Simulation modeling and health-care decision making

Development of dynamic health care delivery heatmaps for end-of-life cancer care: a cohort study

OBJECTIVE

Measures of variation in end-of-life (EOL) care intensity across hospitals are typically summarised using unidimensional measures. These measures do not capture the full dimensionality of complex clinical care trajectories over time that are needed to inform quality improvement efforts. The objective is to develop a novel visual map of EOL care trajectories that illustrates multidimensional utilisation over time.

SETTING

United States' National Cancer Institute or National Comprehensive Cancer Network (NCI/NCCN)-designated hospitals.

PARTICIPANTS

We identified Medicare claims for fee-for-service beneficiaries with poor prognosis cancers who died between April and December 2016 and received the preponderance of treatment in the last 6 months of life at an NCI/NCCN-designated hospital.

DESIGN

For each beneficiary, we transformed each Medicare claim into two elements to generate a two-dimensional individual-level heatmap. On the y-axis, each claim was classified into a categorical description of the service delivered by a healthcare resource. On the x-axis, the date for each claim was converted into the day number prior to death it occurred on. We then summed up individual-level heatmaps of patients attributed to each hospital to generate two-dimensional hospital-level heatmaps. We used four case studies to illustrate the feasibility of interpreting these heatmaps and to shed light on how they might be used to guide value-based, quality improvement initiatives.

RESULTS

We identified nine distinct EOL care delivery patterns from hospital-level heatmaps based on signal intensity and patterns for inpatient, outpatient and home-based hospice services. We illustrate that in most cases, heatmaps illustrating patterns of multidimensional healthcare utilisation over time provide more information about care trajectories and highlight more heterogeneity than current unidimensional measures.

CONCLUSIONS

This study illustrates the feasibility of representing multidimensional EOL utilisation over time as a heatmap. These heatmaps may provide potentially actionable insights into hospital-level care delivery patterns, and the approach may generalise to other serious illness populations.

Symbiotic simulation for the operational management of inpatient beds: model development and validation using Δ-method

In many modern hospitals, resources are shared between patients who require immediate care, and must be dealt with as they arrive (emergency patients), and those whose care requirements are partly known to the hospital some time in advance (elective patients). Catering for these two types of patients is a challenging short-term operational decision-making problem, since some portion of each resource must be set aside for emergency patients when planning for the number and type of elective patients to admit. This paper shows how symbiotic simulation can help hospitals with important short-term operational decision making. We demonstrate how a symbiotic simulation model can be developed from an existing simulation model by adding the ability to load the state of the physical system at run-time and by making use of conditional length-of-stay distributions. The model is parameterised using 18 months of patient administrative data from an Anonymised General Hospital. Further, we propose a new Δ-Method that is suitable for validating a stochastic symbiotic simulation model. We demonstrate the benefit of our symbiotic simulation by showing how it can be used as an early warning system, and how additional patient-level information which might only become available after admission, can affect the predicted bed census.

How the health-seeking behaviour of pregnant women affects neonatal outcomes: findings of system dynamics modelling in Pakistan

Background

Limited studies have explored how health-seeking behaviour during pregnancy through to delivery affect neonatal outcomes. We modelled health-seeking behaviour across urban and rural settings in Pakistan, where poor neonatal outcomes persist with wide disparities.

Methods and findings

A system dynamics model was developed and parameterised. Following validation tests, the model was used to determine neonatal mortality for pregnant women considering their decisions to access, refuse and switch antenatal care services in four provider sectors: public, private, traditional and charitable. Four health-seeking scenarios were tested across different pregnancy trimesters. Health-seeking behaviour in different subgroups by geographical locations and social network effect was modelled. The largest reduction in neonatal mortality was achieved with antenatal care provided by skilled providers in public, private or charitable sectors, combined with the use of institutional delivery. Women’s social networks had strong influences on if, when and where to seek care. Interventions by Lady Health Workers had a minimal impact on health-seeking behaviour and neonatal outcomes after trimester 1. Optimal benefits were achieved for urban women when antenatal care was accessed within trimester 2, but for rural women within trimester 1. Antenatal care access delayed to trimester 3 had no protective impact on neonatal mortality.

Conclusions

System dynamics modelling enables capturing the complexity of health-seeking behaviours and impact on outcomes, informing intervention design, implementation of targeted policies and uptake of services specific to urban/rural settings considering structural enablers/barriers to access, cultural contexts and strong social network influences.

Simulation Modelling in Healthcare: An Umbrella Review of Systematic Literature Reviews

BACKGROUND

Numerous studies examine simulation modelling in healthcare. These studies present a bewildering array of simulation techniques and applications, making it challenging to characterise the literature.

OBJECTIVE

The aim of this paper is to provide an overview of the level of activity of simulation modelling in healthcare and the key themes.

METHODS

We performed an umbrella review of systematic literature reviews of simulation modelling in healthcare. Searches were conducted of academic databases (JSTOR, Scopus, PubMed, IEEE, SAGE, ACM, Wiley Online Library, ScienceDirect) and grey literature sources, enhanced by citation searches. The articles were included if they performed a systematic review of simulation modelling techniques in healthcare. After quality assessment of all included articles, data were extracted on numbers of studies included in each review, types of applications, techniques used for simulation modelling, data sources and simulation software.

RESULTS

The search strategy yielded a total of 117 potential articles. Following sifting, 37 heterogeneous reviews were included. Most reviews achieved moderate quality rating on a modified AMSTAR (A Measurement Tool used to Assess systematic Reviews) checklist. All the review articles described the types of applications used for simulation modelling; 15 reviews described techniques used for simulation modelling; three reviews described data sources used for simulation modelling; and six reviews described software used for simulation modelling. The remaining reviews either did not report or did not provide enough detail for the data to be extracted.

CONCLUSION

Simulation modelling techniques have been used for a wide range of applications in healthcare, with a variety of software tools and data sources. The number of reviews published in recent years suggest an increased interest in simulation modelling in healthcare.

Decreasing patient length of stay via new flexible exam room allocation policies in ambulatory care clinics

To address prolonged lengths of stay (LOS) in ambulatory care clinics, we analyze the impact of implementing flexible and dynamic policies for assigning exam rooms to providers. In contrast to the traditional approaches of assigning specific rooms to each provider or pooling rooms among all practitioners, we characterize the impact of alternate compromise policies that have not been explored in previous studies. Since ambulatory care patients may encounter multiple different providers in a single visit, room allocation can be determined separately for each encounter accordingly. For the first phase of the visit, conducted by the medical assistant, we define a dynamic room allocation policy that adjusts room assignments based on the current state of the clinic. For the second phase of the visit, conducted by physicians, we define a series of room sharing policies which vary based on two dimensions, the number of shared rooms and the number of physicians sharing each room. Using a discrete event simulation model of an outpatient cardiovascular clinic, we analyze the benefits and costs associated with the proposed room allocation policies. Our findings show that it is not necessary to fully share rooms among providers in order to reduce patient LOS and physician idle time. Instead, most of the benefit of pooling can be achieved by implementation of a compromise room allocation approach, limiting the need for significant organizational changes within the clinic. Also, in order to achieve most of the benefits of room allocation policies, it is necessary to increase flexibility in the two dimensions simultaneously. These findings are shown to be consistent in settings with alternate patient scheduling and distinctions between physicians.

Core discrete event simulation model for the evaluation of health care technologies in major depressive disorder

OBJECTIVE

A review of existing economic models in major depressive disorder (MDD) highlighted the need for models with longer time horizons that also account for heterogeneity in treatment pathways between patients. A core discrete event simulation model was developed to estimate health and cost outcomes associated with alternative treatment strategies.

METHODS

This model simulated short- and long-term clinical events (partial response, remission, relapse, recovery, and recurrence), adverse events, and treatment changes (titration, switch, addition, and discontinuation) over up to 5 years. Several treatment pathways were defined on the basis of fictitious antidepressants with three levels of efficacy, tolerability, and price (low, medium, and high) from first line to third line. The model was populated with input data from the literature for the UK setting. Model outputs include time in different health states, quality-adjusted life-years (QALYs), and costs from National Health Service and societal perspectives. The codes are open source.

RESULTS

Predicted costs and QALYs from this model are within the range of results from previous economic evaluations. The largest cost components from the payer perspective were physician visits and hospitalizations. Key parameters driving the predicted costs and QALYs were utility values, effectiveness, and frequency of physician visits. Differences in QALYs and costs between two strategies with different effectiveness increased approximately twofold when the time horizon increased from 1 to 5 years.

CONCLUSION

The discrete event simulation model can provide a more comprehensive evaluation of different therapeutic options in MDD, compared with existing Markov models, and can be used to compare a wide range of health care technologies in various groups of patients with MDD.

A proposed model for economic evaluations of major depressive disorder

In countries like UK and Australia, the comparability of model-based analyses is an essential aspect of reimbursement decisions for new pharmaceuticals, medical services and technologies. Within disease areas, the use of models with alternative structures, type of modelling techniques and/or data sources for common parameters reduces the comparability of evaluations of alternative technologies for the same condition. The aim of this paper is to propose a decision analytic model to evaluate long-term costs and benefits of alternative management options in patients with depression. The structure of the proposed model is based on the natural history of depression and includes clinical events that are important from both clinical and economic perspectives. Considering its greater flexibility with respect to handling time, discrete event simulation (DES) is an appropriate simulation platform for modelling studies of depression. We argue that the proposed model can be used as a reference model in model-based studies of depression improving the quality and comparability of studies.

A critical review of model-based economic studies of depression: modelling techniques, model structure and data sources

Depression is the most common mental health disorder and is recognized as a chronic disease characterized by multiple acute episodes/relapses. Although modelling techniques play an increasingly important role in the economic evaluation of depression interventions, comparatively little attention has been paid to issues around modelling studies with a focus on potential biases. This, however, is important as different modelling approaches, variations in model structure and input parameters may produce different results, and hence different policy decisions. This paper presents a critical review of literature on recently published model-based cost-utility studies of depression. Taking depression as an illustrative example, through this review, we discuss a number of specific issues in relation to the use of decision-analytic models including the type of modelling techniques, structure of models and data sources. The potential benefits and limitations of each modelling technique are discussed and factors influencing the choice of modelling techniques are addressed. This review found that model-based studies of depression used various simulation techniques. We note that a discrete-event simulation may be the preferred technique for the economic evaluation of depression due to the greater flexibility with respect to handling time compared with other individual-based modelling techniques. Considering prognosis and management of depression, the structure of the reviewed models are discussed. We argue that a few reviewed models did not include some important structural aspects such as the possibility of relapse or the increased risk of suicide in patients with depression. Finally, the appropriateness of data sources used to estimate input parameters with a focus on transition probabilities is addressed. We argue that the above issues can potentially bias results and reduce the comparability of economic evaluations.

A review on the relation between simulation and improvement in hospitals

Background

Simulation applications on operations management in hospitals are frequently published and claim to support decision-making on operations management subjects. However, the reported implementation rates of recommendations are low and the actual impact of the changes recommended by the modeler has hardly been examined. This paper examines: 1) the execution rate of simulation study recommendations, 2) the research methods used to evaluate implementation of recommendations, 3) factors contributing to implementation, and 4) the differences regarding implementation between literature and practice.

Results

Altogether 16 hospitals executed the recommendations (at least partially). Implementation results were hardly reported upon; 1 study described a before-and-after design, 2 a partial before and after design. Factors that help implementation were grouped according to 1) technical quality, of which data availability, validation/verification with historic data/expert opinion, and the development of the conceptual model were mentioned most frequently 2) process quality, with client involvement and 3) outcome quality with, presentation of results. The survey response rate of traceable authors was 61%, 18 authors implemented the results at least partially. Among these responses, evaluation methods were relatively better with 3 time series designs and 2 before-and-after designs.

Conclusions

Although underreported in literature, implementation of recommendations seems limited; this review provides recommendations on project design, implementation conditions and evaluation methods to increase implementation.

Methods

A literature review in PubMed and Business Source Elite on stochastic simulation applications on operations management in individual hospitals published between 1997 and 2008. From those reporting implementation, cross references were added. In total, 89 papers were included. A scoring list was used for data extraction. Two reviewers evaluated each paper separately; in case of discrepancies, they jointly determined the scores. The findings were validated with a survey to the original authors.

Operational Performance Analysis of a Public Hospital Laboratory

Efficient utilization of scarce resources is an issue for any healthcare system. In developing countries, proper tools, techniques, and resources must be widely used in healthcare operational planning. Considering the necessity of effective resource planning, this study focuses on the rural healthcare system of Bangladesh and concentrates on the sub-district government hospital laboratory. The authors’ determine possible ways to improve operations of laboratory facilities. To analyze existing system efficiency, sample laboratory data is fed into a simulation model. This paper identifies several possible ways for future expansion and suggests using simulation for better planning and analysis.

Design and analysis of a health care clinic for homeless people using simulations

PURPOSE

Improving quality of care is important in health care management. For health care clinics, reducing patient waiting time and improving throughput with efficient utilization of the workforce are important issues to achieve better quality of care. This paper seeks to introduce a simulation study on design and analysis of a health clinic for homeless patients in Lexington, Kentucky, USA.

DESIGN/METHODOLOGY/APPROACH

Using the simulation model, the patient flow of the clinic and analyze quality of care for different staffing levels is simulated. In addition, the dependence of distributions on service times is investigated. Moreover, the impact of service time variability on quality of care (e.g. patient waiting time) is analyzed.

FINDINGS

The necessary staffing level and utilizations to reduce patient waiting times and improve throughput to achieve better quality of care are obtained. In addition, it is shown that the system performance is primarily dependent on the mean and coefficients of variation, rather than a complete distribution, of service times. In addition, a piece-wise linear approximation formula is proposed so that patient waiting time in the clinic can be estimated for any variability with only two simulations.

RESEARCH LIMITATIONS/IMPLICATIONS

The simulation method may need long model development time and long simulation executing time for complex systems.

PRACTICAL IMPLICATIONS

The quality of care delivery in a health care clinic can be evaluated using simulations. The results presented in the paper provide an easier approach for medical practitioners to evaluate different scenarios, examine needed resources, and carry out what-if analysis to predictthe impact of any changes in the system, to determine an optimal system configuration.

ORIGINALITY/VALUE

The paper shows that such models provide a quantitative tool for clinic operations and management to achieve better care quality. Moreover, it can be easily adapted to model other health care facilities, such as hospitals, emergency rooms, operating rooms, supply chain in health care industry.

Predicting joint replacement waiting times

Currently, the median waiting time for total hip and knee replacement in Ontario is greater than 6 months. Waiting longer than 6 months is not recommended and may result in lower post-operative benefits. We developed a simulation model to estimate the proportion of patients who would receive surgery within the recommended waiting time for surgery over a 10-year period considering a wide range of demand projections and varying the number of available surgeries. Using an estimate that demand will grow by approximately 8.7% each year for 10 years, we determined that increasing available supply by 10% each year was unable to maintain the status quo for 10 years. Reducing waiting times within 10 years required that the annual supply of surgeries increased by 12% or greater. Allocating surgeries across regions in proportion to each region's waiting time resulted in a more efficient distribution of surgeries and a greater reduction in waiting times in the long-term compared to allocation strategies based only on the region's population size.

Visualizing the demand for various resources as a function of the master surgery schedule: a case study

This paper presents a software system that visualizes the impact of the master surgery schedule on the demand for various resources throughout the rest of the hospital. The master surgery schedule can be seen as the engine that drives the hospital. Therefore, it is very important for decision makers to have a clear image on how the demand for resources is linked to the surgery schedule. The software presented in this paper enables schedulers to instantaneously view the impact of, e.g., an exchange of two block assignments in the master surgery schedule on the expected resource consumption pattern. A case study entailing a large Belgian surgery unit illustrates how the software can be used to assist in building better surgery schedules.

A clinically based discrete-event simulation of end-stage liver disease and the organ allocation process

BACKGROUND

The optimal allocation of scarce donor livers is a contentious health care issue requiring careful analysis. The objective of this article was to design a biologically based discrete-event simulation to test proposed changes in allocation policies.

METHODS

The authors used data from multiple sources to simulate end-stage liver disease and the complex allocation system. To validate the model, they compared simulation output with historical data.

RESULTS

Simulation outcomes were within 1% to 2% of actual results for measures such as new candidates, donated livers, and transplants by year. The model overestimated the yearly size of the waiting list by 5% in the last year of the simulation and the total number of pretransplant deaths by 10%.

CONCLUSION

The authors created a discrete-event simulation model that represents the biology of end-stage liver disease and the health care organization of transplantation in the United States.

Advances in detecting and responding to threats from bioterrorism and emerging infectious disease

Much progress has been made in recent years to strengthen local, state, national and international capacities to detect and respond to bioterrorism events and naturally occurring outbreaks of disease. New tools and systems are available to estimate the potential impact of a biological event and predict resource needs for effective response, enable earlier detection of an attack or outbreak, enhance diagnostic capacity and facilitate rapid intervention to mitigate the impact of an event on a community. These advances have required new approaches to preparedness, planning and surveillance, as well as new partnerships and collaborations across a range of disciplines. We examine some of these developments, discuss potential uses and limitations of these approaches, and identify priorities for the future.

Modeling the public health response to bioterrorism: using discrete event simulation to design antibiotic distribution centers

BACKGROUND

Post-exposure prophylaxis is a critical component of the public health response to bioterrorism. Computer simulation modeling may assist in designing antibiotic distribution centers for this task.

METHODS

The authors used discrete event simulation modeling to determine staffing levels for entry screening, triage, medical evaluation, and drug dispensing stations in a hypothetical antibiotic distribution center operating in low, medium, and high disease prevalence bioterrorism response scenarios. Patient arrival rates and processing times were based on prior mass prophylaxis campaigns. Multiple sensitivity analyses examined the relationship between average staff utilization rate (UR) (i.e., percentage of time occupied in patient contact) and capacity of the model to handle surge arrivals.

RESULTS

Distribution center operation required from 93 staff for the low-prevalence scenario to 111 staff for the high-prevalence scenario to process approximately 1000 people per hour within the baseline model assumptions. Excess capacity to process surge arrivals approximated (1-UR) for triage staffing.

CONCLUSIONS

Discrete event simulation modeling is a useful tool in developing the public health infrastructure for bioterrorism response. Live exercises to validate the assumptions and outcomes presented here may improve preparedness to respond to bioterrorism.

A simulation modelling approach to evaluating alternative policies for the management of the waiting list for liver transplantation

A shortage of donor liver grafts unfortunately results in approximately 10% of patients dying whilst listed for a liver transplant in Europe and the United States. Thus it is imperative that all available organs are used as efficiently as possible. This paper reports upon the application of a simulation modelling approach to assess the impact of several alternative allocation policies upon the cost effectiveness of this technology at one liver transplant centre in the UK. The impact of changes in allocation criteria on the estimated net life expectancy, average net costs and overall cost effectiveness of the transplantation programme were evaluated. The incremental cost effectiveness ratio (ICER) for the base case allocation policy, based upon the time spent on the waiting list (i.e., longest wait first) was 11,557 pounds sterling at 1999 prices. The ICERs associated with an allocation policy based upon age (lowest age first), and an allocation policy based upon the severity of the pre-transplant condition of the patient (with most severely ill patients given a lower priority) were lower than the base case at 10,424 pounds sterling and 9,077 pounds sterling, respectively. The results of this modelling study suggest that the overall cost effectiveness of the liver transplantation programme could be improved if the current allocation policy were modified to give more weight to the age of the patient and the reduced chances of success of the most severely ill patients.

Selecting a decision model for economic evaluation: a case study and review

The increased use of modelling techniques as a methodological tool in the economic evaluation of health care technologies has, in the main, been limited to two approaches--decision trees and Markov chain models. The former are suited to modelling simple scenarios that occur over a short time period, whilst Markov chain models allow longer time periods to be modelled, in continuous time, where the timing of an event is uncertain. In the context of economic evaluation, a less well developed technique is discrete event simulation, which may allow even greater flexibility. Taking the economic evaluation of adjuvant therapies for breast cancer as an illustrative example, the process of building a decision tree, a Markov chain model, and a discrete event simulation model are described. The potential benefits and problems of each approach are discussed. The suitability of the modelling techniques to economic evaluations of health care programmes in general is then discussed. This section aims to illustrate the areas in which the alternative modelling methods may be most appropriately employed.

Presenting clinical trial information: a comparison of methods

The study objective was to assess the relative effects of 2 approaches to teaching about a clinical trial, in terms of patients' satisfaction, information understanding, and whether or not they would enter such a trial. One hundred patients receiving radiation therapy for a variety of cancer diagnoses were randomized to receive information about a hypothetical trial, either by audio tape or interactive computer program. A day later, information understanding was assessed. One week later, method satisfaction and whether respondents would enter such a trial were assessed. There were no differences in understanding or satisfaction. Members of the computer program group tended to report a more positive attitude towards trial entry (chi 2 = 4.0; 1 df; P = 0.05). Overall, refusers tended to be women with higher understanding scores. The results suggest that teaching with interactive components might not adversely affect trial accrual. Further work involving an actual trial entry decision is merited; the sex of the respondent should be controlled in designing this future work.