteria. Simulation tools are often used to avoid expensive and time-consuming experiments, but even simulations may be inefficient and lead to inadequate results if they are based only on trial and error. In this work, the interactive multiobjective optimization method NIMBUS is applied in designing a high-power ultrasonic transducer. The performance of the transducer is simulated with a finite element model, and three design goals are formulated as objective functions to be minimized. To find an appropriate compromise solution, additional preference information is needed from a decision maker, who in our case is an expert in transducer design. A realistic design problem is formulated, and an interactive solution process is described. Our findings demonstrate that interactive multiobjective optimization methods, combined with numerical simulation models, can efficiently help in finding new solution approaches and possibilities as well as new understanding of real-life problems as entirenesses. In this case, the decision maker found a solution that was better with respect to all three objectives than the conventional unoptimized design.