Activity Diagrams for Admissions in Hospital Management - DOC

Document Sample
Activity Diagrams for Admissions in Hospital Management - DOC Powered By Docstoc
   I.   Problem Formulation
  II.   Setting of Objectives and Overall Project Plan
 III.   Model Building
 IV.    Data Collection
  V.    Coding
 VI.    Verification
VII.    Validation
VIII.   Experimental Design
 IX.    Production Runs and Analysis
  X.    Repeat of 9 if necessary
 XI.    Documentation of Program and Reporting of Results
XII.    Implementation of Proposed System

        “Healthcare applications tend to fall into two major categories (1) „analytic‟
         decisions with uncertain components; and (2) comparison of alternative systems
         for determining resource or scheduling requirements.” (1)
        “Simulation can also be used for modeling input to more complex types of
         analytic decisions, such as those employing Markov models or decision trees,
         which are often used for representing patient flow through various disease states.”
        “…the analysis of operational systems generally requires a comparison of
         alternative systems to obtain the desired information.” (1)

        “Increased pressure from competitions, health care reform, reimbursement
         difficulties, and rising health care costs are primarily responsible for the high
         level of interest in this, and other ED operating efficiency issues.” (5)

        “Simulation modeling can capture the complexities of patient flow and allows
         decision makers to compare proposed capacity planning scenarios in a rational,
         objective manner.” (10)

        “…simulation models could be used to assist hospital planners in evaluating the
         effectiveness of a geriatric department by experimenting with different policy
         parameters such as level of emptiness, number of beds available for each
         compartment, conversion rates, length of stay and admissions. They can also be
         used to demonstrate to hospital planners and clinicians the long term effects of
         any radical change in the current system.” (12)


      “The first objective of this study was to determine the bottlenecks for in-in-patient
      “The second objective was to study the impact of bed availability on the waiting
       time of admitted patients in ED before being transferred to assigned beds in other
       units of the hospital.”(2)

      “Staffing and utilization of ED nurse and physician resources is mainly a concern
       because of expense, but it is also significant because of its impact on patient
       throughput and overall system performance.”(5)
      “The main goal of this simulation analysis is to select the best scheduling
       alternative (i.e. minimize the total average patient time within the ED)” (5)

      “The ultimate objective of the research reported herein is to design an admissions
       scheduling system which can control hospital occupancy.” (6)

Objectives of Oncology Ward Simulation(8)
    “The first objective was to model, analyze and improve patient flow processes
       and increase capacity in the main facility for both the medical oncology practice
       and the Ambulatory Treatment Center.”
    “A second and perhaps the primary objective was to translate this model to a new
       building which was being designed.”

      “If model output is extremely sensitive to the estimates [distribution estimates for
       certain variables in the model], then you may want to recommend to your client
       that they put some time and effort into initiating a primary data collection effort,
       or at least pay close attention to the estimation process.” (1)

      Attention must be given, once the problem and goals are defined, to the type and
       amount of data necessary for model construction. (2)
      “First of all, data pertaining to daily volume of ED and other units of the hospital
       were collected. Based on this data, the percentages of admitted and discharged
       ED patients were determined. The distribution of the arrivals was determined as
       Poisson using the statistical software package STAT:FIT.” (2)
      Patient types were defined, with six distinct levels. Higher levels were used to
       denote a longer stay in the ED; data was collected on the arrival cycles for these
       six types of patients. (2)

Data Collection Methods (5)
    Patient Visit Time Study – information about each stage of ED department visit
      collected through the use time study
    Service Distribution Time Study – info on amount of time staff spent with
      patients – gathered from a separate time study completed over a week
       Patient Arrival Processes – determine arrival rates of patients – gathered from a
        computerized patient tracking system database
       Transport and Routing times – distances between locations measured and random
        walking velocities approximated

Patient Types and Classifications (10)
     A patient type or class refers to a group of patients consuming a similar level of
        hospital resources. Such resources include the sequence of hospital units visited
        and the corresponding length of stay in those units. We will refer to this as
        patient classification problem.”
     “…too many patient types become very unmanageable from a modeling
     “The first stage of their method is concerned with splitting or grouping patient
        subpopulations based on statistical tests on the equality of transition probability
        matrices and length of stay distributions.”
     “…objective…is to explore the potential of using data mining techniques,
        specifically clustering techniques such as K-means, to help guide the development
        of patient type definitions for the purpose of building computer simulations or
        analytical models of patient flow in hospitals.”
     “It has been suggested by researchers that occupancy in obstetrical units can be
        reasonably well modeled by the Poisson distribution having a mean equal to the
        mean arrival rate of patients multiplied by the average length of stay.”
     “The study…identified patient types having fundamentally different resource
        needs, either with respect to the path taken through the system, the distribution of
        time spent in each area, or the amount of nursing care needed.”
     “These patient types were developed primarily through clinical expert opinion and
        accumulated domain knowledge of the modelers.”
     Diagnosis Related Groups used by health care to assess and bill costs based on
        type of treatments required; “…for clustering patient diagnoses and procedures
        into a manageable number of clinically meaningful categories.”
     “DRGs alone do not capture some distinctions that can be important for detailed
        patient flow simulation. For example, DRGs do not capture distinctions between
        scheduled and unscheduled patients or between patients entering the hospital
        through the emergency department versus an elective admission.”
     When examining the data sets, “…the OB/Gyn population has paths that are
        relatively homogenous. When one begins to widen the scope of the model and to
        capture more general medical/surgical patient populations, the problem becomes
        much more difficult.”

   “…the model is defined not in terms of a series of equations, but in terms of the
    physical movement of a transaction (e.g. patient, lab test) over time through different
    facilities or resources.”(1)
   “…the simulation model includes resources as one of the input variables, and the
    analyst must try different values of the input variable(s) and examine the effects of
    those values on the output variables of interest (e.g. utilization, delays in service,
    turnaways, etc.) Different values are tried until one is found that produces acceptable
    performance.” (1)
   “In business process reengineering, a simulation model of current processes can be
    constructed, then used to identify bottlenecks and underutilized resources.” (1)
   “The simpler the design of the model, the faster it will get completed, the sooner you
    will have some results, and the happier your clients will be.” (1)

Modeling an ED (2) Insert Copy of Flow Chart
  I.      Defining Locations and Corresponding Entity Types Within ED
          A. Core Areas
          B. Fast Track Area
          C. Preliminary Evaluation
          D. Observation
  II.     Defining Locations and Entity Types for Outside ED
          A. M/S/O/O
          B. Step-Down
          C. Women-Children
          D. ICU
          E. OR
          F. Post Anesthesia Care Unit
  III.    Simulation project was modeled as a queuing model with capacity constraints
          at each location

Modeling inflow in an ED (5)
  I.      Sources of incoming patients
          A. Walk-in arrivals
                    1. Registration
                    2. Triage area
                    3. BP, temp, initial assessment
                    4. Wait for admission to designated wing or treatment unit in ED
          B. Ambulance patient arrivals
                    1. Classified trauma or non-trauma
                    2. NT go directly to ER and follow same flow as walk-ins
                    3. Trauma are taken to adult wing of ED for immediate attention
          C. Helicopter patient arrivals
                    1. Immediately taken to adult wing of ED
                    2. Minor emergencies are treated generally over a long period of
                       time once stabilized
                    3. Major emergencies are generally taken to OR
  II.     Emergency Department Wings
          A. Adult Care Wing
          B. Chest Pain Center
           C. Pediatric Care Wing
           D. Minor Emergency Area
   III.    Means of Exiting ED
           A. Admitted to hospital
           B. Discharged from ED
           C. Balking
           D. Death
   IV.     Modeling the ED
           A. Use of patient flow diagrams and corresponding activity flow diagrams
           B. Arena and pre-constructed modules used
           C. Simplifying assumptions
                    i. All patients remain at same acuity level throughout stay
                   ii. All trauma patients treated as high acuity level patients rather than
                       create separate entity type
                  iii. At midnight all minor emergency patients removed from mea
                       waiting area and redefined as adult wing patients
   V.      Entity types
           A. ED patients
           B. Phone calls and other indirect care activities
           C. Logical entities for initializing the model and generating patient arrival
              rates for walk-in, ambulance, and helicopter arrivals

Design of Hospital Admissions Scheduling System (6)
   I.     General Process
          A. Developed for two different hospitals
          B. Patient flow diagrams were developed for both hospitals, with lists of each
             location a patient could go from the current location
          C. “The distribution of admissions across the possible flow patterns for each
             admitting bed service is one of the model inputs, and is based on historical
          D. “Secondary patient flow patterns for each site were identified and modeled
             for patients who cannot follow the primary, or desired, flow patterns
             because of a lack of available beds.”
   II.    Theoretical vs. Empirical Distributions
          A. “Use of theoretical rather than empirical distributions for these variables
             facilitates model implementation because a hospital can use summary data
             as model input, rather than have to perform a detailed analysis of patient-
             specific data. Specifically, theoretical distribution‟s parameters as model
             input; while empirical distributions require the hospital to develop a
             distribution from individual patient records.”

Modeling a Cancer Treatment Center (8)
Insert Flow Charts Here
  I. Flow of patients
          I. Patient arrives for appointment
         II. Port Y/N?
                 a. Y = taken to port room to draw blood
                 b. N = taken to blood drawing room
       III. Back to waiting room
       IV. To exam room when made available
        V. Patient evaluation after blood results
                 a. ATC treatment
                 b. Sent home
       VI. Check out
      VII. Wait time of 20 to 70 minutes before seeing a doctor; goal to hit 30
 II. Simulation Modeling
           A. Done in Arena
           B. Visio Technical used to draw layout of facility

Mathematical Modeling (3)
    “If the time taken for each stage is known it should be possible to construct a
      mathematical model that represents the activity of the department.”
    “A model that accurately reflected a real A&E department would allow an
      exploration of the effect of different configurations of service delivery without the
      large organizational changes that are required if changes are tried out in the real

Cancer Treatment Center (8)
    “Simulated patient‟s arrivals were generated from historical data gathered from
       the hospital personnel in charge of patient scheduling.”
    “…the procedure used to determine the duration of these [list of various times at
       each stage of the process] activities was through expert opinion.”
    “Uniforms distributions were used in many situations as well as triangular
       distributions containing as parameters the expert opinions about minimum,
       maximum and most likely duration of each activity.”

Geriatric Patient Model (12)
    “Discrete event simulation concerns the modeling of a system as it evolves over
        time by a representation in which the state variables change instantaneously at
        separate points in time.”
    “In this model the entities are patients.”
    “In this model the activities are the three compartments and the queues.”
    “System state: this is the collection of state variables necessary to describe the
        system at a certain point in time, for example, the number of available beds, the
        waiting time in a queue, etc.”
    “In our model, beds are the server units and patients are the customers.”
    “In the behavioral theory of flow of beds occupied by patients in the different
        streams of acute, rehabilitative and long-stay care are separated by decision
        making thresholds.”
    “A queuing system is characterized by three components: arrival process, service
        mechanism, and queue discipline.”
      “…interval times can be considered to be independent, identically distributed
       variables. Hence they can be described by an exponential distribution. Thus, the
       number of patients arriving in the hospital follows a Poison distribution.”
      “The service mechanism is described by the number of servers, the number of
       queues, and the probability distribution of customer service times.”
      “The service times in each compartment for each patient are IID random variables
       that can be described by an exponential distribution.
      UNCONSTRAINED MODEL (Insert flowchart)
      BASIC MODEL (Insert flowchart)


      “Demonstration of a simulation model‟s validity – i.e. its ability to accurately
       represent the system under investigation – is key to the acceptance of simulation
       as a technique” (1)

ED Model Verification/Validation (2)
     o Discussion of model and behavior with hospital administrators
     o “The performance measure, average waiting time of admitted patients in ED
         before bed placement, was compared with the historical data. The histogram
         for historical data and that for simulation model results….had similar shapes.”

List of Suggestions for Verification (5)
        o Have someone familiar with the system (other than the developer) check the
           computer simulation model for problems.
        o Generate an activity flow diagram of the system. This should include logic
           for all possible activities an entity may encounter while in the system.
        o Examine the reasonableness of the model output for a variety of input
           parameter values. A wide variety of output statistics should be used for this
        o If possible, animate the computer model to verify that what is seen in the
           animation imitates the behavior of the actual system.
Three Step Approach for Validation (5)
        o Face validation – asking model users and others who are knowledgeable about
           the actual system being modeled, whether or not the model and behavior are
        o Validation of model assumptions – “Any data assumptions used in
           constructing the simulation model or specifying the model‟s input parameters”
           should be validated.
        o Validating I/O transformations – Achieved by testing the simulation model‟s
           ability to predict the future (or past) behavior of the real world system being

Model Validation (6)
      “A graphical methodology, rather than hypothesis testing, was used for evaluating
       the validity of the admissions scheduling model because observations of daily
       census are autocorrelated, and because limited data are available from the actual
       system. When observations are autocorrelated, classical statistical tests based on
       identically and independently distributed observations are not directly applicable.
       Furthermore, because a simulation model is only an approximation of the actual
       system, the null hypothesis that model behavior and system behavior are the same
       will almost certainly be false…it may still be valid for the purpose for which it is
       intended. This is especially true for models that are designed primarily for
       comparing alternatives than for predicting absolute answers, as is the admissions
       scheduling model.”

Cancer Treatment Center (8)
    “Validation is the process of raising to an acceptable level of user‟s confidence
       that any simulation derived inference about the system is correct.”
    “..we used other techniques that involve simulated animation and the customer
       (hospital personnel) directly in the validation process.”
    “We applied this strategy by showing the simulation animation and results to the
       process improvement team and asked them their opinion about different aspects
       of the system, such as queue length, number of busy exam rooms, number of busy
       treatment chairs, etc.”


      “In order to evaluate the impact of bed availability on average waiting time of
       admitted patients in the ED before bed placement, different „what-if‟ scenarios
       were tried by adding additional in-patient beds to various locations.”(2)

Alternatives tested for staffing schedules (5)
   1. Schedules based on ED manager decisions
   2. Maintain an 8 hour double coverage shift and stagger when the shift change
       would occur (used to staff at peak inflow)
   3. The addition of a shift in which there is double coverage (increase from one two
       to double coverage shifts)
   4. Shift where the double coverage shift would fall (different from when the rest of
       the staffing changes might occur)

Barriers to Implementation (9)
    “…health care providers are reluctant to embrace computer-based models of
       patient care processes, because these processes are simply too complicated to be
       reduced to representation by a model.”
    Slow acceptance due to “…lack of incentives and…managers‟ continued
     dependence on deterministic decision making.”
    Historically, “…there were few incentives to provide services as efficiently as
     possible, thus precluding the need for a tool such as simulation, which is most
     frequently used for determining the efficient allocation of scarce resources.”
    “Health care managers have instead relied primarily on simpler, more
     deterministic analyses. Common control systems consist of databases of
     comparative information.”
    List of Barriers
     o A natural resistance to change
     o Resistance to “…the dehumanizing nature of time and motion analysis.”
     o A poorly conducted simulation
    “Often simulation of a proposal purchase or new facility is viewed by these
     engineers as requiring too great an investment time to be completed to support
     administration‟s decision making. Additionally, these management engineers do
     not possess significant training in simulation and therefore require tools that can
     reduce the perceived learning curve.”
    “The greatest barrier to implementing simulation in the health care industry is the
     people most interested in promoting simulation. Their strong technical nature
     does not allow them either to properly appreciate simulation or to promote the use
     of the tool within their own organization.”
     o “…in promoting simulation, they may perpetuate a number of myths, which
         can serve as barriers to the acceptance of simulation. These include such
         beliefs as, „Simulation projects must be difficult;‟ „Analyze simulation results
         with an eye towards precision;‟ and „Build models of everything.‟”
     o “The more critical cause of the failure of simulation in health care is the
         techno-babble utilized by these engineers to sell simulation. This techno-
         babble causes management and operations personnel to lose interest in this
         technology. As a result, a tool that can be critical to the success of the health
         care industry is largely ignored.”
    “A major barrier to implementation of these recommendations is a failure of the
     sponsoring manager to follow through on the studies‟ recommendations.”
     o May resent the executive [who decides to make the decisions]
     o May not have complete authority over all affected areas
     o Not strong enough to overcome obstacles that will arise
     o Cooperation from external personnel is required but impossible to obtain
     o Not enough time to complete regular responsibilities and additional charges

    “[After implementation] consider collecting data on the effect of the change on
     the system‟s performance and then comparing actual performance with model
    “A likely reason for observed discrepancies between model predictions and actual
     performance is a corresponding discrepancy between predicted and actual input
     data.” (1)
Decision-Making System (6)
    The scheduling system is incorporated into the simulation model, then
       combinations of different values of the parameters are systematically tried as
       model input, along with the hospital‟s historic or projected values for the other
       input variables.”
    “Model output includes predicted values of hospital performance, which will be
       reviewed by hospital clinical and administrative staff. The values of the
       scheduling parameters which result in a desirable combination of values for daily
       census, overall occupancy, cancellations, and turnaways become the parameters
       of the scheduling system to be implemented in the hospital.”
    “…simple software program which performs a series of calculations using the
       parameter values plus daily data on bed availability. The results from the
       calculations will be used y admitting personnel in their daily admitting and
       scheduling decisions.”
           o “How many elective admissions should be scheduled, by bed service, by
               day of the week?”
           o “…how many patients should be called in?”
           o “…how many scheduled admissions should be canceled (to ensure bed
               availability for emergency patients)?”

Shared By:
Description: Activity Diagrams for Admissions in Hospital Management document sample