Loading...
Publication
Application of discrete event simulation in the performance analysis of an outpatient unit
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.71 MB | Adobe PDF |
Advisor(s)
Abstract(s)
O Departamento da Consulta Externa do Grupo HPA Saúde, Unidade Hospitalar de Gambelas, oferece aos seus pacientes uma ampla oferta em especialidades médicas e cirúrgicas. No início de 2021, a consulta externa sofreu uma expansão e reestruturação, dando resposta ao elevado aumento na procura de exames, permitindo dar resposta aos pacientes, tendo como premissa a manutenção da elevada qualidade do serviço prestado. Esta reestruturação do serviço de ambulatório, conseguiu aumentar a capacidade de resposta, a qualidade das suas instalações e os respetivos serviços prestados. A Unidade de Exames Especiais é uma das unidades integrantes deste departamento, onde são realizados exames específicos de várias especialidades tais como, cardiologia, gastroenterologia, urologia e pneumologia. A unidade foi reestruturada e iniciou funções em maio de 2021. A nível da gestão operacional, esta reestruturação e ampliação da capacidade física da unidade, trouxe vários desafios, como a alteração dos fluxos de pacientes, a redistribuição de recursos físicos e humanos, e a necessidade de fazer um adequado planeamento da capacidade da unidade. O planeamento da capacidade é complexo e é um grande desafio para os gestores, pois o desequilíbrio entre a oferta e a procura pode desencadear aumentos de tempo de espera de pacientes ou desperdício de recursos. Qualquer um destes desfechos negativos põe em causa o desempenho do serviço, tornando-o menos eficiente. Realizar um adequado planeamento da capacidade não é, no entanto, tarefa fácil dada a imprevisibilidade associada a estes serviços, a sazonalidade da procura, as flutuações de pacientes ao longo do dia e os padrões de comportamento dos próprios pacientes. Ainda assim, torna-se muito importante estudar qual o impacto que a alteração ou o ajustamento de determinados parâmetros ao nível deste planeamento poderá ter no desempenho da Unidade de Exames Especiais, para que sejam implementadas estratégias de melhoria da eficiência dos serviços e, consequentemente, do seu desempenho. Este estudo tem como objetivo explorar a aplicação de um modelo de Simulação de Eventos Discretos (DES) para analisar o impacto do planeamento de capacidade no desempenho da Unidade de Exames Especiais. Em particular, este estudo tem como objetivos específicos, identificar quais as variáveis que mais condicionam o planeamento de capacidade e ainda identificar relações/interações entre o planeamento de capacidade e o tempo de espera dos pacientes. A aplicação do modelo DES, irá permitir recriar o funcionamento do sistema real e realizar várias simulações, de forma a compreender como é que o sistema funciona e de que forma estratégias alternativas poderão afetar o desempenho do sistema. Irá também permitir perceber qual é o impacto do planeamento de capacidade no desempenho do serviço e qual é a relação entre a capacidade de planeamento e o tempo de espera dos pacientes. Dada a complexidade da unidade e o caráter exploratório deste estudo, apenas foram considerados para o estudo os exames realizados pela especialidade de gastroenterologia. Foram recolhidos dados referentes a todos os exames realizados às quintas-feiras na Unidade de Exames Especiais entre janeiro e abril de 2022. Nesse período, 261 pacientes tiveram exames agendados, mas 8 não compareceram. Os dados tratados referem-se, portanto, a um total de 253 pacientes. Esses dados foram extraídos do processo eletrónico do paciente, registos de enfermagem do processo clínico e do programa de agendamento de consultas. O modelo DES foi aplicado para modelar o fluxo de pacientes da unidade utilizando o software SIMUL8, versão 29.0. Após a análise dos dados e a identificação das distribuições estatística mais adequadas para representar o comportamento de alguns parâmetros, o modelo foi devidamente configurado. Para validar o modelo DES, os resultados e a animação da simulação foram verificados por elementos da equipa da unidade. Uma análise comparativa foi também realizada através da comparação de dados reais com dados e resultados gerados pela simulação. A maioria das variáveis apresentou erros relativos substancialmente baixos, por exemplo, o tempo médio de permanência no sistema, obteve um erro relativo de 0,07. Os tempos médios dos procedimentos, também obtiveram bons resultados. Nomeadamente, o procedimento de colonoscopia dupla, colonoscopia baixa e endoscopia alta, obtiveram erros relativos negligenciáveis (de 0,01, 0,00 e 0,00 respetivamente). No entanto a variável desvio padrão do tempo no sistema apresentou um erro relativo superior a 10%. A diferença observada entre os valores reais e simulados podem estar associadas ao processo de desconstrução do modelo. O processo de preparação do paciente e o processo de recobro foram desconstruídos, em atividades menores e específicas, com o objetivo de compreender melhor a taxa de utilização dos recursos nestas mesmas atividades. Essa discrepância pode ser atribuída à utilização de uma distribuição média para atividades em vez de empregar uma distribuição adequada, o que permitiria diminuir a variabilidade da amostra resultando num valor de desvio padrão mais baixo. Foram testados oito cenários, com o objetivo de identificar limitações e gargalos, maximizar a capacidade e melhorar o agendamento de exames e a otimização de recursos. No Cenário nº 1, foi otimizado o número de agendamentos através da redução dos intervalos entre agendamentos de 30 minutos para 20 minutos em comparação ao modelo de simulação inicial, mantendo o horário de trabalho. O objetivo era maximizar a eficiência do modelo. Observou-se um aumento notável no número de exames realizados e na taxa de utilização de recursos. No entanto, isso levou a um aumento significativo no tempo de espera e no tempo de permanência do paciente no sistema. A sobrecarga do sistema neste cenário permitiu identificar um comportamento de gargalo na atividade do vestiário com uma taxa de congestionamento de 26,1% o que indica que a atividade de vestiário sofreu um alto índice de congestionamento e se tornou um fator limitante no fluxo de pacientes no processo. A adição de um segundo vestiário no Cenário nº 2, mostrou uma diminuição dos tempos de espera e da taxa de congestionamento no vestiário, que reduziu de 26,1% para 15,96%, validando a opinião de especialistas de que esta atividade assume um comportamento de gargalo. Os Cenários nº 3 e nº 4 foram conduzidos para testar a otimização de recursos com o objetivo de identificar limitações e comportamentos de gargalo em outras atividades ou recursos, visto que o Cenário nº 2 ainda observava congestionamento na atividade de vestiário. No Cenário nº 3, a adição de um enfermeiro extra para o processo de preparação e recobro teve um efeito positivo no sistema. Isso resultou em uma redução no tempo de preparação do paciente e no tempo de permanência no sistema. O congestionamento na atividade vestiário, diminuiu de 15,96% para 13,34%, o que pode ser interpretado como um potencial comportamento de gargalo do recurso enfermeiro alocado à preparação do paciente e recobro, nomeadamente quando está apenas um enfermeiro disponível. Neste cenário nº 3, o acréscimo de um enfermeiro, contribuiu para uma redução de 2,62% no congestionamento da atividade de vestiário. No entanto, verificou-se também que com esta adição extra, a taxa de utilização de recursos não foi eficiente, pois a taxa de utilização do enfermeiro do processo de preparação e recobro do paciente atingiu apenas 43%. No Cenário nº 4, a adição de mais duas camas, ajudou a aliviar o comportamento de gargalo na atividade de vestiário. Isso levou a uma diminuição significativa da taxa de congestionamento nessa atividade de 13,34% para 2,38%. No entanto, também resultou em um aumento notável no tempo de espera para os exames e um discreto aumento no tempo de permanência no sistema. Com base nas constatações do Cenário nº 4, que indicou um possível gargalo no processo de procedimento de exame, foi realizado um cenário de otimização de fluxos e de recursos. Nesse cenário (i.e. Cenário nº5), foi adicionada uma segunda sala de procedimentos, com uma equipa independente composta por médicos e enfermeiros. A adição da segunda sala de procedimentos resultou em uma diminuição significativa tanto no tempo de permanência do paciente no sistema quanto no tempo de espera. Observou-se, também, uma redução significativa da taxa de congestionamento da atividade vestiário, validando a suspeita que a atividade sala de procedimento assume um comportamento de gargalo, e que a adição de uma segunda sala, pode reduzir significativamente a taxa de congestionamento. No entanto, observou-se uma taxa de utilização dos recursos ineficiente. No Cenário nº 6, devido às baixas taxas de utilização de recursos observadas no Cenário nº 5, foi utilizado o mesmo modelo de simulação mas com o objetivo de maximizar a capacidade através da otimização do fluxo, recursos e agendamento. Em particular, o Cenário nº 6 consubstanciou-se na otimização do agendamento (exames x agendados a cada 12,5 minutos) com o intuito de maximizar a capacidade, mantendo a otimização de fluxos (dois vestiários e duas salas de procedimento) e recursos adequados ao incremento das salas de procedimento, duplicando o número de elementos das equipas e adicionando duas camas extras. Este cenário apresentou um alto nível de otimização de recursos, utilizando efetivamente os recursos disponíveis, otimizando o fluxo de pacientes e implementando uma abordagem de agendamento aprimorada. Isso resultou em maior eficiência, redução do tempo de espera e maior produtividade em relação ao número de exames realizados. O cenário nº 6 emergiu como o cenário mais eficiente testado, no entanto, a sua implementação poderá não ser muito realista dado que obrigaria a modificações físicas do espaço, ajustamentos no fluxo dos pacientes e uma duplicação de alguns recursos. Com o intuito de testar um cenário mais realista, foi criado o cenário nº 7 usando a mesma modelação que o cenário nº 1, mas otimizando o agendamento de exames (exames agendados a cada 24,5 minutos) e os recursos (mantendo os mesmos que no cenário nº 1, mas com taxas de utilização elevadas). Este cenário teve em conta as restrições causadas pelos comportamentos gargalos identificados em cenários anteriores e demonstrou um bom desempenho ao reduzir efetivamente os tempos de espera e o tempo de permanência no sistema. Ao priorizar o agendamento e a otimização de recursos, o Cenário nº 7 visava obter uma alocação eficiente de recursos, mantendo uma abordagem realista e viável. Os resultados da simulação indicam que este cenário permitiria melhorar o fluxo de pacientes e reduzir os tempos de espera, contribuindo para melhorar o desempenho geral do sistema. Finalmente, o Cenário nº 8, segue as orientações da Sociedade Portuguesa de Endoscopia Digestiva, que sugerem que existam tempos mínimos de agendamento de acordo com o tipo de exame. Neste cenário o agendamento dos procedimentos endoscópicos foi realizado seguindo os tempos recomendados e que são de 40 minutos para os procedimentos de endoscopia digestiva alta e colonoscopia, 30 minutos para os procedimentos de colonoscopia e 15 minutos para os procedimentos de endoscopia digestiva alta. O objetivo deste cenário alternativo era otimizar o agendamento de procedimentos endoscópicos mantendo o desempenho operacional do modelo. Com este cenário, observou-se uma redução no número de exames realizados bem como na taxa de utilização de recursos. A diminuição do tempo de espera teve um papel significativo na redução da média e desvio padrão do tempo gasto no sistema, bem como do tempo xi máximo. As melhorias alcançadas no Cenário nº 8 indicam que este cenário poderia ter um impacto positivo na satisfação do paciente, pois contribui para a redução do tempo de espera e para um processo mais eficiente. Considerando as restrições e limitações inerentes à implementação no mundo real, o Cenário nº 7 é aquele que parece oferecer uma solução mais prática e viável, com potencial para otimizar o processo de calendarização de exames, levando a uma maior eficiência e melhoria na experiência do paciente.
The Outpatient Department of the HPA Saúde Group offers its patients a wide range of medical and surgical specialties. To provide better quality services with the current increase in patient needs, the hospital, at the beginning of 2021, expanded its facilities and restructured the outpatient service, increasing its responsiveness as well as the quality of its facilities and services provided to patients. The Special Exams Unit is one of the units that integrates this department, where specific exams are conducted in various specialties such as cardiology, gastroenterology, urology, and pulmonology. The unit was restructured and started operations in May 2021. In terms of operational management, this restructuring and expansion of the unit's physical capacity brought challenges, such as changing patient flows, redistribution of physical and human resources, and regarding the need for proper capacity planning. Capacity planning is complex and a major challenge for managers, as the imbalance between demand and supply can trigger increases in patient waiting times or waste of resources. Any of these negative outcomes jeopardize the performance of the service, making it less efficient. The intervention threshold in capacity planning is very tenuous, given factors such as unpredictability, seasonality, patient fluctuations throughout the day, and patient behavior patterns. Consequently, it is important to study the impact of this planning on the performance of the Special Exams Unit, so that strategies can be designed to improve the efficiency of the services, and as a result, their performance. This study aims to explore the Application of Discrete Event Simulation (DES) to study the impact of capacity planning on the performance of the Special Exams Unit. Its specific objectives are to identify which variables affect capacity planning and to identify relationships/interactions between capacity planning and patient waiting time. The application of DES will allow us to replicate the workings of the Unit and perform simulations to understand how the system works and how alternative strategies can impact the performance of the system. It will also help us identify what the impact of capacity planning is on service performance and what the relationship between capacity planning and patient waiting time is. The Special Exams unit operates five days a week, Monday through Friday. The exams are usually scheduled between 8:30 and 12:30 a.m. and between 14:00 to 18:30 p.m., and there may be changes in the schedule according to the number of exams scheduled. Given the complexity of the Unit and the exploratory nature of this study, only exams performed by the gastroenterology specialty were considered for the study. In order to develop and validate the simulation model, data were collected relating to all the exams performed on Thursdays at the Special Exams Unit between January and April 2022. During this period, 261 patients had exams scheduled, but 8 did not show up. The data used in the simulation refer, therefore, to data collected for 253 individual patients. These data were extracted from electronic record data from the hospital, nursing records from the patients’ handling process, and from the appointment-scheduling system. These data were recorded and analyzed using Microsoft Excel version 365 and the Statistical Package for the Social Sciences version 28.0.1.0. The Stat::fit® version 2 program was used to analyze the statistical distributions of the collected data. DES was applied in the modelling of the patients’ flow at the Special Exams Unit using the software program SIMUL8, version 29.0. The patients’ flow was mapped through interviews with experts from the Unit and through direct observation of the real process system. After data and statistical distribution analyses were performed, the simulation model was developed. To validate the DES model, the results from the baseline simulation were verified by experts from the Unit. The simulation animation was also analyzed by members of staff from the Unit, to confirm that patients in the model were following an appropriate flow. A comparative analysis was also performed through comparing real data and simulation output data. Most of the variables had acceptable results with minimal values of relative error. For example, the patients’ average time in the system showed a relative error of 0.07, the average time of Upper and Lower GI endoscopy, Lower GI endoscopy and Upper GI endoscopy, had relative errors of 0.01, xiv 0.00 and 0.00, respectively. One variable show, however, a relative error superior to 10%. For example, the indicator Standard Deviation regarding the total time spent by the patients in the system presents an error higher than 10%. These differences observed between real and simulated values might be associated with the model construction process. Some processes were deconstructed, in smaller and specific activities, with the objective of understanding better the resource utilization rate in the system. These discrepancies can eventually be attributed to the utilization of an average distribution to represent the time taken by certain activities instead of employing a better adjusted distribution due to the lack of some data. Eight scenarios were tested, with the objective of identifying limitations and bottlenecks, maximizing capacity and improving scheduling appointment and resource optimization. In Scenario No. 1, the number of appointments was optimized by reducing the scheduling intervals from 30 minutes to 20 minutes, while maintaining the same working hours used in the baseline simulation. The goal was to maximize the appointment efficiency of the model. A noticeable increase in the number of exams performed and in resource utilization was observed. However, this came at the cost of a significant increase in queuing time and the length of stay in the system. The overload on the system during this scenario allowed for the identification of a bottleneck behavior within the dressing room activity. The result of a block route of 26.1% in this activity, indicates that patients’ flow experienced an elevated level of congestion and became a limiting factor in the overall process flow. Adding a second dressing room on Scenario No.2, showed a decrease in the queuing times and also a decrease of the blocked route on the dressing room from 26.1% to 15.96%, validating experts’ opinion that this activity assumes a bottleneck behavior. Scenarios No. 3 and No. 4 were conducted to test resource optimization with the objective of identifying limitations and bottleneck behaviors in other activities or resources, considering that Scenario No. 2 still observed congestion in the dressing room activity. In Scenario No. 3, the addition of an extra nurse for the preparation and recovery process had a positive effect on the system. It resulted in a reduction in patient preparation xv time and length of stay on the system. The blocked route on the dressing room decreased from 15.96% to 13.80%. This reduction can be interpreted as the patient preparation and recovery process nurse behaving as a bottleneck resource, when only one nurse is available. In scenario No.3, the addition of one more nurse, contributed to the reduction of 2.62% in the blockage of the dressing room activity. However, it was also found that the resource utilization rate was not efficient, as the utilization rate of the patient preparation and recovery process nurse reached only 43%. On Scenario No. 4, the addition of two more beds helped alleviate the bottleneck behavior in the dressing room. This led to a significant decrease in the blocked route of this activity from 13.80% to 2.64%. However, it also resulted in a notable increase in the queuing time for exams and a slight increase in the patients’ length of stay in the system. Based on the findings from Scenario No. 4, which indicated a potential bottleneck in the exams’ procedure, a flow and resource optimization scenario was conducted (Scenario No. 5). In this scenario, a second procedure room was introduced, along with an independent team including medical professionals and nurses. The addition of the second procedure room resulted in a significant decrease in both patient length of stay in the system and queuing time. This confirmed that this activity can act as a bottleneck in the system. On Scenario No. 6, due to low resource utilization rates observed on Scenario No. 5, we used the same model but with the objective of maximizing capacity through resource and scheduling optimization. Regarding scheduling optimization, we assumed that a patient would be admitted each 12,5 minutes. Regarding resource optimization, we assumed that the same resources of Scenario No. 5 would be used (i.e., 2 dressing rooms and 2 procedure rooms, each one with one independent team and 6 beds). In this scenario we observed a reduction on the queuing times, shorter patient length of stay, and a notable increase in the number of exams performed. Scenario No. 6 displayed an elevated level of resource optimization by effectively utilizing the available resources, optimizing the flow of patients, and implementing an enhanced scheduling approach. This resulted in an improvement in efficiency, reduced waiting times, and increased throughput in terms of the number of exams conducted. xvi Scenario No. 6 emerged as the most efficient scenario tested, however, due to hypothetical constraints such as the need for physical modifications of the space and potential complications in the patients’ flow, a more realistic scenario, Scenario No. 7, was explored. It aimed to optimize the model from Scenario No. 1 by focusing on scheduling and resource optimization. In Scenario No. 7, appointments were scheduled at every 24,5 minutes, ensuring high resource utilization while considering the limitations and bottlenecks identified in previous scenarios. This scenario demonstrated reliable performance by effectively reducing queuing times and the patients’ length of stay in the system. By prioritizing scheduling and resource optimization, Scenario No. 7 aimed to achieve efficient resource allocation while maintaining a realistic and feasible approach. This scenario allowed for an improvement in patient flow and a reduction in the waiting times, contributing to an enhanced overall performance of the system. Finally, Scenario No. 8, follows the Portuguese Society of Digestive Endoscopy guidelines, which recommend different minimum scheduling times based on the type of procedure. For the Upper GI endoscopy, the recommendation is of 15 minutes, for the Lower GI endoscopy, it is 30 minutes and for the Upper and Lower GI endoscopy, it is 40 minutes. This scenario was tested to offer an alternative scheduling appointment optimization, using a Schedule of Arrivals. A dedicated spreadsheet was developed for manual control of patients' arrivals, following the guideline recommendations. This Scenario demonstrated a reduction in the number of exams conducted and in the resource utilization rate. The decrease in queuing times played a significant role in reducing the average and standard deviation of the time patients spent in the system, as well as the maximum time. The improvements achieved through Scenario No. 8 have a positive impact on patient satisfaction, as they contribute to shorter waiting times and to a more efficient process. Considering the constraints and limitations inherent to real-world implementations, we believe that Scenario No. 7 provides a practical and viable solution that optimizes the scheduling process, resulting in an improvement in efficiency and in patients’ experience.
The Outpatient Department of the HPA Saúde Group offers its patients a wide range of medical and surgical specialties. To provide better quality services with the current increase in patient needs, the hospital, at the beginning of 2021, expanded its facilities and restructured the outpatient service, increasing its responsiveness as well as the quality of its facilities and services provided to patients. The Special Exams Unit is one of the units that integrates this department, where specific exams are conducted in various specialties such as cardiology, gastroenterology, urology, and pulmonology. The unit was restructured and started operations in May 2021. In terms of operational management, this restructuring and expansion of the unit's physical capacity brought challenges, such as changing patient flows, redistribution of physical and human resources, and regarding the need for proper capacity planning. Capacity planning is complex and a major challenge for managers, as the imbalance between demand and supply can trigger increases in patient waiting times or waste of resources. Any of these negative outcomes jeopardize the performance of the service, making it less efficient. The intervention threshold in capacity planning is very tenuous, given factors such as unpredictability, seasonality, patient fluctuations throughout the day, and patient behavior patterns. Consequently, it is important to study the impact of this planning on the performance of the Special Exams Unit, so that strategies can be designed to improve the efficiency of the services, and as a result, their performance. This study aims to explore the Application of Discrete Event Simulation (DES) to study the impact of capacity planning on the performance of the Special Exams Unit. Its specific objectives are to identify which variables affect capacity planning and to identify relationships/interactions between capacity planning and patient waiting time. The application of DES will allow us to replicate the workings of the Unit and perform simulations to understand how the system works and how alternative strategies can impact the performance of the system. It will also help us identify what the impact of capacity planning is on service performance and what the relationship between capacity planning and patient waiting time is. The Special Exams unit operates five days a week, Monday through Friday. The exams are usually scheduled between 8:30 and 12:30 a.m. and between 14:00 to 18:30 p.m., and there may be changes in the schedule according to the number of exams scheduled. Given the complexity of the Unit and the exploratory nature of this study, only exams performed by the gastroenterology specialty were considered for the study. In order to develop and validate the simulation model, data were collected relating to all the exams performed on Thursdays at the Special Exams Unit between January and April 2022. During this period, 261 patients had exams scheduled, but 8 did not show up. The data used in the simulation refer, therefore, to data collected for 253 individual patients. These data were extracted from electronic record data from the hospital, nursing records from the patients’ handling process, and from the appointment-scheduling system. These data were recorded and analyzed using Microsoft Excel version 365 and the Statistical Package for the Social Sciences version 28.0.1.0. The Stat::fit® version 2 program was used to analyze the statistical distributions of the collected data. DES was applied in the modelling of the patients’ flow at the Special Exams Unit using the software program SIMUL8, version 29.0. The patients’ flow was mapped through interviews with experts from the Unit and through direct observation of the real process system. After data and statistical distribution analyses were performed, the simulation model was developed. To validate the DES model, the results from the baseline simulation were verified by experts from the Unit. The simulation animation was also analyzed by members of staff from the Unit, to confirm that patients in the model were following an appropriate flow. A comparative analysis was also performed through comparing real data and simulation output data. Most of the variables had acceptable results with minimal values of relative error. For example, the patients’ average time in the system showed a relative error of 0.07, the average time of Upper and Lower GI endoscopy, Lower GI endoscopy and Upper GI endoscopy, had relative errors of 0.01, xiv 0.00 and 0.00, respectively. One variable show, however, a relative error superior to 10%. For example, the indicator Standard Deviation regarding the total time spent by the patients in the system presents an error higher than 10%. These differences observed between real and simulated values might be associated with the model construction process. Some processes were deconstructed, in smaller and specific activities, with the objective of understanding better the resource utilization rate in the system. These discrepancies can eventually be attributed to the utilization of an average distribution to represent the time taken by certain activities instead of employing a better adjusted distribution due to the lack of some data. Eight scenarios were tested, with the objective of identifying limitations and bottlenecks, maximizing capacity and improving scheduling appointment and resource optimization. In Scenario No. 1, the number of appointments was optimized by reducing the scheduling intervals from 30 minutes to 20 minutes, while maintaining the same working hours used in the baseline simulation. The goal was to maximize the appointment efficiency of the model. A noticeable increase in the number of exams performed and in resource utilization was observed. However, this came at the cost of a significant increase in queuing time and the length of stay in the system. The overload on the system during this scenario allowed for the identification of a bottleneck behavior within the dressing room activity. The result of a block route of 26.1% in this activity, indicates that patients’ flow experienced an elevated level of congestion and became a limiting factor in the overall process flow. Adding a second dressing room on Scenario No.2, showed a decrease in the queuing times and also a decrease of the blocked route on the dressing room from 26.1% to 15.96%, validating experts’ opinion that this activity assumes a bottleneck behavior. Scenarios No. 3 and No. 4 were conducted to test resource optimization with the objective of identifying limitations and bottleneck behaviors in other activities or resources, considering that Scenario No. 2 still observed congestion in the dressing room activity. In Scenario No. 3, the addition of an extra nurse for the preparation and recovery process had a positive effect on the system. It resulted in a reduction in patient preparation xv time and length of stay on the system. The blocked route on the dressing room decreased from 15.96% to 13.80%. This reduction can be interpreted as the patient preparation and recovery process nurse behaving as a bottleneck resource, when only one nurse is available. In scenario No.3, the addition of one more nurse, contributed to the reduction of 2.62% in the blockage of the dressing room activity. However, it was also found that the resource utilization rate was not efficient, as the utilization rate of the patient preparation and recovery process nurse reached only 43%. On Scenario No. 4, the addition of two more beds helped alleviate the bottleneck behavior in the dressing room. This led to a significant decrease in the blocked route of this activity from 13.80% to 2.64%. However, it also resulted in a notable increase in the queuing time for exams and a slight increase in the patients’ length of stay in the system. Based on the findings from Scenario No. 4, which indicated a potential bottleneck in the exams’ procedure, a flow and resource optimization scenario was conducted (Scenario No. 5). In this scenario, a second procedure room was introduced, along with an independent team including medical professionals and nurses. The addition of the second procedure room resulted in a significant decrease in both patient length of stay in the system and queuing time. This confirmed that this activity can act as a bottleneck in the system. On Scenario No. 6, due to low resource utilization rates observed on Scenario No. 5, we used the same model but with the objective of maximizing capacity through resource and scheduling optimization. Regarding scheduling optimization, we assumed that a patient would be admitted each 12,5 minutes. Regarding resource optimization, we assumed that the same resources of Scenario No. 5 would be used (i.e., 2 dressing rooms and 2 procedure rooms, each one with one independent team and 6 beds). In this scenario we observed a reduction on the queuing times, shorter patient length of stay, and a notable increase in the number of exams performed. Scenario No. 6 displayed an elevated level of resource optimization by effectively utilizing the available resources, optimizing the flow of patients, and implementing an enhanced scheduling approach. This resulted in an improvement in efficiency, reduced waiting times, and increased throughput in terms of the number of exams conducted. xvi Scenario No. 6 emerged as the most efficient scenario tested, however, due to hypothetical constraints such as the need for physical modifications of the space and potential complications in the patients’ flow, a more realistic scenario, Scenario No. 7, was explored. It aimed to optimize the model from Scenario No. 1 by focusing on scheduling and resource optimization. In Scenario No. 7, appointments were scheduled at every 24,5 minutes, ensuring high resource utilization while considering the limitations and bottlenecks identified in previous scenarios. This scenario demonstrated reliable performance by effectively reducing queuing times and the patients’ length of stay in the system. By prioritizing scheduling and resource optimization, Scenario No. 7 aimed to achieve efficient resource allocation while maintaining a realistic and feasible approach. This scenario allowed for an improvement in patient flow and a reduction in the waiting times, contributing to an enhanced overall performance of the system. Finally, Scenario No. 8, follows the Portuguese Society of Digestive Endoscopy guidelines, which recommend different minimum scheduling times based on the type of procedure. For the Upper GI endoscopy, the recommendation is of 15 minutes, for the Lower GI endoscopy, it is 30 minutes and for the Upper and Lower GI endoscopy, it is 40 minutes. This scenario was tested to offer an alternative scheduling appointment optimization, using a Schedule of Arrivals. A dedicated spreadsheet was developed for manual control of patients' arrivals, following the guideline recommendations. This Scenario demonstrated a reduction in the number of exams conducted and in the resource utilization rate. The decrease in queuing times played a significant role in reducing the average and standard deviation of the time patients spent in the system, as well as the maximum time. The improvements achieved through Scenario No. 8 have a positive impact on patient satisfaction, as they contribute to shorter waiting times and to a more efficient process. Considering the constraints and limitations inherent to real-world implementations, we believe that Scenario No. 7 provides a practical and viable solution that optimizes the scheduling process, resulting in an improvement in efficiency and in patients’ experience.
Description
Keywords
Planeamento de capacidade Discrete event simulation Fluxo de pacientes Unidade de ambulatório Endoscopia